Release: v4.0.1

* conda

* Guide

* correct tag

* Update README.md

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Update docs/source/installation.md

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Sylvain's comments

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

.circleci/TROUBLESHOOT.md

@@ -1,7 +0,0 @@
-# Troubleshooting
-
-This is a document explaining how to deal with various issues on Circle-CI. The entries may include actually solutions or pointers to Issues that cover those.
-
-## Circle CI
-
-* pytest worker runs out of resident RAM and gets killed by `cgroups`: https://github.com/huggingface/transformers/issues/11408

.circleci/deploy.sh

@@ -0,0 +1,56 @@
+cd docs
+
+function deploy_doc(){
+	echo "Creating doc at commit $1 and pushing to folder $2"
+	git checkout $1
+	if [ ! -z "$2" ]
+	then
+		if [ "$2" == "master" ]; then
+		    echo "Pushing master"
+			make clean && make html && scp -r -oStrictHostKeyChecking=no _build/html/* $doc:$dir/$2/
+			cp -r _build/html/_static .
+		elif ssh -oStrictHostKeyChecking=no $doc "[ -d $dir/$2 ]"; then
+			echo "Directory" $2 "already exists"
+			scp -r -oStrictHostKeyChecking=no _static/* $doc:$dir/$2/_static/
+		else
+			echo "Pushing version" $2
+			make clean && make html
+			rm -rf _build/html/_static
+			cp -r _static _build/html
+			scp -r -oStrictHostKeyChecking=no _build/html $doc:$dir/$2
+		fi
+	else
+		echo "Pushing stable"
+		make clean && make html
+		rm -rf _build/html/_static
+		cp -r _static _build/html
+		scp -r -oStrictHostKeyChecking=no _build/html/* $doc:$dir
+	fi
+}
+
+# You can find the commit for each tag on https://github.com/huggingface/transformers/tags
+deploy_doc "master" master
+deploy_doc "b33a385" v1.0.0
+deploy_doc "fe02e45" v1.1.0
+deploy_doc "89fd345" v1.2.0
+deploy_doc "fc9faa8" v2.0.0
+deploy_doc "3ddce1d" v2.1.1
+deploy_doc "3616209" v2.2.0
+deploy_doc "d0f8b9a" v2.3.0
+deploy_doc "6664ea9" v2.4.0
+deploy_doc "fb560dc" v2.5.0
+deploy_doc "b90745c" v2.5.1
+deploy_doc "fbc5bf1" v2.6.0
+deploy_doc "6f5a12a" v2.7.0
+deploy_doc "11c3257" v2.8.0
+deploy_doc "e7cfc1a" v2.9.0
+deploy_doc "7cb203f" v2.9.1
+deploy_doc "10d7239" v2.10.0 
+deploy_doc "b42586e" v2.11.0
+deploy_doc "7fb8bdf" v3.0.2
+deploy_doc "4b3ee9c" v3.1.0
+deploy_doc "3ebb1b3" v3.2.0
+deploy_doc "0613f05" v3.3.1
+deploy_doc "eb0e0ce" v3.4.0
+deploy_doc "818878d" v3.5.1
+deploy_doc "28d3ccd" # v4.0.1 Latest stable release

.gitattributes

@@ -1,4 +0,0 @@
-*.py	eol=lf
-*.rst	eol=lf
-*.md	eol=lf
-*.mdx   eol=lf

.github/ISSUE_TEMPLATE/---new-benchmark.md

@@ -0,0 +1,22 @@
+---
+name: "\U0001F5A5 New benchmark"
+about: Benchmark a part of this library and share your results
+title: "[Benchmark]"
+labels: ''
+assignees: ''
+
+---
+
+# 🖥 Benchmarking `transformers`
+
+## Benchmark
+
+Which part of `transformers` did you benchmark?
+
+## Set-up
+
+What did you run your benchmarks on? Please include details, such as: CPU, GPU? If using multiple GPUs, which parallelization did you use?
+
+## Results
+
+Put your results here!

.github/ISSUE_TEMPLATE/--new-model-addition.md

@@ -0,0 +1,20 @@
+---
+name: "\U0001F31F New model addition"
+about: Submit a proposal/request to implement a new Transformer-based model
+title: ''
+labels: New model
+assignees: ''
+
+---
+
+# 🌟 New model addition
+
+## Model description
+
+<!-- Important information -->
+
+## Open source status
+
+* [ ] the model implementation is available: (give details)
+* [ ] the model weights are available: (give details)
+* [ ] who are the authors: (mention them, if possible by @gh-username)

.github/ISSUE_TEMPLATE/bug-report.md

@@ -0,0 +1,81 @@
+---
+name: "\U0001F41B Bug Report"
+about: Submit a bug report to help us improve transformers
+title: ''
+labels: ''
+assignees: ''
+
+---
+
+
+## Environment info
+<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
+     Don't forget to fill out the missing fields in that output! -->
+     
+- `transformers` version:
+- Platform:
+- Python version:
+- PyTorch version (GPU?):
+- Tensorflow version (GPU?):
+- Using GPU in script?:
+- Using distributed or parallel set-up in script?:
+
+### Who can help
+<!-- Your issue will be replied to more quickly if you can figure out the right person to tag with @
+ If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
+ Please tag fewer than 3 people.
+ 
+ albert, bert, GPT2, XLM: @LysandreJik 
+ tokenizers: @mfuntowicz
+ Trainer: @sgugger
+ Speed and Memory Benchmarks: @patrickvonplaten
+ Model Cards: @julien-c
+ TextGeneration: @TevenLeScao 
+ examples/distillation: @VictorSanh
+ nlp datasets: [different repo](https://github.com/huggingface/nlp)
+ rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
+ Text Generation: @patrickvonplaten @TevenLeScao
+ Blenderbot: @patrickvonplaten
+ Bart: @patrickvonplaten
+ Marian: @patrickvonplaten
+ Pegasus: @patrickvonplaten
+ mBART: @patrickvonplaten
+ T5: @patrickvonplaten
+ Longformer/Reformer: @patrickvonplaten
+ TransfoXL/XLNet: @TevenLeScao
+ RAG: @patrickvonplaten, @lhoestq
+ FSMT: @stas00
+ examples/seq2seq: @patil-suraj
+ examples/bert-loses-patience: @JetRunner
+ tensorflow: @jplu
+ examples/token-classification: @stefan-it
+ documentation: @sgugger
+ -->
+
+## Information
+
+Model I am using (Bert, XLNet ...):
+
+The problem arises when using:
+* [ ] the official example scripts: (give details below)
+* [ ] my own modified scripts: (give details below)
+
+The tasks I am working on is:
+* [ ] an official GLUE/SQUaD task: (give the name)
+* [ ] my own task or dataset: (give details below)
+
+## To reproduce
+
+Steps to reproduce the behavior:
+
+1.
+2.
+3.
+
+<!-- If you have code snippets, error messages, stack traces please provide them here as well.
+     Important! Use code tags to correctly format your code. See https://help.github.com/en/github/writing-on-github/creating-and-highlighting-code-blocks#syntax-highlighting
+     Do not use screenshots, as they are hard to read and (more importantly) don't allow others to copy-and-paste your code.-->
+
+## Expected behavior
+
+<!-- A clear and concise description of what you would expect to happen. -->

.github/ISSUE_TEMPLATE/bug-report.yml

@@ -1,119 +0,0 @@
-name: "\U0001F41B Bug Report"
-description: Submit a bug report to help us improve transformers
-labels: [ "bug" ]
-body:
-  - type: textarea
-    id: system-info
-    attributes:
-      label: System Info
-      description: Please share your system info with us. You can run the command `transformers-cli env` and copy-paste its output below.
-      placeholder: transformers version, platform, python version, ...
-    validations:
-      required: true
-
-  - type: textarea
-    id: who-can-help
-    attributes:
-      label: Who can help?
-      description: |
-        Your issue will be replied to more quickly if you can figure out the right person to tag with @
-        If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
-        Please tag fewer than 3 people.
-        
-        Models:
-
-          - ALBERT, BERT, XLM, DeBERTa, DeBERTa-v2, ELECTRA, MobileBert, SqueezeBert: `@LysandreJik`
-          - T5, Pegasus, EncoderDecoder: `@patrickvonplaten`
-          - Blenderbot, MBART, BART, Marian, Pegasus: `@patil-suraj`
-          - Reformer, TransfoXL, XLNet, FNet: `@patrickvonplaten`
-          - Longformer, BigBird: `@ydshieh`
-          - FSMT: `@stas00`
-          - Funnel: `@sgugger`
-          - GPT-2, GPT: `@patil-suraj`, `@patrickvonplaten`, `@LysandreJik`
-          - RAG, DPR: `@patrickvonplaten`, `@lhoestq`
-          - TensorFlow: `@Rocketknight1`
-          - JAX/Flax: `@patil-suraj`
-          - TAPAS, LayoutLM, LayoutLMv2, LUKE, ViT, BEiT, DEiT, DETR, CANINE: `@NielsRogge`
-          - GPT-Neo, GPT-J, CLIP: `@patil-suraj`
-          - Wav2Vec2, HuBERT, UniSpeech, UniSpeechSAT, SEW, SEW-D: `@patrickvonplaten`, `@anton-l`
-          - SpeechEncoderDecoder, Speech2Text, Speech2Text2: `@sanchit-gandhi`, `@patrickvonplaten`, `@anton-l`
-          
-          If the model isn't in the list, ping `@LysandreJik` who will redirect you to the correct contributor.
-
-        Library:
-          - Benchmarks: `@patrickvonplaten`
-          - Deepspeed: `@stas00`
-          - Ray/raytune: `@richardliaw`, `@amogkam`
-          - Text generation: `@patrickvonplaten`, `@Narsil`, `@gante`
-          - Tokenizers: `@SaulLu`
-          - Trainer: `@sgugger`
-          - Pipelines: `@Narsil`
-          - Speech: `@patrickvonplaten`, `@anton-l`, `@sanchit-gandhi`
-          - Vision: `@NielsRogge`, `@sgugger`
-
-        Documentation: `@sgugger`, `@stevhliu`
-
-        Model hub:
-
-          - for issues with a model, report at https://discuss.huggingface.co/ and tag the model's creator.
-
-        HF projects:
-
-          - datasets: [different repo](https://github.com/huggingface/datasets)
-          - rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
-
-        Examples:
-
-          - maintained examples (not research project or legacy): `@sgugger`, `@patil-suraj`
-
-        For research projetcs, please ping the contributor directly. For example, on the following projects:
-
-          - research_projects/bert-loses-patience: `@JetRunner`
-          - research_projects/distillation: `@VictorSanh`
-      placeholder: "@Username ..."
-
-  - type: checkboxes
-    id: information-scripts-examples
-    attributes:
-      label: Information
-      description: 'The problem arises when using:'
-      options:
-        - label: "The official example scripts"
-        - label: "My own modified scripts"
-
-  - type: checkboxes
-    id: information-tasks
-    attributes:
-      label: Tasks
-      description: "The tasks I am working on are:"
-      options:
-        - label: "An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...)"
-        - label: "My own task or dataset (give details below)"
-
-  - type: textarea
-    id: reproduction
-    validations:
-      required: true
-    attributes:
-      label: Reproduction
-      description: |
-        Please provide a code sample that reproduces the problem you ran into. It can be a Colab link or just a code snippet.
-        If you have code snippets, error messages, stack traces please provide them here as well.
-        Important! Use code tags to correctly format your code. See https://help.github.com/en/github/writing-on-github/creating-and-highlighting-code-blocks#syntax-highlighting
-        Do not use screenshots, as they are hard to read and (more importantly) don't allow others to copy-and-paste your code.
-
-      placeholder: |
-        Steps to reproduce the behavior:
-          
-          1.
-          2.
-          3.
-          
-
-  - type: textarea
-    id: expected-behavior
-    validations:
-      required: true
-    attributes:
-      label: Expected behavior
-      description: "A clear and concise description of what you would expect to happen."

.github/ISSUE_TEMPLATE/config.yml

@@ -1,12 +0,0 @@
-blank_issues_enabled: true
-version: 2.1
-contact_links:
-  - name: Model checkpoints on the Hugging Face Hub
-    url: https://huggingface.co/models
-    about: Open a Pull request / Discussion related to a specific model checkpoint directly on the Hugging Face Hub
-  - name: Website Related
-    url: https://github.com/huggingface/hub-docs/issues
-    about: Feature requests and bug reports related to the website
-  - name: Forum
-    url: https://discuss.huggingface.co/
-    about: General usage questions and community discussions

.github/ISSUE_TEMPLATE/feature-request.md

@@ -0,0 +1,25 @@
+---
+name: "\U0001F680 Feature request"
+about: Submit a proposal/request for a new transformers feature
+title: ''
+labels: ''
+assignees: ''
+
+---
+
+# 🚀 Feature request
+
+<!-- A clear and concise description of the feature proposal.
+     Please provide a link to the paper and code in case they exist. -->
+
+## Motivation
+
+<!-- Please outline the motivation for the proposal. Is your feature request
+     related to a problem? e.g., I'm always frustrated when [...]. If this is related
+     to another GitHub issue, please link here too. -->
+
+## Your contribution
+
+<!-- Is there any way that you could help, e.g. by submitting a PR?
+     Make sure to read the CONTRIBUTING.MD readme:
+     https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md -->

.github/ISSUE_TEMPLATE/feature-request.yml

@@ -1,31 +0,0 @@
-name: "\U0001F680 Feature request"
-description: Submit a proposal/request for a new transformers feature
-labels: [ "feature" ]
-body:
-  - type: textarea
-    id: feature-request
-    validations:
-      required: true
-    attributes:
-      label: Feature request
-      description: |
-        A clear and concise description of the feature proposal. Please provide a link to the paper and code in case they exist.
-
-  - type: textarea
-    id: motivation
-    validations:
-      required: true
-    attributes:
-      label: Motivation
-      description: |
-        Please outline the motivation for the proposal. Is your feature request related to a problem? e.g., I'm always frustrated when [...]. If this is related to another GitHub issue, please link here too.
-        
-
-  - type: textarea
-    id: contribution
-    validations:
-      required: true
-    attributes:
-      label: Your contribution
-      description: |
-        Is there any way that you could help, e.g. by submitting a PR? Make sure to read the CONTRIBUTING.MD [readme](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md)

.github/ISSUE_TEMPLATE/migration.md

@@ -0,0 +1,58 @@
+---
+name: "\U0001F4DA Migration from pytorch-pretrained-bert or pytorch-transformers"
+about: Report a problem when migrating from pytorch-pretrained-bert or pytorch-transformers
+  to transformers
+title: ''
+labels: Migration
+assignees: ''
+
+---
+
+# 📚 Migration
+
+## Information
+
+<!-- Important information -->
+
+Model I am using (Bert, XLNet ...):
+
+Language I am using the model on (English, Chinese ...):
+
+The problem arises when using:
+* [ ] the official example scripts: (give details below)
+* [ ] my own modified scripts: (give details below)
+
+The tasks I am working on is:
+* [ ] an official GLUE/SQUaD task: (give the name)
+* [ ] my own task or dataset: (give details below)
+
+## Details
+
+<!-- A clear and concise description of the migration issue.
+    If you have code snippets, please provide it here as well.
+    Important! Use code tags to correctly format your code. See https://help.github.com/en/github/writing-on-github/creating-and-highlighting-code-blocks#syntax-highlighting
+    Do not use screenshots, as they are hard to read and (more importantly) don't allow others to copy-and-paste your code.
+    -->
+
+## Environment info
+<!-- You can run the command `python transformers-cli env` and copy-and-paste its output below.
+     Don't forget to fill out the missing fields in that output! -->
+ 
+- `transformers` version:
+- Platform:
+- Python version:
+- PyTorch version (GPU?):
+- Tensorflow version (GPU?):
+- Using GPU in script?:
+- Using distributed or parallel set-up in script?:
+
+<!-- IMPORTANT: which version of the former library do you use? -->
+* `pytorch-transformers` or `pytorch-pretrained-bert` version (or branch):
+
+
+## Checklist
+
+- [ ] I have read the migration guide in the readme.
+ ([pytorch-transformers](https://github.com/huggingface/transformers#migrating-from-pytorch-transformers-to-transformers);
+  [pytorch-pretrained-bert](https://github.com/huggingface/transformers#migrating-from-pytorch-pretrained-bert-to-transformers))
+- [ ] I checked if a related official extension example runs on my machine.

.github/ISSUE_TEMPLATE/migration.yml

@@ -1,72 +0,0 @@
-name: "\U0001F4DA Migration from pytorch-pretrained-bert or pytorch-transformers"
-description: Report a problem when migrating from pytorch-pretrained-bert or pytorch-transformers to transformers
-labels: [ "migration" ]
-body:
-  - type: textarea
-    id: system-info
-    attributes:
-      label: System Info
-      description: Please share your system info with us. You can run the command `transformers-cli env` and copy-paste its output below.
-      render: shell
-      placeholder: transformers version, platform, python version, ...
-    validations:
-      required: true
-
-  - type: checkboxes
-    id: information-scripts-examples
-    attributes:
-      label: Information
-      description: 'The problem arises when using:'
-      options:
-        - label: "The official example scripts"
-        - label: "My own modified scripts"
-
-  - type: checkboxes
-    id: information-tasks
-    attributes:
-      label: Tasks
-      description: "The tasks I am working on are:"
-      options:
-        - label: "An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...)"
-        - label: "My own task or dataset (give details below)"
-
-  - type: textarea
-    id: reproduction
-    validations:
-      required: true
-    attributes:
-      label: Reproduction
-      description: |
-        Please provide a code sample that reproduces the problem you ran into. It can be a Colab link or just a code snippet.
-        If you have code snippets, error messages, stack traces please provide them here as well.
-        Important! Use code tags to correctly format your code. See https://help.github.com/en/github/writing-on-github/creating-and-highlighting-code-blocks#syntax-highlighting
-        Do not use screenshots, as they are hard to read and (more importantly) don't allow others to copy-and-paste your code.
-
-      placeholder: |
-        Steps to reproduce the behavior:
-          
-          1.
-          2.
-          3.
-          
-
-  - type: textarea
-    id: expected-behavior
-    validations:
-      required: true
-    attributes:
-      label: Expected behavior
-      description: "A clear and concise description of what you would expect to happen."
-      render: shell
-
-  - type: checkboxes
-    id: checklist
-    attributes:
-      label: Checklist
-      options:
-        - label: "I have read the migration guide in the readme.
- ([pytorch-transformers](https://github.com/huggingface/transformers#migrating-from-pytorch-transformers-to-transformers);
-  [pytorch-pretrained-bert](https://github.com/huggingface/transformers#migrating-from-pytorch-pretrained-bert-to-transformers))"
-          required: true
-        - label: "I checked if a related official extension example runs on my machine."
-          required: true

.github/ISSUE_TEMPLATE/new-model-addition.yml

@@ -1,31 +0,0 @@
-name: "\U0001F31F New model addition"
-description: Submit a proposal/request to implement a new model
-labels: [ "New model" ]
-
-body:
-  - type: textarea
-    id: description-request
-    validations:
-      required: true
-    attributes:
-      label: Model description
-      description: |
-        Put any and all important information relative to the model
-
-  - type: checkboxes
-    id: information-tasks
-    attributes:
-      label: Open source status
-      description: |
-          Please note that if the model implementation isn't available or if the weights aren't open-source, we are less likely to implement it in `transformers`.
-      options:
-        - label: "The model implementation is available"
-        - label: "The model weights are available"
-
-  - type: textarea
-    id: additional-info
-    attributes:
-      label: Provide useful links for the implementation
-      description: |
-        Please provide information regarding the implementation, the weights, and the authors.
-        Please mention the authors by @gh-username if you're aware of their usernames.

.github/ISSUE_TEMPLATE/question-help.md

@@ -0,0 +1,26 @@
+---
+name: "❓ Questions & Help"
+about: Post your general questions on the Hugging Face forum: https://discuss.huggingface.co/
+title: ''
+labels: ''
+assignees: ''
+
+---
+
+# ❓ Questions & Help
+
+<!-- The GitHub issue tracker is primarly intended for bugs, feature requests,
+     new models, benchmarks, and migration questions. For all other questions,
+     we direct you to the Hugging Face forum: https://discuss.huggingface.co/ .
+     -->
+
+## Details
+
+<!-- Description of your issue -->
+
+<!-- You should first ask your question on the forum, and only if
+     you didn't get an answer after a few days ask it here on GitHub. -->
+
+**A link to original question on the forum**:
+
+<!-- Your issue will be closed if you don't fill this part. -->

.github/PULL_REQUEST_TEMPLATE.md

@@ -17,58 +17,46 @@ Fixes # (issue)
 
 ## Before submitting
 - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
-- [ ] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests),
+- [ ] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md#start-contributing-pull-requests),
       Pull Request section?
 - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link
       to it if that's the case.
 - [ ] Did you make sure to update the documentation with your changes? Here are the
-      [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and
-      [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation).
+      [documentation guidelines](https://github.com/huggingface/transformers/tree/master/docs), and
+      [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/master/docs#writing-source-documentation).
 - [ ] Did you write any new necessary tests?
 
 
 ## Who can review?
 
 Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
-members/contributors who may be interested in your PR.
+members/contributors which may be interested in your PR.
 
 <!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
 
  If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
  Please tag fewer than 3 people.
 
-Models:
-
-- albert, bert, xlm: @LysandreJik
-- blenderbot, bart, marian, pegasus, encoderdecoder,  t5: @patrickvonplaten, @patil-suraj
-- longformer, reformer, transfoxl, xlnet: @patrickvonplaten
-- fsmt: @stas00
-- funnel: @sgugger
-- gpt2: @patrickvonplaten, @LysandreJik
-- rag: @patrickvonplaten, @lhoestq
-- tensorflow: @LysandreJik
-
-Library:
-
-- benchmarks: @patrickvonplaten
-- deepspeed: @stas00
-- ray/raytune: @richardliaw, @amogkam
-- text generation: @patrickvonplaten
-- tokenizers: @n1t0, @LysandreJik
-- trainer: @sgugger
-- pipelines: @LysandreJik
-
-Documentation: @sgugger
-
-HF projects:
-
-- datasets: [different repo](https://github.com/huggingface/datasets)
-- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
-
-Examples:
-
-- maintained examples (not research project or legacy): @sgugger, @patil-suraj
-- research_projects/bert-loses-patience: @JetRunner
-- research_projects/distillation: @VictorSanh
-
+ albert, bert, XLM: @LysandreJik
+ GPT2: @LysandreJik, @patrickvonplaten
+ tokenizers: @mfuntowicz
+ Trainer: @sgugger
+ Benchmarks: @patrickvonplaten
+ Model Cards: @julien-c
+ examples/distillation: @VictorSanh
+ nlp datasets: [different repo](https://github.com/huggingface/nlp)
+ rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
+ Text Generation: @patrickvonplaten, @TevenLeScao
+ Blenderbot, Bart, Marian, Pegasus: @patrickvonplaten
+ T5: @patrickvonplaten
+ Rag: @patrickvonplaten, @lhoestq
+ EncoderDecoder: @patrickvonplaten
+ Longformer, Reformer: @patrickvonplaten
+ TransfoXL, XLNet: @TevenLeScao, @patrickvonplaten
+ examples/seq2seq: @patil-suraj
+ examples/bert-loses-patience: @JetRunner
+ tensorflow: @jplu
+ examples/token-classification: @stefan-it
+ documentation: @sgugger
+ FSTM: @stas00
  -->

.github/conda/meta.yaml

@@ -14,10 +14,8 @@ requirements:
   host:
     - python
     - pip
-    - numpy >=1.17
+    - numpy
     - dataclasses
-    - importlib_metadata
-    - huggingface_hub
     - packaging
     - filelock
     - requests
@@ -25,14 +23,11 @@ requirements:
     - sacremoses
     - regex !=2019.12.17
     - protobuf
-    - tokenizers >=0.11.1,!=0.11.3,<0.13
-    - pyyaml >=5.1
+    - tokenizers ==0.9.4
   run:
     - python
-    - numpy >=1.17
+    - numpy
     - dataclasses
-    - importlib_metadata
-    - huggingface_hub
     - packaging
     - filelock
     - requests
@@ -40,8 +35,7 @@ requirements:
     - sacremoses
     - regex !=2019.12.17
     - protobuf
-    - tokenizers >=0.11.1,!=0.11.3,<0.13
-    - pyyaml >=5.1
+    - tokenizers ==0.9.4
 
 test:
   imports:

.github/stale.yml

@@ -0,0 +1,17 @@
+# Number of days of inactivity before an issue becomes stale
+daysUntilStale: 60
+# Number of days of inactivity before a stale issue is closed
+daysUntilClose: 7
+# Issues with these labels will never be considered stale
+exemptLabels:
+  - pinned
+  - security
+# Label to use when marking an issue as stale
+staleLabel: wontfix
+# Comment to post when marking an issue as stale. Set to `false` to disable
+markComment: >
+  This issue has been automatically marked as stale because it has not had
+  recent activity. It will be closed if no further activity occurs. Thank you
+  for your contributions.
+# Comment to post when closing a stale issue. Set to `false` to disable
+closeComment: false

.github/workflows/TROUBLESHOOT.md

@@ -1,9 +0,0 @@
-# Troubleshooting
-
-This is a document explaining how to deal with various issues on github-actions self-hosted CI. The entries may include actually solutions or pointers to Issues that cover those.
-
-## GitHub Actions (self-hosted CI)
-
-* Deepspeed
-
-  - if jit build hangs, clear out `rm -rf ~/.cache/torch_extensions/` reference: https://github.com/huggingface/transformers/pull/12723

.github/workflows/add-model-like.yml

@@ -1,78 +0,0 @@
-name: Add model like runner
-
-on:
-  push:
-    branches:
-      - main
-  pull_request:
-    paths:
-      - "src/**"
-      - "tests/**"
-      - ".github/**"
-    types: [opened, synchronize, reopened]
-
-jobs:
-  run_tests_templates_like:
-    name: "Add new model like template tests"
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v2
-
-      - name: Install dependencies
-        run: |
-          sudo apt -y update && sudo apt install -y libsndfile1-dev
-
-      - name: Load cached virtual environment
-        uses: actions/cache@v2
-        id: cache
-        with:
-          path: ~/venv/
-          key: v4-tests_model_like-${{ hashFiles('setup.py') }}
-
-      - name: Create virtual environment on cache miss
-        if: steps.cache.outputs.cache-hit != 'true'
-        run: |
-          python -m venv ~/venv && . ~/venv/bin/activate
-          pip install --upgrade pip!=21.3
-          pip install -e .[dev]
-
-      - name: Check transformers location
-        # make `transformers` available as package (required since we use `-e` flag) and check it's indeed from the repo.
-        run: |
-          . ~/venv/bin/activate
-          python setup.py develop
-          transformer_loc=$(pip show transformers | grep "Location: " | cut -c11-)
-          transformer_repo_loc=$(pwd .)
-          if [ "$transformer_loc" != "$transformer_repo_loc/src" ]; then
-              echo "transformers is from $transformer_loc but it shoud be from $transformer_repo_loc/src."
-              echo "A fix is required. Stop testing."
-              exit 1
-          fi
-
-      - name: Create model files
-        run: |
-          . ~/venv/bin/activate
-          transformers-cli add-new-model-like --config_file tests/fixtures/add_distilbert_like_config.json --path_to_repo .
-          make style
-          make fix-copies
-
-      - name: Run all PyTorch modeling test
-        run: |
-          . ~/venv/bin/activate
-          python -m pytest -n 2 --dist=loadfile -s --make-reports=tests_new_models tests/bert_new/test_modeling_bert_new.py
-
-      - name: Run style changes
-        run: |
-          . ~/venv/bin/activate
-          make style && make quality && make repo-consistency
-
-      - name: Failure short reports
-        if: ${{ always() }}
-        run: cat reports/tests_new_models/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: run_all_tests_new_models_test_reports
-          path: reports/tests_new_models

.github/workflows/build-docker-images.yml

@@ -1,196 +0,0 @@
-name: Build docker images (scheduled)
-
-on:
-  push:
-    branches:
-      - docker-image*
-  repository_dispatch:
-  workflow_call:
-  schedule:
-    - cron: "0 1 * * *"
-
-concurrency:
-  group: docker-images-builds
-  cancel-in-progress: false
-
-jobs:
-  latest-docker:
-    name: "Latest PyTorch + TensorFlow [dev]"
-    runs-on: ubuntu-latest
-    steps:
-      -
-        name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v1
-      -
-        name: Check out code
-        uses: actions/checkout@v2
-      -
-        name: Login to DockerHub
-        uses: docker/login-action@v1
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_PASSWORD }}
-      -
-        name: Build and push
-        uses: docker/build-push-action@v2
-        with:
-          context: ./docker/transformers-all-latest-gpu
-          build-args: |
-            REF=main
-          push: true
-          tags: huggingface/transformers-all-latest-gpu
-
-  latest-with-torch-nightly-docker:
-    name: "Nightly PyTorch + Stable TensorFlow"
-    runs-on: ubuntu-latest
-    steps:
-      -
-        name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v1
-      -
-        name: Check out code
-        uses: actions/checkout@v2
-      -
-        name: Login to DockerHub
-        uses: docker/login-action@v1
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_PASSWORD }}
-      -
-        name: Build and push
-        uses: docker/build-push-action@v2
-        with:
-          context: ./docker/transformers-all-latest-gpu
-          build-args: |
-            REF=main
-            PYTORCH=pre
-          push: true
-          tags: huggingface/transformers-all-latest-torch-nightly-gpu
-
-  latest-torch-deepspeed-docker:
-    name: "Latest PyTorch + DeepSpeed"
-    runs-on: ubuntu-latest
-    steps:
-      -
-        name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v1
-      -
-        name: Check out code
-        uses: actions/checkout@v2
-      -
-        name: Login to DockerHub
-        uses: docker/login-action@v1
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_PASSWORD }}
-      -
-        name: Build and push
-        uses: docker/build-push-action@v2
-        with:
-          context: ./docker/transformers-pytorch-deepspeed-latest-gpu
-          build-args: |
-            REF=main
-          push: true
-          tags: huggingface/transformers-pytorch-deepspeed-latest-gpu
-
-  nightly-torch-deepspeed-docker:
-    name: "Nightly PyTorch + DeepSpeed"
-    runs-on: ubuntu-latest
-    steps:
-      -
-        name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v1
-      -
-        name: Check out code
-        uses: actions/checkout@v2
-      -
-        name: Login to DockerHub
-        uses: docker/login-action@v1
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_PASSWORD }}
-      -
-        name: Build and push
-        uses: docker/build-push-action@v2
-        with:
-          context: ./docker/transformers-pytorch-deepspeed-nightly-gpu
-          build-args: |
-            REF=main
-          push: true
-          tags: huggingface/transformers-pytorch-deepspeed-nightly-gpu
-
-  doc-builder:
-    name: "Doc builder"
-    runs-on: ubuntu-latest
-    steps:
-      -
-        name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v1
-      -
-        name: Check out code
-        uses: actions/checkout@v2
-      -
-        name: Login to DockerHub
-        uses: docker/login-action@v1
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_PASSWORD }}
-      -
-        name: Build and push
-        uses: docker/build-push-action@v2
-        with:
-          context: ./docker/transformers-doc-builder
-          push: true
-          tags: huggingface/transformers-doc-builder
-
-  latest-pytorch:
-    name: "Latest PyTorch [dev]"
-    runs-on: ubuntu-latest
-    steps:
-      -
-        name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v1
-      -
-        name: Check out code
-        uses: actions/checkout@v2
-      -
-        name: Login to DockerHub
-        uses: docker/login-action@v1
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_PASSWORD }}
-      -
-        name: Build and push
-        uses: docker/build-push-action@v2
-        with:
-          context: ./docker/transformers-pytorch-gpu
-          build-args: |
-            REF=main
-          push: true
-          tags: huggingface/transformers-pytorch-gpu
-
-  latest-tensorflow:
-    name: "Latest TensorFlow [dev]"
-    runs-on: ubuntu-latest
-    steps:
-      -
-        name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v1
-      -
-        name: Check out code
-        uses: actions/checkout@v2
-      -
-        name: Login to DockerHub
-        uses: docker/login-action@v1
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_PASSWORD }}
-      -
-        name: Build and push
-        uses: docker/build-push-action@v2
-        with:
-          context: ./docker/transformers-tensorflow-gpu
-          build-args: |
-            REF=main
-          push: true
-          tags: huggingface/transformers-tensorflow-gpu

.github/workflows/build-past-ci-docker-images.yml

@@ -1,108 +0,0 @@
-name: Build docker images (Past CI)
-
-on:
-  push:
-    branches:
-      - past-ci-docker-image*
-
-concurrency:
-  group: docker-images-builds
-  cancel-in-progress: false
-
-jobs:
-  past-pytorch-docker:
-    name: "Past PyTorch Docker"
-    strategy:
-      fail-fast: false
-      matrix:
-        version: ["1.11", "1.10", "1.9", "1.8", "1.7", "1.6", "1.5", "1.4"]
-    runs-on: ubuntu-latest
-    steps:
-      -
-        name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v1
-      -
-        name: Check out code
-        uses: actions/checkout@v2
-      -
-        name: Login to DockerHub
-        uses: docker/login-action@v1
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_PASSWORD }}
-      -
-        name: Build and push
-        uses: docker/build-push-action@v2
-        with:
-          context: ./docker/transformers-past-gpu
-          build-args: |
-            REF=main
-            FRAMEWORK=pytorch
-            VERSION=${{ matrix.version }}
-          push: true
-          tags: huggingface/transformers-pytorch-past-${{ matrix.version }}-gpu
-
-  past-tensorflow-docker:
-    name: "Past TensorFlow Docker"
-    strategy:
-      fail-fast: false
-      matrix:
-        version: ["2.8", "2.7", "2.6", "2.5"]
-    runs-on: ubuntu-latest
-    steps:
-      -
-        name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v1
-      -
-        name: Check out code
-        uses: actions/checkout@v2
-      -
-        name: Login to DockerHub
-        uses: docker/login-action@v1
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_PASSWORD }}
-      -
-        name: Build and push
-        uses: docker/build-push-action@v2
-        with:
-          context: ./docker/transformers-past-gpu
-          build-args: |
-            REF=main
-            FRAMEWORK=tensorflow
-            VERSION=${{ matrix.version }}
-          push: true
-          tags: huggingface/transformers-tensorflow-past-${{ matrix.version }}-gpu
-
-  past-tensorflow-docker-2-4:
-    name: "Past TensorFlow Docker"
-    strategy:
-      fail-fast: false
-      matrix:
-        version: ["2.4"]
-    runs-on: ubuntu-latest
-    steps:
-      -
-        name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v1
-      -
-        name: Check out code
-        uses: actions/checkout@v2
-      -
-        name: Login to DockerHub
-        uses: docker/login-action@v1
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_PASSWORD }}
-      -
-        name: Build and push
-        uses: docker/build-push-action@v2
-        with:
-          context: ./docker/transformers-past-gpu
-          build-args: |
-            REF=main
-            BASE_DOCKER_IMAGE=nvidia/cuda:11.0.3-cudnn8-devel-ubuntu20.04
-            FRAMEWORK=tensorflow
-            VERSION=${{ matrix.version }}
-          push: true
-          tags: huggingface/transformers-tensorflow-past-${{ matrix.version }}-gpu

.github/workflows/build_documentation.yml

@@ -1,20 +0,0 @@
-name: Build documentation
-
-on:
-  push:
-    branches:
-      - main
-      - doc-builder*
-      - v*-release
-      - use_templates
-
-jobs:
-   build:
-    uses: huggingface/doc-builder/.github/workflows/build_main_documentation.yml@main
-    with:
-      commit_sha: ${{ github.sha }}
-      package: transformers
-      notebook_folder: transformers_doc
-      languages: en es it pt
-    secrets:
-      token: ${{ secrets.HUGGINGFACE_PUSH }}

.github/workflows/build_pr_documentation.yml

@@ -1,17 +0,0 @@
-name: Build PR Documentation
-
-on:
-  pull_request:
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
-  cancel-in-progress: true
-
-jobs:
-  build:
-    uses: huggingface/doc-builder/.github/workflows/build_pr_documentation.yml@main
-    with:
-      commit_sha: ${{ github.event.pull_request.head.sha }}
-      pr_number: ${{ github.event.number }}
-      package: transformers
-      languages: en es it pt

.github/workflows/delete_doc_comment.yml

@@ -1,13 +0,0 @@
-name: Delete dev documentation
-
-on:
-  pull_request:
-    types: [ closed ]
-
-
-jobs:
-  delete:
-    uses: huggingface/doc-builder/.github/workflows/delete_doc_comment.yml@main
-    with:
-      pr_number: ${{ github.event.number }}
-      package: transformers

.github/workflows/doctests.yml

@@ -1,78 +0,0 @@
-name: Doctests
-
-on:
-  push:
-    branches:
-      - doctest*
-  repository_dispatch:
-  schedule:
-    - cron: "0 0 * * *"
-
-
-env:
-  HF_HOME: /mnt/cache
-  TRANSFORMERS_IS_CI: yes
-  RUN_SLOW: yes
-  OMP_NUM_THREADS: 16
-  MKL_NUM_THREADS: 16
-  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
-  TF_FORCE_GPU_ALLOW_GROWTH: true
-
-jobs:
-  run_doctests:
-    runs-on: [self-hosted, doc-tests-gpu]
-    container:
-      image: huggingface/transformers-all-latest-gpu
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    steps:
-      - uses: actions/checkout@v2
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: GPU visibility
-        run: |
-          python3 utils/print_env.py
-
-      - name: Prepare files for doctests
-        run: |
-          python3 utils/prepare_for_doc_test.py src docs
-
-      - name: Run doctests
-        run: |
-          python3 -m pytest -v --make-reports doc_tests_gpu --doctest-modules $(cat utils/documentation_tests.txt) -sv --doctest-continue-on-failure --doctest-glob="*.mdx"
-
-      - name: Clean files after doctests
-        run: |
-          python3 utils/prepare_for_doc_test.py src docs --remove_new_line
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat reports/doc_tests_gpu/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: doc_tests_gpu_test_reports
-          path: reports/doc_tests_gpu
-
-
-  send_results:
-    name: Send results to webhook
-    runs-on: ubuntu-latest
-    if: always()
-    needs: [run_doctests]
-    steps:
-      - uses: actions/checkout@v2
-      - uses: actions/download-artifact@v2
-      - name: Send message to Slack
-        env:
-          CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
-          CI_SLACK_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY_DOCS }}
-          CI_SLACK_CHANNEL_ID_DAILY: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY_DOCS }}
-          CI_SLACK_CHANNEL_DUMMY_TESTS: ${{ secrets.CI_SLACK_CHANNEL_DUMMY_TESTS }}
-        run: |
-          pip install slack_sdk
-          python utils/notification_service_doc_tests.py

.github/workflows/github-torch-hub.yml

@@ -0,0 +1,44 @@
+name: Torch hub integration
+
+on: 
+  push:
+    branches:
+      - "*"
+
+jobs:
+  torch_hub_integration:
+    runs-on: ubuntu-latest
+    env:
+      # TODO quickfix but may need more investigation
+      ACTIONS_ALLOW_UNSECURE_COMMANDS: True
+    steps:
+    # no checkout necessary here.
+    - name: Extract branch name
+      run: echo "::set-env name=BRANCH::${GITHUB_REF#refs/heads/}"
+    - name: Check branch name
+      run: echo $BRANCH
+    - name: Set up Python
+      uses: actions/setup-python@v1
+      with:
+        python-version: 3.7
+
+    - name: Loading cache
+      uses: actions/cache@v2
+      id: cache
+      with:
+        path: ~/.cache/pip
+        key: v0-torch_hub-${{ hashFiles('setup.py') }}
+
+    - name: Install dependencies
+      run: |
+        pip install --upgrade pip
+        pip install torch
+        pip install numpy filelock protobuf requests tqdm regex sentencepiece sacremoses tokenizers packaging
+
+    - name: Torch hub list
+      run: |
+        python -c "import torch; print(torch.hub.list('huggingface/transformers:$BRANCH'))"
+
+    - name: Torch hub help
+      run: |
+        python -c "import torch; print(torch.hub.help('huggingface/transformers:$BRANCH', 'modelForSequenceClassification'))"

.github/workflows/model-templates.yml

@@ -1,81 +0,0 @@
-name: Model templates runner
-
-on:
-  repository_dispatch:
-  schedule:
-    - cron: "0 2 * * *"
-
-jobs:
-  run_tests_templates:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@v2
-
-      - name: Install dependencies
-        run: |
-          sudo apt -y update && sudo apt install -y libsndfile1-dev
-
-      - name: Load cached virtual environment
-        uses: actions/cache@v2
-        id: cache
-        with:
-          path: ~/venv/
-          key: v4-tests_templates-${{ hashFiles('setup.py') }}
-
-      - name: Create virtual environment on cache miss
-        if: steps.cache.outputs.cache-hit != 'true'
-        run: |
-          python -m venv ~/venv && . ~/venv/bin/activate
-          pip install --upgrade pip!=21.3
-          pip install -e .[dev]
-
-      - name: Check transformers location
-        # make `transformers` available as package (required since we use `-e` flag) and check it's indeed from the repo.
-        run: |
-          . ~/venv/bin/activate
-          python setup.py develop
-          transformer_loc=$(pip show transformers | grep "Location: " | cut -c11-)
-          transformer_repo_loc=$(pwd .)
-          if [ "$transformer_loc" != "$transformer_repo_loc/src" ]; then
-              echo "transformers is from $transformer_loc but it shoud be from $transformer_repo_loc/src."
-              echo "A fix is required. Stop testing."
-              exit 1
-          fi
-
-      - name: Create model files
-        run: |
-          . ~/venv/bin/activate
-          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/encoder-bert-tokenizer.json --path=templates/adding_a_new_model
-          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/pt-encoder-bert-tokenizer.json --path=templates/adding_a_new_model
-          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/standalone.json --path=templates/adding_a_new_model
-          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/tf-encoder-bert-tokenizer.json --path=templates/adding_a_new_model
-          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/tf-seq-2-seq-bart-tokenizer.json --path=templates/adding_a_new_model
-          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/pt-seq-2-seq-bart-tokenizer.json --path=templates/adding_a_new_model
-          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/flax-encoder-bert-tokenizer.json --path=templates/adding_a_new_model
-          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/flax-seq-2-seq-bart-tokenizer.json --path=templates/adding_a_new_model
-          make style
-          python utils/check_table.py --fix_and_overwrite
-          python utils/check_dummies.py --fix_and_overwrite
-          python utils/check_copies.py --fix_and_overwrite
-
-      - name: Run all non-slow tests
-        run: |
-          . ~/venv/bin/activate
-          python -m pytest -n 2 --dist=loadfile -s --make-reports=tests_templates tests/*template*
-
-      - name: Run style changes
-        run: |
-          . ~/venv/bin/activate
-          make style && make quality && make repo-consistency
-
-      - name: Failure short reports
-        if: ${{ always() }}
-        run: cat reports/tests_templates/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: run_all_tests_templates_test_reports
-          path: reports/tests_templates

.github/workflows/release-conda.yml

@@ -4,8 +4,6 @@ on:
   push:
     tags:
       - v*
-    branches:
-      - conda_*
 
 env:
   ANACONDA_API_TOKEN: ${{ secrets.ANACONDA_API_TOKEN }}
@@ -26,7 +24,6 @@ jobs:
         with:
           auto-update-conda: true
           auto-activate-base: false
-          python-version: 3.8
           activate-environment: "build-transformers"
           channels: huggingface
 
@@ -40,8 +37,7 @@ jobs:
       - name: Build conda packages
         run: |
           conda info
-          conda list
-          conda-build .github/conda
+          conda build .github/conda
 
       - name: Upload to Anaconda
-        run: anaconda upload `conda-build .github/conda --output` --force
+        run: anaconda upload `conda build .github/conda --output` --force

.github/workflows/self-nightly-scheduled.yml

@@ -1,236 +0,0 @@
-name: Self-hosted runner (nightly)
-
-# Note that each job's dependencies go into a corresponding docker file.
-#
-# For example for `run_all_tests_torch_cuda_extensions_gpu` the docker image is
-# `huggingface/transformers-pytorch-deepspeed-latest-gpu`, which can be found at
-# `docker/transformers-pytorch-deepspeed-latest-gpu/Dockerfile`
-
-on:
-  repository_dispatch:
-  schedule:
-    - cron: "0 16 * * *"
-
-env:
-  HF_HOME: /mnt/cache
-  TRANSFORMERS_IS_CI: yes
-  OMP_NUM_THREADS: 8
-  MKL_NUM_THREADS: 8
-  RUN_SLOW: yes
-  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
-  TF_FORCE_GPU_ALLOW_GROWTH: true
-  RUN_PT_TF_CROSS_TESTS: 1
-
-jobs:
-  setup:
-    name: Setup
-    strategy:
-      matrix:
-        machine_type: [single-gpu, multi-gpu]
-    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
-    container:
-      image: huggingface/transformers-all-latest-torch-nightly-gpu
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    outputs:
-      matrix: ${{ steps.set-matrix.outputs.matrix }}
-    steps:
-      - name: Update clone
-        working-directory: /transformers
-        run: |
-          git fetch && git checkout ${{ github.sha }}
-
-      - name: Cleanup
-        working-directory: /transformers
-        run: |
-          rm -rf tests/__pycache__
-          rm -rf tests/models/__pycache__
-          rm -rf reports
-
-      - id: set-matrix
-        name: Identify models to test
-        working-directory: /transformers/tests
-        run: |
-          echo "::set-output name=matrix::$(python3 -c 'import os; tests = os.getcwd(); model_tests = os.listdir(os.path.join(tests, "models")); d1 = sorted(list(filter(os.path.isdir, os.listdir(tests)))); d2 = sorted(list(filter(os.path.isdir, [f"models/{x}" for x in model_tests]))); d1.remove("models"); d = d2 + d1; print(d)')"
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-  run_tests_single_gpu:
-    name: Model tests
-    strategy:
-      fail-fast: false
-      matrix:
-        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
-        machine_type: [single-gpu]
-    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
-    container:
-      image: huggingface/transformers-all-latest-torch-nightly-gpu
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    needs: setup
-    steps:
-      - name: Echo folder ${{ matrix.folders }}
-        shell: bash
-        # For folders like `models/bert`, set an env. var. (`matrix_folders`) to `models_bert`, which will be used to
-        # set the artifact folder names (because the character `/` is not allowed).
-        run: |
-          echo "${{ matrix.folders }}"
-          matrix_folders=${{ matrix.folders }}
-          matrix_folders=${matrix_folders/'models/'/'models_'}
-          echo "$matrix_folders"
-          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
-
-      - name: Update clone
-        working-directory: /transformers
-        run: git fetch && git checkout ${{ github.sha }}
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /transformers
-        run: |
-          python3 utils/print_env.py
-
-      - name: Run all tests on GPU
-        working-directory: /transformers
-        run: python3 -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
-          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
-
-  run_tests_multi_gpu:
-    name: Model tests
-    strategy:
-      fail-fast: false
-      matrix:
-        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
-        machine_type: [multi-gpu]
-    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
-    container:
-      image: huggingface/transformers-all-latest-torch-nightly-gpu
-      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    needs: setup
-    steps:
-      - name: Echo folder ${{ matrix.folders }}
-        shell: bash
-        # For folders like `models/bert`, set an env. var. (`matrix_folders`) to `models_bert`, which will be used to
-        # set the artifact folder names (because the character `/` is not allowed).
-        run: |
-          echo "${{ matrix.folders }}"
-          matrix_folders=${{ matrix.folders }}
-          matrix_folders=${matrix_folders/'models/'/'models_'}
-          echo "$matrix_folders"
-          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
-
-      - name: Update clone
-        working-directory: /transformers
-        run: git fetch && git checkout ${{ github.sha }}
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /transformers
-        run: |
-          python3 utils/print_env.py
-
-      - name: Run all tests on GPU
-        working-directory: /transformers
-        run: python3 -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
-          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
-
-  run_all_tests_torch_cuda_extensions_gpu:
-    name: Torch CUDA extension tests
-    strategy:
-      fail-fast: false
-      matrix:
-        machine_type: [single-gpu, multi-gpu]
-    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
-    needs: setup
-    container:
-      image: huggingface/transformers-pytorch-deepspeed-nightly-gpu
-      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    steps:
-      - name: Update clone
-        working-directory: /workspace/transformers
-        run: git fetch && git checkout ${{ github.sha }}
-
-      # To avoid unknown test failures
-      - name: Pre build DeepSpeed *again*
-        working-directory: /workspace
-        run: |
-          python3 -m pip uninstall -y deepspeed
-          rm -rf DeepSpeed
-          git clone https://github.com/microsoft/DeepSpeed && cd DeepSpeed && rm -rf build
-          DS_BUILD_CPU_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install . --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /workspace/transformers
-        run: |
-          python utils/print_env.py
-
-      - name: Run all tests on GPU
-        working-directory: /workspace/transformers
-        run: |
-          python -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu tests/deepspeed tests/extended
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports
-          path: /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu
-
-  send_results:
-    name: Send results to webhook
-    runs-on: ubuntu-latest
-    if: always()
-    needs: [setup, run_tests_single_gpu, run_tests_multi_gpu, run_all_tests_torch_cuda_extensions_gpu]
-    steps:
-      - uses: actions/checkout@v2
-      - uses: actions/download-artifact@v2
-      - name: Send message to Slack
-        env:
-          CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
-          CI_SLACK_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID }}
-          CI_SLACK_CHANNEL_ID_DAILY: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY }}
-          CI_SLACK_CHANNEL_DUMMY_TESTS: ${{ secrets.CI_SLACK_CHANNEL_DUMMY_TESTS }}
-          CI_SLACK_REPORT_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID_PAST_FUTURE }}
-          CI_EVENT: nightly-build
-        # We pass `needs.setup.outputs.matrix` as the argument. A processing in `notification_service.py` to change
-        # `models/bert` to `models_bert` is required, as the artifact names use `_` instead of `/`.
-        run: |
-          pip install slack_sdk
-          python utils/notification_service.py "${{ needs.setup.outputs.matrix }}"

.github/workflows/self-past-caller.yml

@@ -1,136 +0,0 @@
-name: Self-hosted runner (past-ci-caller)
-
-on:
-  push:
-    branches:
-      - run-past-ci*
-
-jobs:
-  run_past_ci_pytorch_1-11:
-    name: PyTorch 1.11
-    if: always()
-    uses: ./.github/workflows/self-past.yml
-    with:
-      framework: pytorch
-      version: "1.11"
-    secrets: inherit
-
-  run_past_ci_pytorch_1-10:
-    name: PyTorch 1.10
-    if: always()
-    needs: [run_past_ci_pytorch_1-11]
-    uses: ./.github/workflows/self-past.yml
-    with:
-      framework: pytorch
-      version: "1.10"
-    secrets: inherit
-
-  run_past_ci_pytorch_1-9:
-    name: PyTorch 1.9
-    if: always()
-    needs: [run_past_ci_pytorch_1-10]
-    uses: ./.github/workflows/self-past.yml
-    with:
-      framework: pytorch
-      version: "1.9"
-    secrets: inherit
-
-  run_past_ci_pytorch_1-8:
-    name: PyTorch 1.8
-    if: always()
-    needs: [run_past_ci_pytorch_1-9]
-    uses: ./.github/workflows/self-past.yml
-    with:
-      framework: pytorch
-      version: "1.8"
-    secrets: inherit
-
-  run_past_ci_pytorch_1-7:
-    name: PyTorch 1.7
-    if: always()
-    needs: [run_past_ci_pytorch_1-8]
-    uses: ./.github/workflows/self-past.yml
-    with:
-      framework: pytorch
-      version: "1.7"
-    secrets: inherit
-
-  run_past_ci_pytorch_1-6:
-    name: PyTorch 1.6
-    if: always()
-    needs: [run_past_ci_pytorch_1-7]
-    uses: ./.github/workflows/self-past.yml
-    with:
-      framework: pytorch
-      version: "1.6"
-    secrets: inherit
-
-  run_past_ci_pytorch_1-5:
-    name: PyTorch 1.5
-    if: always()
-    needs: [run_past_ci_pytorch_1-6]
-    uses: ./.github/workflows/self-past.yml
-    with:
-      framework: pytorch
-      version: "1.5"
-    secrets: inherit
-
-  run_past_ci_pytorch_1-4:
-    name: PyTorch 1.4
-    if: always()
-    needs: [run_past_ci_pytorch_1-5]
-    uses: ./.github/workflows/self-past.yml
-    with:
-      framework: pytorch
-      version: "1.4"
-    secrets: inherit
-
-  run_past_ci_tensorflow_2-8:
-    name: TensorFlow 2.8
-    if: always()
-    needs: [run_past_ci_pytorch_1-4]
-    uses: ./.github/workflows/self-past.yml
-    with:
-      framework: tensorflow
-      version: "2.8"
-    secrets: inherit
-
-  run_past_ci_tensorflow_2-7:
-    name: TensorFlow 2.7
-    if: always()
-    needs: [run_past_ci_tensorflow_2-8]
-    uses: ./.github/workflows/self-past.yml
-    with:
-      framework: tensorflow
-      version: "2.7"
-    secrets: inherit
-
-  run_past_ci_tensorflow_2-6:
-    name: TensorFlow 2.6
-    if: always()
-    needs: [run_past_ci_tensorflow_2-7]
-    uses: ./.github/workflows/self-past.yml
-    with:
-      framework: tensorflow
-      version: "2.6"
-    secrets: inherit
-
-  run_past_ci_tensorflow_2-5:
-    name: TensorFlow 2.5
-    if: always()
-    needs: [run_past_ci_tensorflow_2-6]
-    uses: ./.github/workflows/self-past.yml
-    with:
-      framework: tensorflow
-      version: "2.5"
-    secrets: inherit
-
-  run_past_ci_tensorflow_2-4:
-    name: TensorFlow 2.4
-    if: always()
-    needs: [run_past_ci_tensorflow_2-5]
-    uses: ./.github/workflows/self-past.yml
-    with:
-      framework: tensorflow
-      version: "2.4"
-    secrets: inherit

.github/workflows/self-past.yml

@@ -1,192 +0,0 @@
-name: Self-hosted runner (past)
-
-# Note that each job's dependencies go into a corresponding docker file.
-#
-# For example for `run_all_tests_torch_cuda_extensions_gpu` the docker image is
-# `huggingface/transformers-pytorch-deepspeed-latest-gpu`, which can be found at
-# `docker/transformers-pytorch-deepspeed-latest-gpu/Dockerfile`
-
-on:
-  workflow_call:
-    inputs:
-      framework:
-        required: true
-        type: string
-      version:
-        required: true
-        type: string
-
-env:
-  HF_HOME: /mnt/cache
-  TRANSFORMERS_IS_CI: yes
-  OMP_NUM_THREADS: 8
-  MKL_NUM_THREADS: 8
-  RUN_SLOW: yes
-  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
-  TF_FORCE_GPU_ALLOW_GROWTH: true
-  RUN_PT_TF_CROSS_TESTS: 1
-
-jobs:
-  setup:
-    name: Setup
-    runs-on: ubuntu-latest
-    outputs:
-      matrix: ${{ steps.set-matrix.outputs.matrix }}
-    steps:
-      - name: Checkout transformers
-        uses: actions/checkout@v2
-        with:
-          fetch-depth: 2
-
-      - name: Cleanup
-        run: |
-          rm -rf tests/__pycache__
-          rm -rf tests/models/__pycache__
-          rm -rf reports
-
-      - id: set-matrix
-        name: Identify models to test
-        run: |
-          cd tests
-          echo "::set-output name=matrix::$(python3 -c 'import os; tests = os.getcwd(); model_tests = os.listdir(os.path.join(tests, "models")); d1 = sorted(list(filter(os.path.isdir, os.listdir(tests)))); d2 = sorted(list(filter(os.path.isdir, [f"models/{x}" for x in model_tests]))); d1.remove("models"); d = d2 + d1; print(d)')"
-
-  run_tests_single_gpu:
-    name: Model tests
-    strategy:
-      fail-fast: false
-      matrix:
-        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
-        machine_type: [single-gpu]
-    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker-past-ci') }}
-    container:
-      image: huggingface/transformers-${{ inputs.framework }}-past-${{ inputs.version }}-gpu
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    needs: setup
-    steps:
-      - name: Update clone
-        working-directory: /transformers
-        run: git fetch && git checkout ${{ github.sha }}
-
-      - name: Echo folder ${{ matrix.folders }}
-        shell: bash
-        # For folders like `models/bert`, set an env. var. (`matrix_folders`) to `models_bert`, which will be used to
-        # set the artifact folder names (because the character `/` is not allowed).
-        run: |
-          echo "${{ matrix.folders }}"
-          matrix_folders=${{ matrix.folders }}
-          matrix_folders=${matrix_folders/'models/'/'models_'}
-          echo "$matrix_folders"
-          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /transformers
-        run: |
-          python3 utils/print_env.py
-
-      - name: Run all tests on GPU
-        working-directory: /transformers
-        run: python3 -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
-          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
-
-  run_tests_multi_gpu:
-    name: Model tests
-    strategy:
-      fail-fast: false
-      matrix:
-        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
-        machine_type: [multi-gpu]
-    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker-past-ci') }}
-    container:
-      image: huggingface/transformers-${{ inputs.framework }}-past-${{ inputs.version }}-gpu
-      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    needs: setup
-    steps:
-      - name: Update clone
-        working-directory: /transformers
-        run: git fetch && git checkout ${{ github.sha }}
-
-      - name: Echo folder ${{ matrix.folders }}
-        shell: bash
-        # For folders like `models/bert`, set an env. var. (`matrix_folders`) to `models_bert`, which will be used to
-        # set the artifact folder names (because the character `/` is not allowed).
-        run: |
-          echo "${{ matrix.folders }}"
-          matrix_folders=${{ matrix.folders }}
-          matrix_folders=${matrix_folders/'models/'/'models_'}
-          echo "$matrix_folders"
-          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /transformers
-        run: |
-          python3 utils/print_env.py
-
-      - name: Run all tests on GPU
-        working-directory: /transformers
-        run: python3 -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
-          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
-
-  send_results:
-    name: Send results to webhook
-    runs-on: ubuntu-latest
-    if: always()
-    needs: [setup, run_tests_single_gpu, run_tests_multi_gpu]
-    steps:
-      - uses: actions/checkout@v2
-      - uses: actions/download-artifact@v2
-
-      # Create a directory to store test failure tables in the next step
-      - name: Create directory
-        run: mkdir test_failure_tables
-
-      - name: Send message to Slack
-        env:
-          CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
-          CI_SLACK_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID }}
-          CI_SLACK_CHANNEL_ID_DAILY: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY }}
-          CI_SLACK_CHANNEL_DUMMY_TESTS: ${{ secrets.CI_SLACK_CHANNEL_DUMMY_TESTS }}
-          CI_SLACK_REPORT_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID_PAST_FUTURE }}
-          CI_EVENT: Past CI - ${{ inputs.framework }}-${{ inputs.version }}
-        # We pass `needs.setup.outputs.matrix` as the argument. A processing in `notification_service.py` to change
-        # `models/bert` to `models_bert` is required, as the artifact names use `_` instead of `/`.
-        run: |
-          pip install slack_sdk
-          python utils/notification_service.py "${{ needs.setup.outputs.matrix }}"
-
-      # Upload complete failure tables, as they might be big and only truncated versions could be sent to Slack.
-      - name: Failure table artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: test_failure_tables_${{ inputs.framework }}-${{ inputs.version }}
-          path: test_failure_tables

.github/workflows/self-push-caller.yml

@@ -1,52 +0,0 @@
-# Used to trigger self-push CI
-name: Self-hosted runner (push-caller)
-
-on:
-  push:
-    branches:
-      - main
-    paths:
-      - "src/**"
-      - "tests/**"
-      - ".github/**"
-      - "templates/**"
-      - "utils/**"
-
-jobs:
-  check-for-setup:
-      runs-on: ubuntu-latest
-      name: Check if setup was changed
-      outputs:
-        changed: ${{ steps.was_changed.outputs.changed }}
-      steps:
-        - uses: actions/checkout@v3
-          with: 
-            fetch-depth: "2"
-        
-        - name: Get changed files
-          id: changed-files
-          uses: tj-actions/changed-files@v22.2
-        
-        - name: Was setup changed 
-          id: was_changed
-          run: |
-            for file in ${{ steps.changed-files.outputs.all_changed_files }}; do
-              if [ `basename "${file}"` = "setup.py" ]; then
-                echo ::set-output name=changed::"1"
-              fi
-            done
-
-  build-docker-containers:
-    needs: check-for-setup
-    if: (github.event_name == 'push') && (needs.check-for-setup.outputs.changed == '1')
-    uses: ./.github/workflows/build-docker-images.yml
-    secrets: inherit
-
-  run_push_ci:
-    name: Trigger Push CI
-    runs-on: ubuntu-latest
-    if: ${{ always() }}
-    needs: build-docker-containers
-    steps:
-      - name: Trigger push CI via workflow_run
-        run: echo "Trigger push CI via workflow_run"

.github/workflows/self-push.yml

@@ -1,505 +1,273 @@
 name: Self-hosted runner (push)
 
 on:
-  workflow_run:
-    workflows: ["Self-hosted runner (push-caller)"]
-    branches: ["main"]
-    types: [completed]
   push:
     branches:
-      - ci_*
-      - ci-*
+      - master
+      - model-templates
     paths:
       - "src/**"
       - "tests/**"
       - ".github/**"
       - "templates/**"
-      - "utils/**"
+  # pull_request:
   repository_dispatch:
 
-env:
-  HF_HOME: /mnt/cache
-  TRANSFORMERS_IS_CI: yes
-  OMP_NUM_THREADS: 8
-  MKL_NUM_THREADS: 8
-  PYTEST_TIMEOUT: 60
-  TF_FORCE_GPU_ALLOW_GROWTH: true
-  RUN_PT_TF_CROSS_TESTS: 1
 
 jobs:
-  setup:
-    name: Setup
-    runs-on: ubuntu-latest
-    outputs:
-      matrix: ${{ steps.set-matrix.outputs.matrix }}
-      test_map: ${{ steps.set-matrix.outputs.test_map }}
+  run_tests_torch_gpu:
+    runs-on: [self-hosted, gpu, single-gpu]
     steps:
-      # Necessary to get the correct branch name and commit SHA for `workflow_run` event
-      # We also take into account the `push` event (we might want to test some changes in a branch)
-      - name: Prepare custom environment variables
-        shell: bash
-        # `CI_BRANCH_PUSH`: The branch name from the push event
-        # `CI_BRANCH_WORKFLOW_RUN`: The name of the branch on which this workflow is triggered by `workflow_run` event
-        # `CI_BRANCH`: The non-empty branch name from the above two (one and only one of them is empty)
-        # `CI_SHA_PUSH`: The commit SHA from the push event
-        # `CI_SHA_WORKFLOW_RUN`: The commit SHA that triggers this workflow by `workflow_run` event
-        # `CI_SHA`: The non-empty commit SHA from the above two (one and only one of them is empty)
-        run: |
-          CI_BRANCH_PUSH=${{ github.event.ref }}
-          CI_BRANCH_PUSH=${CI_BRANCH_PUSH/'refs/heads/'/''}
-          CI_BRANCH_WORKFLOW_RUN=${{ github.event.workflow_run.head_branch }}
-          CI_SHA_PUSH=${{ github.event.head_commit.id }}
-          CI_SHA_WORKFLOW_RUN=${{ github.event.workflow_run.head_sha }}
-          echo $CI_BRANCH_PUSH
-          echo $CI_BRANCH_WORKFLOW_RUN
-          echo $CI_SHA_PUSH
-          echo $CI_SHA_WORKFLOW_RUN
-          [[ ! -z "$CI_BRANCH_PUSH" ]] && echo "CI_BRANCH=$CI_BRANCH_PUSH" >> $GITHUB_ENV || echo "CI_BRANCH=$CI_BRANCH_WORKFLOW_RUN" >> $GITHUB_ENV
-          [[ ! -z "$CI_SHA_PUSH" ]] && echo "CI_SHA=$CI_SHA_PUSH" >> $GITHUB_ENV || echo "CI_SHA=$CI_SHA_WORKFLOW_RUN" >> $GITHUB_ENV
-
-      - name: print environment variables
-        run: |
-          echo "env.CI_BRANCH = ${{ env.CI_BRANCH }}"
-          echo "env.CI_SHA = ${{ env.CI_SHA }}"
-
-      - name: Checkout transformers
-        uses: actions/checkout@v2
-        with:
-          fetch-depth: 2
-
-      - name: Update clone using environment variables
-        run: |
-          echo "original branch = $(git branch --show-current)"
-          git fetch && git checkout ${{ env.CI_BRANCH }}
-          echo "updated branch = $(git branch --show-current)"
-          git checkout ${{ env.CI_SHA }}
-          echo "log = $(git log -n 1)"
-
-      - name: Cleanup
-        run: |
-          rm -rf tests/__pycache__
-          rm -rf tests/models/__pycache__
-          rm -rf reports
-
-      - name: Fetch the tests to run
-        # TODO: add `git-python` in the docker images
-        run: |
-          pip install --upgrade git-python
-          python utils/tests_fetcher.py --diff_with_last_commit | tee test_preparation.txt
-
-      - name: Report fetched tests
-        uses: actions/upload-artifact@v2
-        with:
-          name: test_fetched
-          path: test_preparation.txt
-
-      - id: set-matrix
-        name: Organize tests into models
-        # The `keys` is used as GitHub actions matrix for jobs, i.e. `models/bert`, `tokenization`, `pipeline`, etc.
-        # The `test_map` is used to get the actual identified test files under each key.
-        # If no test to run (so no `test_map.json` file), create a dummy map (empty matrix will fail)
-        run: |
-          if [ -f test_map.json ]; then
-              keys=$(python3 -c 'import json; fp = open("test_map.json"); test_map = json.load(fp); fp.close(); d = list(test_map.keys()); print(d)')
-              test_map=$(python3 -c 'import json; fp = open("test_map.json"); test_map = json.load(fp); fp.close(); print(test_map)')
-          else
-              keys=$(python3 -c 'keys = ["dummy"]; print(keys)')
-              test_map=$(python3 -c 'test_map = {"dummy": []}; print(test_map)')
-          fi
-          echo $keys
-          echo $test_map
-          echo "::set-output name=matrix::$keys"
-          echo "::set-output name=test_map::$test_map"
-
-  run_tests_single_gpu:
-    name: Model tests
-    needs: setup
-    # `dummy` means there is no test to run
-    if: contains(fromJson(needs.setup.outputs.matrix), 'dummy') != true
-    strategy:
-      fail-fast: false
-      matrix:
-        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
-        machine_type: [single-gpu]
-    runs-on: [self-hosted, docker-gpu, '${{ matrix.machine_type }}']
-    container:
-      image: huggingface/transformers-all-latest-gpu
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    steps:
-      # Necessary to get the correct branch name and commit SHA for `workflow_run` event
-      # We also take into account the `push` event (we might want to test some changes in a branch)
-      - name: Prepare custom environment variables
-        shell: bash
-        # For the meaning of these environment variables, see the job `Setup`
-        run: |
-          CI_BRANCH_PUSH=${{ github.event.ref }}
-          CI_BRANCH_PUSH=${CI_BRANCH_PUSH/'refs/heads/'/''}
-          CI_BRANCH_WORKFLOW_RUN=${{ github.event.workflow_run.head_branch }}
-          CI_SHA_PUSH=${{ github.event.head_commit.id }}
-          CI_SHA_WORKFLOW_RUN=${{ github.event.workflow_run.head_sha }}
-          echo $CI_BRANCH_PUSH
-          echo $CI_BRANCH_WORKFLOW_RUN
-          echo $CI_SHA_PUSH
-          echo $CI_SHA_WORKFLOW_RUN
-          [[ ! -z "$CI_BRANCH_PUSH" ]] && echo "CI_BRANCH=$CI_BRANCH_PUSH" >> $GITHUB_ENV || echo "CI_BRANCH=$CI_BRANCH_WORKFLOW_RUN" >> $GITHUB_ENV
-          [[ ! -z "$CI_SHA_PUSH" ]] && echo "CI_SHA=$CI_SHA_PUSH" >> $GITHUB_ENV || echo "CI_SHA=$CI_SHA_WORKFLOW_RUN" >> $GITHUB_ENV
-
-      - name: print environment variables
-        run: |
-          echo "env.CI_BRANCH = ${{ env.CI_BRANCH }}"
-          echo "env.CI_SHA = ${{ env.CI_SHA }}"
-
-      - name: Update clone using environment variables
-        working-directory: /transformers
-        run: |
-          echo "original branch = $(git branch --show-current)"
-          git fetch && git checkout ${{ env.CI_BRANCH }}
-          echo "updated branch = $(git branch --show-current)"
-          git checkout ${{ env.CI_SHA }}
-          echo "log = $(git log -n 1)"
-
-      - name: Echo folder ${{ matrix.folders }}
-        shell: bash
-        # For folders like `models/bert`, set an env. var. (`matrix_folders`) to `models_bert`, which will be used to
-        # set the artifact folder names (because the character `/` is not allowed).
-        run: |
-          echo "${{ matrix.folders }}"
-          echo "${{ fromJson(needs.setup.outputs.test_map)[matrix.folders] }}"
-          matrix_folders=${{ matrix.folders }}
-          matrix_folders=${matrix_folders/'models/'/'models_'}
-          echo "$matrix_folders"
-          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /transformers
-        run: |
-          python3 utils/print_env.py
-
-      - name: Run all non-slow selected tests on GPU
-        working-directory: /transformers
-        run: |
-          python3 -m pytest -n 2 --dist=loadfile -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} ${{ fromJson(needs.setup.outputs.test_map)[matrix.folders] }}
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
-          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
-
-  run_tests_multi_gpu:
-    name: Model tests
-    needs: setup
-    # `dummy` means there is no test to run
-    if: contains(fromJson(needs.setup.outputs.matrix), 'dummy') != true
-    strategy:
-      fail-fast: false
-      matrix:
-        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
-        machine_type: [multi-gpu]
-    runs-on: [self-hosted, docker-gpu, '${{ matrix.machine_type }}']
-    container:
-      image: huggingface/transformers-all-latest-gpu
-      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    steps:
-      # Necessary to get the correct branch name and commit SHA for `workflow_run` event
-      # We also take into account the `push` event (we might want to test some changes in a branch)
-      - name: Prepare custom environment variables
-        shell: bash
-        # For the meaning of these environment variables, see the job `Setup`
-        run: |
-          CI_BRANCH_PUSH=${{ github.event.ref }}
-          CI_BRANCH_PUSH=${CI_BRANCH_PUSH/'refs/heads/'/''}
-          CI_BRANCH_WORKFLOW_RUN=${{ github.event.workflow_run.head_branch }}
-          CI_SHA_PUSH=${{ github.event.head_commit.id }}
-          CI_SHA_WORKFLOW_RUN=${{ github.event.workflow_run.head_sha }}
-          echo $CI_BRANCH_PUSH
-          echo $CI_BRANCH_WORKFLOW_RUN
-          echo $CI_SHA_PUSH
-          echo $CI_SHA_WORKFLOW_RUN
-          [[ ! -z "$CI_BRANCH_PUSH" ]] && echo "CI_BRANCH=$CI_BRANCH_PUSH" >> $GITHUB_ENV || echo "CI_BRANCH=$CI_BRANCH_WORKFLOW_RUN" >> $GITHUB_ENV
-          [[ ! -z "$CI_SHA_PUSH" ]] && echo "CI_SHA=$CI_SHA_PUSH" >> $GITHUB_ENV || echo "CI_SHA=$CI_SHA_WORKFLOW_RUN" >> $GITHUB_ENV
-
-      - name: print environment variables
-        run: |
-          echo "env.CI_BRANCH = ${{ env.CI_BRANCH }}"
-          echo "env.CI_SHA = ${{ env.CI_SHA }}"
-
-      - name: Update clone using environment variables
-        working-directory: /transformers
-        run: |
-          echo "original branch = $(git branch --show-current)"
-          git fetch && git checkout ${{ env.CI_BRANCH }}
-          echo "updated branch = $(git branch --show-current)"
-          git checkout ${{ env.CI_SHA }}
-          echo "log = $(git log -n 1)"
-
-      - name: Echo folder ${{ matrix.folders }}
-        shell: bash
-        # For folders like `models/bert`, set an env. var. (`matrix_folders`) to `models_bert`, which will be used to
-        # set the artifact folder names (because the character `/` is not allowed).
-        run: |
-          echo "${{ matrix.folders }}"
-          echo "${{ fromJson(needs.setup.outputs.test_map)[matrix.folders] }}"
-          matrix_folders=${{ matrix.folders }}
-          matrix_folders=${matrix_folders/'models/'/'models_'}
-          echo "$matrix_folders"
-          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /transformers
-        run: |
-          python3 utils/print_env.py
-
-      - name: Run all non-slow selected tests on GPU
-        env:
-          MKL_SERVICE_FORCE_INTEL: 1
-        working-directory: /transformers
-        run: |
-          python3 -m pytest -n 2 --dist=loadfile -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} ${{ fromJson(needs.setup.outputs.test_map)[matrix.folders] }}
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
-          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
-
-  run_tests_torch_cuda_extensions_single_gpu:
-    name: Torch CUDA extension tests
-    needs: setup
-    if: contains(fromJson(needs.setup.outputs.matrix), 'deepspeed') || contains(fromJson(needs.setup.outputs.matrix), 'extended')
-    strategy:
-      fail-fast: false
-      matrix:
-        machine_type: [single-gpu]
-    runs-on: [self-hosted, docker-gpu, '${{ matrix.machine_type }}']
-    container:
-      image: huggingface/transformers-pytorch-deepspeed-latest-gpu
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    steps:
-      # Necessary to get the correct branch name and commit SHA for `workflow_run` event
-      # We also take into account the `push` event (we might want to test some changes in a branch)
-      - name: Prepare custom environment variables
-        shell: bash
-        # For the meaning of these environment variables, see the job `Setup`
-        run: |
-          CI_BRANCH_PUSH=${{ github.event.ref }}
-          CI_BRANCH_PUSH=${CI_BRANCH_PUSH/'refs/heads/'/''}
-          CI_BRANCH_WORKFLOW_RUN=${{ github.event.workflow_run.head_branch }}
-          CI_SHA_PUSH=${{ github.event.head_commit.id }}
-          CI_SHA_WORKFLOW_RUN=${{ github.event.workflow_run.head_sha }}
-          echo $CI_BRANCH_PUSH
-          echo $CI_BRANCH_WORKFLOW_RUN
-          echo $CI_SHA_PUSH
-          echo $CI_SHA_WORKFLOW_RUN
-          [[ ! -z "$CI_BRANCH_PUSH" ]] && echo "CI_BRANCH=$CI_BRANCH_PUSH" >> $GITHUB_ENV || echo "CI_BRANCH=$CI_BRANCH_WORKFLOW_RUN" >> $GITHUB_ENV
-          [[ ! -z "$CI_SHA_PUSH" ]] && echo "CI_SHA=$CI_SHA_PUSH" >> $GITHUB_ENV || echo "CI_SHA=$CI_SHA_WORKFLOW_RUN" >> $GITHUB_ENV
-
-      - name: print environment variables
-        run: |
-          echo "env.CI_BRANCH = ${{ env.CI_BRANCH }}"
-          echo "env.CI_SHA = ${{ env.CI_SHA }}"
-
-      - name: Update clone using environment variables
-        working-directory: /workspace/transformers
-        run: |
-          echo "original branch = $(git branch --show-current)"
-          git fetch && git checkout ${{ env.CI_BRANCH }}
-          echo "updated branch = $(git branch --show-current)"
-          git checkout ${{ env.CI_SHA }}
-          echo "log = $(git log -n 1)"
-
-      # To avoid unknown test failures
-      - name: Pre build DeepSpeed *again*
-        working-directory: /workspace
-        run: |
-          python3 -m pip uninstall -y deepspeed
-          DS_BUILD_CPU_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install deepspeed --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /workspace/transformers
-        run: |
-          python utils/print_env.py
-
-      - name: Run all non-slow selected tests on GPU
-        working-directory: /workspace/transformers
-        # TODO: Here we pass all tests in the 2 folders for simplicity. It's better to pass only the identified tests.
-        run: |
-          python -m pytest -n 1 --dist=loadfile -v --make-reports=${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu tests/deepspeed tests/extended
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports
-          path: /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu
-
-  run_tests_torch_cuda_extensions_multi_gpu:
-    name: Torch CUDA extension tests
-    needs: setup
-    if: contains(fromJson(needs.setup.outputs.matrix), 'deepspeed') || contains(fromJson(needs.setup.outputs.matrix), 'extended')
-    strategy:
-      fail-fast: false
-      matrix:
-        machine_type: [multi-gpu]
-    runs-on: [self-hosted, docker-gpu, '${{ matrix.machine_type }}']
-    container:
-      image: huggingface/transformers-pytorch-deepspeed-latest-gpu
-      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    steps:
-      # Necessary to get the correct branch name and commit SHA for `workflow_run` event
-      # We also take into account the `push` event (we might want to test some changes in a branch)
-      - name: Prepare custom environment variables
-        shell: bash
-        # For the meaning of these environment variables, see the job `Setup`
-        run: |
-          CI_BRANCH_PUSH=${{ github.event.ref }}
-          CI_BRANCH_PUSH=${CI_BRANCH_PUSH/'refs/heads/'/''}
-          CI_BRANCH_WORKFLOW_RUN=${{ github.event.workflow_run.head_branch }}
-          CI_SHA_PUSH=${{ github.event.head_commit.id }}
-          CI_SHA_WORKFLOW_RUN=${{ github.event.workflow_run.head_sha }}
-          echo $CI_BRANCH_PUSH
-          echo $CI_BRANCH_WORKFLOW_RUN
-          echo $CI_SHA_PUSH
-          echo $CI_SHA_WORKFLOW_RUN
-          [[ ! -z "$CI_BRANCH_PUSH" ]] && echo "CI_BRANCH=$CI_BRANCH_PUSH" >> $GITHUB_ENV || echo "CI_BRANCH=$CI_BRANCH_WORKFLOW_RUN" >> $GITHUB_ENV
-          [[ ! -z "$CI_SHA_PUSH" ]] && echo "CI_SHA=$CI_SHA_PUSH" >> $GITHUB_ENV || echo "CI_SHA=$CI_SHA_WORKFLOW_RUN" >> $GITHUB_ENV
-
-      - name: print environment variables
-        run: |
-          echo "env.CI_BRANCH = ${{ env.CI_BRANCH }}"
-          echo "env.CI_SHA = ${{ env.CI_SHA }}"
-
-      - name: Update clone using environment variables
-        working-directory: /workspace/transformers
-        run: |
-          echo "original branch = $(git branch --show-current)"
-          git fetch && git checkout ${{ env.CI_BRANCH }}
-          echo "updated branch = $(git branch --show-current)"
-          git checkout ${{ env.CI_SHA }}
-          echo "log = $(git log -n 1)"
-
-      # To avoid unknown test failures
-      - name: Pre build DeepSpeed *again*
-        working-directory: /workspace
-        run: |
-          python3 -m pip uninstall -y deepspeed
-          DS_BUILD_CPU_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install deepspeed --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /workspace/transformers
-        run: |
-          python utils/print_env.py
-
-      - name: Run all non-slow selected tests on GPU
-        working-directory: /workspace/transformers
-        # TODO: Here we pass all tests in the 2 folders for simplicity. It's better to pass only the identified tests.
-        run: |
-          python -m pytest -n 1 --dist=loadfile -v --make-reports=${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu tests/deepspeed tests/extended
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports
-          path: /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu
-
-  send_results:
-    name: Send results to webhook
-    runs-on: ubuntu-latest
-    if: always()
-    needs: [
-        setup,
-        run_tests_single_gpu,
-        run_tests_multi_gpu,
-        run_tests_torch_cuda_extensions_single_gpu,
-        run_tests_torch_cuda_extensions_multi_gpu
-    ]
-    steps:
-      # Necessary to get the correct branch name and commit SHA for `workflow_run` event
-      # We also take into account the `push` event (we might want to test some changes in a branch)
-      - name: Prepare custom environment variables
-        shell: bash
-        # For the meaning of these environment variables, see the job `Setup`
-        run: |
-          CI_BRANCH_PUSH=${{ github.event.ref }}
-          CI_BRANCH_PUSH=${CI_BRANCH_PUSH/'refs/heads/'/''}
-          CI_BRANCH_WORKFLOW_RUN=${{ github.event.workflow_run.head_branch }}
-          CI_SHA_PUSH=${{ github.event.head_commit.id }}
-          CI_SHA_WORKFLOW_RUN=${{ github.event.workflow_run.head_sha }}
-          echo $CI_BRANCH_PUSH
-          echo $CI_BRANCH_WORKFLOW_RUN
-          echo $CI_SHA_PUSH
-          echo $CI_SHA_WORKFLOW_RUN
-          [[ ! -z "$CI_BRANCH_PUSH" ]] && echo "CI_BRANCH=$CI_BRANCH_PUSH" >> $GITHUB_ENV || echo "CI_BRANCH=$CI_BRANCH_WORKFLOW_RUN" >> $GITHUB_ENV
-          [[ ! -z "$CI_SHA_PUSH" ]] && echo "CI_SHA=$CI_SHA_PUSH" >> $GITHUB_ENV || echo "CI_SHA=$CI_SHA_WORKFLOW_RUN" >> $GITHUB_ENV
-
-      - name: print environment variables
-        run: |
-          echo "env.CI_BRANCH = ${{ env.CI_BRANCH }}"
-          echo "env.CI_SHA = ${{ env.CI_SHA }}"
-
       - uses: actions/checkout@v2
-
-      - name: Update clone using environment variables
+      - name: Python version
         run: |
-          echo "original branch = $(git branch --show-current)"
-          git fetch && git checkout ${{ env.CI_BRANCH }}
-          echo "updated branch = $(git branch --show-current)"
-          git checkout ${{ env.CI_SHA }}
-          echo "log = $(git log -n 1)"
+          which python
+          python --version
+          pip --version
 
-      - uses: actions/download-artifact@v2
-      - name: Send message to Slack
+      - name: Current dir
+        run: pwd
+      - run: nvidia-smi
+
+      - name: Loading cache.
+        uses: actions/cache@v2
+        id: cache
+        with:
+          path: .env
+          key: v1.1-tests_torch_gpu-${{ hashFiles('setup.py') }}
+
+      - name: Create new python env (on self-hosted runners we have to handle isolation ourselves)
+        run: |
+          python -m venv .env
+          source .env/bin/activate
+          which python
+          python --version
+          pip --version
+
+      - name: Install dependencies
+        run: |
+          source .env/bin/activate
+          pip install --upgrade pip
+          pip install .[torch,sklearn,testing,onnxruntime,sentencepiece]
+          pip install git+https://github.com/huggingface/datasets
+
+      - name: Are GPUs recognized by our DL frameworks
+        run: |
+          source .env/bin/activate
+          python -c "import torch; print('Cuda available:', torch.cuda.is_available())"
+          python -c "import torch; print('Number of GPUs available:', torch.cuda.device_count())"
+
+      - name: Create model files
+        run: |
+          source .env/bin/activate
+          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/encoder-bert-tokenizer.json --path=templates/adding_a_new_model
+          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/pt-encoder-bert-tokenizer.json --path=templates/adding_a_new_model
+          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/standalone.json --path=templates/adding_a_new_model
+          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/tf-encoder-bert-tokenizer.json --path=templates/adding_a_new_model
+
+      - name: Run all non-slow tests on GPU
         env:
-          CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
-          CI_SLACK_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID }}
-          CI_SLACK_CHANNEL_ID_DAILY: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY }}
-          CI_SLACK_CHANNEL_DUMMY_TESTS: ${{ secrets.CI_SLACK_CHANNEL_DUMMY_TESTS }}
-          CI_SLACK_REPORT_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID }}
-          CI_EVENT: push
-          CI_TITLE_PUSH: ${{ github.event.head_commit.message }}
-          CI_TITLE_WORKFLOW_RUN: ${{ github.event.workflow_run.head_commit.message }}
-          CI_SHA: ${{ env.CI_SHA }}
-        # We pass `needs.setup.outputs.matrix` as the argument. A processing in `notification_service.py` to change
-        # `models/bert` to `models_bert` is required, as the artifact names use `_` instead of `/`.
+          OMP_NUM_THREADS: 1
+          CUDA_VISIBLE_DEVICES: 0
         run: |
-          pip install slack_sdk
-          python utils/notification_service.py "${{ needs.setup.outputs.matrix }}"
+          source .env/bin/activate
+          python -m pytest -n 2 --dist=loadfile -s --make-reports=tests_torch_gpu tests
+
+      - name: Failure short reports
+        if: ${{ always() }}
+        run: cat reports/tests_torch_gpu_failures_short.txt
+        
+      - name: Test suite reports artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: run_all_tests_torch_gpu_test_reports
+          path: reports
+                  
+
+  run_tests_tf_gpu:
+    runs-on: [self-hosted, gpu, single-gpu]
+    steps:
+      - uses: actions/checkout@v2
+      - name: Python version
+        run: |
+          which python
+          python --version
+          pip --version
+      - name: Current dir
+        run: pwd
+      - run: nvidia-smi
+
+      - name: Loading cache.
+        uses: actions/cache@v2
+        id: cache
+        with:
+          path: .env
+          key: v1.1-tests_tf_gpu-${{ hashFiles('setup.py') }}
+
+      - name: Create new python env (on self-hosted runners we have to handle isolation ourselves)
+        run: |
+          python -m venv .env
+          source .env/bin/activate
+          which python
+          python --version
+          pip --version
+
+      - name: Install dependencies
+        run: |
+          source .env/bin/activate
+          pip install --upgrade pip
+          pip install .[tf,sklearn,testing,onnxruntime,sentencepiece]
+          pip install git+https://github.com/huggingface/datasets
+
+      - name: Are GPUs recognized by our DL frameworks
+        run: |
+          source .env/bin/activate
+          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('TF GPUs available:', bool(tf.config.list_physical_devices('GPU')))"
+          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('Number of TF GPUs available:', len(tf.config.list_physical_devices('GPU')))"
+
+      - name: Create model files
+        run: |
+          source .env/bin/activate
+          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/encoder-bert-tokenizer.json --path=templates/adding_a_new_model
+          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/pt-encoder-bert-tokenizer.json --path=templates/adding_a_new_model
+          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/standalone.json --path=templates/adding_a_new_model
+          transformers-cli add-new-model --testing --testing_file=templates/adding_a_new_model/tests/tf-encoder-bert-tokenizer.json --path=templates/adding_a_new_model
+
+      - name: Run all non-slow tests on GPU
+        env:
+          OMP_NUM_THREADS: 1
+          CUDA_VISIBLE_DEVICES: 0
+        run: |
+          source .env/bin/activate
+          python -m pytest -n 2 --dist=loadfile -s --make-reports=tests_tf_gpu tests
+
+      - name: Failure short reports
+        if: ${{ always() }}
+        run: cat reports/tests_tf_gpu_failures_short.txt
+        
+      - name: Test suite reports artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: run_all_tests_tf_gpu_test_reports
+          path: reports
+
+  run_tests_torch_multi_gpu:
+    runs-on: [self-hosted, gpu, multi-gpu]
+    steps:
+      - uses: actions/checkout@v2
+      - name: Python version
+        run: |
+          which python
+          python --version
+          pip --version
+
+      - name: Current dir
+        run: pwd
+      - run: nvidia-smi
+
+      - name: Loading cache.
+        uses: actions/cache@v2
+        id: cache
+        with:
+          path: .env
+          key: v1.1-tests_torch_multi_gpu-${{ hashFiles('setup.py') }}
+
+      - name: Create new python env (on self-hosted runners we have to handle isolation ourselves)
+        run: |
+          python -m venv .env
+          source .env/bin/activate
+          which python
+          python --version
+          pip --version
+      - name: Install dependencies
+        run: |
+          source .env/bin/activate
+          pip install --upgrade pip
+          pip install .[torch,sklearn,testing,onnxruntime,sentencepiece]
+          pip install git+https://github.com/huggingface/datasets
+
+      - name: Are GPUs recognized by our DL frameworks
+        run: |
+          source .env/bin/activate
+          python -c "import torch; print('Cuda available:', torch.cuda.is_available())"
+          python -c "import torch; print('Number of GPUs available:', torch.cuda.device_count())"
+
+      - name: Run all non-slow tests on GPU
+        env:
+          OMP_NUM_THREADS: 1
+        run: |
+          source .env/bin/activate
+          python -m pytest -n 2 --dist=loadfile -s --make-reports=tests_torch_multi_gpu tests
+
+      - name: Failure short reports
+        if: ${{ always() }}
+        run: cat reports/tests_torch_multi_gpu_failures_short.txt          
+
+      - name: Test suite reports artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: run_all_tests_torch_multi_gpu_test_reports
+          path: reports
+
+  run_tests_tf_multi_gpu:
+    runs-on: [self-hosted, gpu, multi-gpu]
+    steps:
+      - uses: actions/checkout@v2
+      - name: Python version
+        run: |
+          which python
+          python --version
+          pip --version
+
+      - name: Current dir
+        run: pwd
+      - run: nvidia-smi
+
+      - name: Loading cache.
+        uses: actions/cache@v2
+        id: cache
+        with:
+          path: .env
+          key: v1.1-tests_tf_multi_gpu-${{ hashFiles('setup.py') }}
+
+      - name: Create new python env (on self-hosted runners we have to handle isolation ourselves)
+        run: |
+          python -m venv .env
+          source .env/bin/activate
+          which python
+          python --version
+          pip --version
+      - name: Install dependencies
+        run: |
+          source .env/bin/activate
+          pip install --upgrade pip
+          pip install .[tf,sklearn,testing,onnxruntime,sentencepiece]
+          pip install git+https://github.com/huggingface/datasets
+
+      - name: Are GPUs recognized by our DL frameworks
+        run: |
+          source .env/bin/activate
+          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('TF GPUs available:', bool(tf.config.list_physical_devices('GPU')))"
+          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('Number of TF GPUs available:', len(tf.config.list_physical_devices('GPU')))"
+
+      - name: Run all non-slow tests on GPU
+        env:
+          OMP_NUM_THREADS: 1
+        run: |
+          source .env/bin/activate
+          python -m pytest -n 2 --dist=loadfile -s --make-reports=tests_tf_multi_gpu tests
+
+      - name: Failure short reports
+        if: ${{ always() }}
+        run: cat reports/tests_tf_multi_gpu_failures_short.txt
+
+      - name: Test suite reports artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: run_all_tests_tf_multi_gpu_test_reports
+          path: reports
+          

.github/workflows/self-scheduled.yml

@@ -1,362 +1,359 @@
+# configuration notes:
+#
+# - `source .env/bin/activate` is currently needed to be run first thing first in each step. Otherwise
+# the step uses the system-wide python interpreter.
+
 name: Self-hosted runner (scheduled)
 
-# Note that each job's dependencies go into a corresponding docker file.
-#
-# For example for `run_all_tests_torch_cuda_extensions_gpu` the docker image is
-# `huggingface/transformers-pytorch-deepspeed-latest-gpu`, which can be found at
-# `docker/transformers-pytorch-deepspeed-latest-gpu/Dockerfile`
-
 on:
+  push:
+    branches:
+      - ci_*
   repository_dispatch:
   schedule:
-    - cron: "0 2 * * *"
-
-env:
-  HF_HOME: /mnt/cache
-  TRANSFORMERS_IS_CI: yes
-  OMP_NUM_THREADS: 8
-  MKL_NUM_THREADS: 8
-  RUN_SLOW: yes
-  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
-  TF_FORCE_GPU_ALLOW_GROWTH: true
-  RUN_PT_TF_CROSS_TESTS: 1
+    - cron: "0 0 * * *"
 
 jobs:
-  setup:
-    name: Setup
-    strategy:
-      matrix:
-        machine_type: [single-gpu, multi-gpu]
-    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
-    container:
-      image: huggingface/transformers-all-latest-gpu
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    outputs:
-      matrix: ${{ steps.set-matrix.outputs.matrix }}
-    steps:
-      - name: Update clone
-        working-directory: /transformers
-        run: |
-          git fetch && git checkout ${{ github.sha }}
-
-      - name: Cleanup
-        working-directory: /transformers
-        run: |
-          rm -rf tests/__pycache__
-          rm -rf tests/models/__pycache__
-          rm -rf reports
-
-      - id: set-matrix
-        name: Identify models to test
-        working-directory: /transformers/tests
-        run: |
-          echo "::set-output name=matrix::$(python3 -c 'import os; tests = os.getcwd(); model_tests = os.listdir(os.path.join(tests, "models")); d1 = sorted(list(filter(os.path.isdir, os.listdir(tests)))); d2 = sorted(list(filter(os.path.isdir, [f"models/{x}" for x in model_tests]))); d1.remove("models"); d = d2 + d1; print(d)')"
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-  run_tests_single_gpu:
-    name: Model tests
-    strategy:
-      fail-fast: false
-      matrix:
-        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
-        machine_type: [single-gpu]
-    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
-    container:
-      image: huggingface/transformers-all-latest-gpu
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    needs: setup
-    steps:
-      - name: Echo folder ${{ matrix.folders }}
-        shell: bash
-        # For folders like `models/bert`, set an env. var. (`matrix_folders`) to `models_bert`, which will be used to
-        # set the artifact folder names (because the character `/` is not allowed).
-        run: |
-          echo "${{ matrix.folders }}"
-          matrix_folders=${{ matrix.folders }}
-          matrix_folders=${matrix_folders/'models/'/'models_'}
-          echo "$matrix_folders"
-          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
-
-      - name: Update clone
-        working-directory: /transformers
-        run: git fetch && git checkout ${{ github.sha }}
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /transformers
-        run: |
-          python3 utils/print_env.py
-
-      - name: Run all tests on GPU
-        working-directory: /transformers
-        run: python3 -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
-          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
-
-  run_tests_multi_gpu:
-    name: Model tests
-    strategy:
-      fail-fast: false
-      matrix:
-        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
-        machine_type: [multi-gpu]
-    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
-    container:
-      image: huggingface/transformers-all-latest-gpu
-      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    needs: setup
-    steps:
-      - name: Echo folder ${{ matrix.folders }}
-        shell: bash
-        # For folders like `models/bert`, set an env. var. (`matrix_folders`) to `models_bert`, which will be used to
-        # set the artifact folder names (because the character `/` is not allowed).
-        run: |
-          echo "${{ matrix.folders }}"
-          matrix_folders=${{ matrix.folders }}
-          matrix_folders=${matrix_folders/'models/'/'models_'}
-          echo "$matrix_folders"
-          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
-
-      - name: Update clone
-        working-directory: /transformers
-        run: git fetch && git checkout ${{ github.sha }}
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /transformers
-        run: |
-          python3 utils/print_env.py
-
-      - name: Run all tests on GPU
-        working-directory: /transformers
-        run: python3 -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
-          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
-
-  run_examples_gpu:
-    name: Examples directory
-    runs-on: [self-hosted, single-gpu-docker]
-    container:
-      image: huggingface/transformers-all-latest-gpu
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    needs: setup
-    steps:
-      - name: Update clone
-        working-directory: /transformers
-        run: git fetch && git checkout ${{ github.sha }}
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /transformers
-        run: |
-          python3 utils/print_env.py
-
-      - name: Run examples tests on GPU
-        working-directory: /transformers
-        run: |
-          pip install -r examples/pytorch/_tests_requirements.txt
-          python3 -m pytest -v --make-reports=single-gpu_examples_gpu examples/pytorch
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /transformers/reports/single-gpu_examples_gpu/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: single-gpu_run_examples_gpu
-          path: /transformers/reports/single-gpu_examples_gpu
-
-  run_pipelines_torch_gpu:
-    name: PyTorch pipelines
-    strategy:
-      fail-fast: false
-      matrix:
-        machine_type: [single-gpu, multi-gpu]
-    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
-    container:
-      image: huggingface/transformers-pytorch-gpu
-      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    needs: setup
-    steps:
-      - name: Update clone
-        working-directory: /transformers
-        run: git fetch && git checkout ${{ github.sha }}
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /transformers
-        run: |
-          python3 utils/print_env.py
-
-      - name: Run all pipeline tests on GPU
-        working-directory: /transformers
-        env:
-          RUN_PIPELINE_TESTS: yes
-        run: |
-          python3 -m pytest -n 1 -v --dist=loadfile -m is_pipeline_test --make-reports=${{ matrix.machine_type }}_tests_torch_pipeline_gpu tests
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_torch_pipeline_gpu/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.machine_type }}_run_tests_torch_pipeline_gpu
-          path: /transformers/reports/${{ matrix.machine_type }}_tests_torch_pipeline_gpu
-
-  run_pipelines_tf_gpu:
-    name: TensorFlow pipelines
-    strategy:
-      fail-fast: false
-      matrix:
-        machine_type: [single-gpu, multi-gpu]
-    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
-    container:
-      image: huggingface/transformers-tensorflow-gpu
-      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    needs: setup
-    steps:
-      - name: Update clone
-        working-directory: /transformers
-        run: |
-          git fetch && git checkout ${{ github.sha }}
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /transformers
-        run: |
-          python3 utils/print_env.py
-
-      - name: Run all pipeline tests on GPU
-        working-directory: /transformers
-        env:
-          RUN_PIPELINE_TESTS: yes
-        run: |
-          python3 -m pytest -n 1 -v --dist=loadfile -m is_pipeline_test --make-reports=${{ matrix.machine_type }}_tests_tf_pipeline_gpu tests
-
-      - name: Failure short reports
-        if: ${{ always() }}
-        run: |
-          cat /transformers/reports/${{ matrix.machine_type }}_tests_tf_pipeline_gpu/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.machine_type }}_run_tests_tf_pipeline_gpu
-          path: /transformers/reports/${{ matrix.machine_type }}_tests_tf_pipeline_gpu
-
-  run_all_tests_torch_cuda_extensions_gpu:
-    name: Torch CUDA extension tests
-    strategy:
-      fail-fast: false
-      matrix:
-        machine_type: [single-gpu, multi-gpu]
-    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
-    needs: setup
-    container:
-      image: huggingface/transformers-pytorch-deepspeed-latest-gpu
-      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    steps:
-      - name: Update clone
-        working-directory: /workspace/transformers
-        run: git fetch && git checkout ${{ github.sha }}
-
-      # To avoid unknown test failures
-      - name: Pre build DeepSpeed *again*
-        working-directory: /workspace
-        run: |
-          python3 -m pip uninstall -y deepspeed
-          DS_BUILD_CPU_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install deepspeed --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check
-
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-
-      - name: Environment
-        working-directory: /workspace/transformers
-        run: |
-          python utils/print_env.py
-
-      - name: Run all tests on GPU
-        working-directory: /workspace/transformers
-        run: |
-          python -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu tests/deepspeed tests/extended
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu/failures_short.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports
-          path: /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu
-
-  send_results:
-    name: Send results to webhook
-    runs-on: ubuntu-latest
-    if: always()
-    needs: [setup, run_tests_single_gpu, run_tests_multi_gpu, run_examples_gpu, run_pipelines_tf_gpu, run_pipelines_torch_gpu, run_all_tests_torch_cuda_extensions_gpu]
+  run_all_tests_torch_gpu:
+    runs-on: [self-hosted, gpu, single-gpu]
     steps:
       - uses: actions/checkout@v2
-      - uses: actions/download-artifact@v2
-      - name: Send message to Slack
-        env:
-          CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
-          CI_SLACK_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID }}
-          CI_SLACK_CHANNEL_ID_DAILY: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY }}
-          CI_SLACK_CHANNEL_DUMMY_TESTS: ${{ secrets.CI_SLACK_CHANNEL_DUMMY_TESTS }}
-          CI_SLACK_REPORT_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY }}
-          CI_EVENT: scheduled
-        # We pass `needs.setup.outputs.matrix` as the argument. A processing in `notification_service.py` to change
-        # `models/bert` to `models_bert` is required, as the artifact names use `_` instead of `/`.
+
+      - name: Loading cache.
+        uses: actions/cache@v2
+        id: cache
+        with:
+          path: .env
+          key: v  1.1-slow_tests_torch_gpu-${{ hashFiles('setup.py') }}
+
+      - name: Python version
         run: |
-          pip install slack_sdk
-          python utils/notification_service.py "${{ needs.setup.outputs.matrix }}"
+          which python
+          python --version
+          pip --version
+
+      - name: Current dir
+        run: pwd
+      - run: nvidia-smi
+
+      - name: Create new python env (on self-hosted runners we have to handle isolation ourselves)
+        if: steps.cache.outputs.cache-hit != 'true'
+        run: |
+          python -m venv .env
+          source .env/bin/activate
+          which python
+          python --version
+          pip --version
+
+      - name: Install dependencies
+        run: |
+          source .env/bin/activate
+          pip install --upgrade pip
+          pip install .[torch,sklearn,testing,onnxruntime,sentencepiece]
+          pip install git+https://github.com/huggingface/datasets
+          pip list
+
+      - name: Are GPUs recognized by our DL frameworks
+        run: |
+          source .env/bin/activate
+          python -c "import torch; print('Cuda available:', torch.cuda.is_available())"
+          python -c "import torch; print('Number of GPUs available:', torch.cuda.device_count())"
+
+      - name: Run all tests on GPU
+        env:
+          OMP_NUM_THREADS: 1
+          RUN_SLOW: yes
+        run: |
+          source .env/bin/activate
+          python -m pytest -n 1 --dist=loadfile -s --make-reports=tests_torch_gpu tests
+
+      - name: Failure short reports
+        if: ${{ always() }}
+        run: cat reports/tests_torch_gpu_failures_short.txt
+        
+      - name: Run examples tests on GPU
+        if: ${{ always() }}
+        env:
+          OMP_NUM_THREADS: 1
+          RUN_SLOW: yes
+        run: |
+          source .env/bin/activate
+          pip install -r examples/requirements.txt
+          python -m pytest -n 1 --dist=loadfile -s --make-reports=examples_torch_gpu examples
+
+      - name: Failure short reports
+        if: ${{ always() }}
+        run: cat reports/examples_torch_gpu_failures_short.txt
+
+      - name: Run all pipeline tests on GPU
+        if: ${{ always() }}
+        env:
+          TF_FORCE_GPU_ALLOW_GROWTH: "true"
+          OMP_NUM_THREADS: 1
+          RUN_SLOW: yes
+          RUN_PIPELINE_TESTS: yes
+        run: |
+          source .env/bin/activate
+          python -m pytest -n 1 --dist=loadfile -s -m is_pipeline_test --make-reports=tests_torch_pipeline_gpu tests
+
+      - name: Failure short reports
+        if: ${{ always() }}
+        run: cat reports/tests_torch_pipeline_gpu_failures_short.txt
+
+      - name: Test suite reports artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: run_all_tests_torch_gpu_test_reports
+          path: reports
+
+
+  run_all_tests_tf_gpu:
+    runs-on: [self-hosted, gpu, single-gpu]
+    steps:
+      - uses: actions/checkout@v2
+
+      - name: Loading cache.
+        uses: actions/cache@v2
+        id: cache
+        with:
+          path: .env
+          key: v1.1-slow_tests_tf_gpu-${{ hashFiles('setup.py') }}
+
+      - name: Python version
+        run: |
+          which python
+          python --version
+          pip --version
+
+      - name: Current dir
+        run: pwd
+      - run: nvidia-smi
+
+      - name: Create new python env (on self-hosted runners we have to handle isolation ourselves)
+        if: steps.cache.outputs.cache-hit != 'true'
+        run: |
+          python -m venv .env
+          source .env/bin/activate
+          which python
+          python --version
+          pip --version
+
+      - name: Install dependencies
+        run: |
+          source .env/bin/activate
+          pip install --upgrade pip
+          pip install .[tf,sklearn,testing,onnxruntime,sentencepiece]
+          pip install git+https://github.com/huggingface/datasets
+          pip list
+
+      - name: Are GPUs recognized by our DL frameworks
+        run: |
+          source .env/bin/activate
+          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('TF GPUs available:', bool(tf.config.list_physical_devices('GPU')))"
+          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('Number of TF GPUs available:', len(tf.config.list_physical_devices('GPU')))"
+
+      - name: Run all tests on GPU
+        env:
+          OMP_NUM_THREADS: 1
+          RUN_SLOW: yes
+        run: |
+          source .env/bin/activate
+          python -m pytest -n 1 --dist=loadfile -s --make-reports=tests_tf_gpu tests
+          
+      - name: Failure short reports
+        if: ${{ always() }}
+        run: cat reports/tests_tf_gpu_failures_short.txt
+
+      - name: Run all pipeline tests on GPU
+        if: ${{ always() }}
+        env:
+          TF_FORCE_GPU_ALLOW_GROWTH: "true"
+          OMP_NUM_THREADS: 1
+          RUN_SLOW: yes
+          RUN_PIPELINE_TESTS: yes
+        run: |
+          source .env/bin/activate
+          python -m pytest -n 1 --dist=loadfile -s -m is_pipeline_test --make-reports=tests_tf_pipelines_gpu tests
+
+      - name: Failure short reports
+        if: ${{ always() }}
+        run: cat reports/tests_tf_pipelines_gpu_failures_short.txt
+
+      - name: Test suite reports artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: run_all_tests_tf_gpu_test_reports
+          path: reports
+          
+  run_all_tests_torch_multi_gpu:
+    runs-on: [self-hosted, gpu, multi-gpu]
+    steps:
+      - uses: actions/checkout@v2
+
+      - name: Loading cache.
+        uses: actions/cache@v2
+        id: cache
+        with:
+          path: .env
+          key: v1.1-slow_tests_torch_multi_gpu-${{ hashFiles('setup.py') }}
+
+      - name: Python version
+        run: |
+          which python
+          python --version
+          pip --version
+
+      - name: Current dir
+        run: pwd
+      - run: nvidia-smi
+
+      - name: Create new python env (on self-hosted runners we have to handle isolation ourselves)
+        if: steps.cache.outputs.cache-hit != 'true'
+        run: |
+          python -m venv .env
+          source .env/bin/activate
+          which python
+          python --version
+          pip --version
+
+      - name: Install dependencies
+        run: |
+          source .env/bin/activate
+          pip install --upgrade pip
+          pip install .[torch,sklearn,testing,onnxruntime,sentencepiece]
+          pip install git+https://github.com/huggingface/datasets
+          pip list
+
+      - name: Are GPUs recognized by our DL frameworks
+        run: |
+          source .env/bin/activate
+          python -c "import torch; print('Cuda available:', torch.cuda.is_available())"
+          python -c "import torch; print('Number of GPUs available:', torch.cuda.device_count())"
+
+      - name: Run all tests on multi-GPU
+        env:
+          OMP_NUM_THREADS: 1
+          RUN_SLOW: yes
+        run: |
+          source .env/bin/activate
+          python -m pytest -n 1 --dist=loadfile -s --make-reports=tests_torch_multi_gpu tests
+
+      - name: Failure short reports
+        if: ${{ always() }}
+        run: cat reports/tests_torch_multi_gpu_failures_short.txt
+
+      - name: Run examples tests on multi-GPU
+        env:
+          OMP_NUM_THREADS: 1
+          RUN_SLOW: yes
+        run: |
+          source .env/bin/activate
+          python -m pytest -n 1 --dist=loadfile -s --make-reports=tests_torch_examples_multi_gpu examples
+
+      - name: Failure short reports
+        if: ${{ always() }}
+        run: cat reports/tests_torch_examples_multi_gpu_failures_short.txt
+
+      - name: Run all pipeline tests on multi-GPU
+        if: ${{ always() }}
+        env:
+          TF_FORCE_GPU_ALLOW_GROWTH: "true"
+          OMP_NUM_THREADS: 1
+          RUN_SLOW: yes
+          RUN_PIPELINE_TESTS: yes
+        run: |
+          source .env/bin/activate
+          python -m pytest -n 1 --dist=loadfile -s -m is_pipeline_test --make-reports=tests_torch_pipeline_multi_gpu tests
+
+      - name: Failure short reports
+        if: ${{ always() }}
+        run: cat reports/tests_torch_pipeline_multi_gpu_failures_short.txt
+
+      - name: Test suite reports artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: run_all_tests_torch_multi_gpu_test_reports
+          path: reports
+
+  run_all_tests_tf_multi_gpu:
+    runs-on: [self-hosted, gpu, multi-gpu]
+    steps:
+      - uses: actions/checkout@v2
+
+      - name: Loading cache.
+        uses: actions/cache@v2
+        id: cache
+        with:
+          path: .env
+          key: v1.1-slow_tests_tf_multi_gpu-${{ hashFiles('setup.py') }}
+
+      - name: Python version
+        run: |
+          which python
+          python --version
+          pip --version
+
+      - name: Current dir
+        run: pwd
+      - run: nvidia-smi
+
+      - name: Create new python env (on self-hosted runners we have to handle isolation ourselves)
+        if: steps.cache.outputs.cache-hit != 'true'
+        run: |
+          python -m venv .env
+          source .env/bin/activate
+          which python
+          python --version
+          pip --version
+
+      - name: Install dependencies
+        run: |
+          source .env/bin/activate
+          pip install --upgrade pip
+          pip install .[tf,sklearn,testing,onnxruntime,sentencepiece]
+          pip install git+https://github.com/huggingface/datasets
+          pip list
+
+      - name: Are GPUs recognized by our DL frameworks
+        run: |
+          source .env/bin/activate
+          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('TF GPUs available:', bool(tf.config.list_physical_devices('GPU')))"
+          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('Number of TF GPUs available:', len(tf.config.list_physical_devices('GPU')))"
+
+      - name: Run all tests on multi-GPU
+        env:
+          OMP_NUM_THREADS: 1
+          RUN_SLOW: yes
+        run: |
+          source .env/bin/activate
+          python -m pytest -n 1 --dist=loadfile -s --make-reports=tests_tf_multi_gpu tests
+
+      - name: Failure short reports
+        if: ${{ always() }}
+        run: cat reports/tests_tf_multi_gpu_failures_short.txt
+
+      - name: Run all pipeline tests on multi-GPU
+        if: ${{ always() }}
+        env:
+          TF_FORCE_GPU_ALLOW_GROWTH: "true"
+          OMP_NUM_THREADS: 1
+          RUN_SLOW: yes
+          RUN_PIPELINE_TESTS: yes
+        run: |
+          source .env/bin/activate
+          python -m pytest -n 1 --dist=loadfile -s -m is_pipeline_test --make-reports=tests_tf_pipeline_multi_gpu tests
+          
+      - name: Failure short reports
+        if: ${{ always() }}
+        run: cat reports/tests_tf_pipeline_multi_gpu_failures_short.txt
+
+      - name: Test suite reports artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: run_all_tests_tf_multi_gpu_test_reports
+          path: reports
+          

.github/workflows/stale.yml

@@ -1,27 +0,0 @@
-name: Stale Bot
-
-on:
-  schedule:
-    - cron: "0 15 * * *"
-
-jobs:
-  close_stale_issues:
-    name: Close Stale Issues
-    if: github.repository == 'huggingface/transformers'
-    runs-on: ubuntu-latest
-    env:
-      GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-    steps:
-    - uses: actions/checkout@v2
-
-    - name: Setup Python
-      uses: actions/setup-python@v1
-      with:
-        python-version: 3.7
-
-    - name: Install requirements
-      run: |
-        pip install PyGithub
-    - name: Close stale issues
-      run: |
-        python scripts/stale.py

.github/workflows/update_metdata.yml

@@ -1,40 +0,0 @@
-name: Update Transformers metadata
-
-on:
-  push:
-    branches:
-      - main
-      - update_transformers_metadata
-
-jobs:
-  build_and_package:
-    runs-on: ubuntu-latest
-    defaults:
-      run:
-        shell: bash -l {0}
-
-    steps:
-      - uses: actions/checkout@v2
-
-      - name: Load cached virtual environment
-        uses: actions/cache@v2
-        id: cache
-        with:
-          path: ~/venv/
-          key: v3-metadata-${{ hashFiles('setup.py') }}
-
-      - name: Create virtual environment on cache miss
-        if: steps.cache.outputs.cache-hit != 'true'
-        run: |
-          python -m venv ~/venv && . ~/venv/bin/activate
-          pip install --upgrade pip
-
-      - name: Setup environment
-        run: |
-          . ~/venv/bin/activate
-          pip install git+https://github.com/huggingface/transformers#egg=transformers[dev]
-
-      - name: Update metadata
-        run: |
-          . ~/venv/bin/activate
-          python utils/update_metadata.py --token ${{ secrets.SYLVAIN_HF_TOKEN }} --commit_sha ${{ github.sha }}

.gitignore

@@ -9,7 +9,8 @@ __pycache__/
 *.so
 
 # tests and logs
-tests/fixtures/cached_*_text.txt
+tests/fixtures/*
+!tests/fixtures/sample_text_no_unicode.txt
 logs/
 lightning_logs/
 lang_code_data/
@@ -158,9 +159,3 @@ tags
 
 # pre-commit
 .pre-commit*
-
-# .lock
-*.lock
-
-# DS_Store (MacOS)
-.DS_Store

CITATION.cff

@@ -1,82 +0,0 @@
-cff-version: "1.2.0"
-date-released: 2020-10
-message: "If you use this software, please cite it using these metadata."
-title: "Transformers: State-of-the-Art Natural Language Processing"
-url: "https://github.com/huggingface/transformers"
-authors: 
-  - family-names: Wolf
-    given-names: Thomas
-  - family-names: Debut
-    given-names: Lysandre
-  - family-names: Sanh
-    given-names: Victor
-  - family-names: Chaumond
-    given-names: Julien
-  - family-names: Delangue
-    given-names: Clement
-  - family-names: Moi
-    given-names: Anthony
-  - family-names: Cistac
-    given-names: Perric
-  - family-names: Ma
-    given-names: Clara
-  - family-names: Jernite
-    given-names: Yacine
-  - family-names: Plu
-    given-names: Julien
-  - family-names: Xu
-    given-names: Canwen
-  - family-names: "Le Scao"
-    given-names: Teven
-  - family-names: Gugger
-    given-names: Sylvain
-  - family-names: Drame
-    given-names: Mariama
-  - family-names: Lhoest
-    given-names: Quentin
-  - family-names: Rush
-    given-names: "Alexander M."
-preferred-citation:
-  type: conference-paper
-  authors:
-  - family-names: Wolf
-    given-names: Thomas
-  - family-names: Debut
-    given-names: Lysandre
-  - family-names: Sanh
-    given-names: Victor
-  - family-names: Chaumond
-    given-names: Julien
-  - family-names: Delangue
-    given-names: Clement
-  - family-names: Moi
-    given-names: Anthony
-  - family-names: Cistac
-    given-names: Perric
-  - family-names: Ma
-    given-names: Clara
-  - family-names: Jernite
-    given-names: Yacine
-  - family-names: Plu
-    given-names: Julien
-  - family-names: Xu
-    given-names: Canwen
-  - family-names: "Le Scao"
-    given-names: Teven
-  - family-names: Gugger
-    given-names: Sylvain
-  - family-names: Drame
-    given-names: Mariama
-  - family-names: Lhoest
-    given-names: Quentin
-  - family-names: Rush
-    given-names: "Alexander M."
-  booktitle: "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations"
-  month: 10
-  start: 38
-  end: 45
-  title: "Transformers: State-of-the-Art Natural Language Processing"
-  year: 2020
-  publisher: "Association for Computational Linguistics"
-  url: "https://www.aclweb.org/anthology/2020.emnlp-demos.6"
-  address: "Online"

CONTRIBUTING.md

@@ -1,19 +1,3 @@
-<!---
-Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
--->
-
 # How to contribute to transformers?
 
 Everyone is welcome to contribute, and we value everybody's contribution. Code
@@ -26,7 +10,7 @@ on the awesome projects it made possible, shout out on Twitter every time it has
 helped you, or simply star the repo to say "thank you".
 
 Whichever way you choose to contribute, please be mindful to respect our
-[code of conduct](https://github.com/huggingface/transformers/blob/main/CODE_OF_CONDUCT.md).
+[code of conduct](https://github.com/huggingface/transformers/blob/master/CODE_OF_CONDUCT.md).
 
 ## You can contribute in so many ways!
 
@@ -36,13 +20,6 @@ There are 4 ways you can contribute to transformers:
 * Contributing to the examples or to the documentation;
 * Submitting issues related to bugs or desired new features.
 
-In particular there is a special [Good First
-Issue](https://github.com/huggingface/transformers/contribute) listing. It will give you a list of
-open Issues that are open to anybody to work on. Just comment in the issue that you'd like to work
-on it. In that same listing you will also find some Issues with `Good Second Issue` label. These are
-typically slightly more complicated than the Issues with just `Good First Issue` label. But if you
-feel you know what you're doing, go for it.
-
 *All are equally valuable to the community.*
 
 ## Submitting a new issue or feature request
@@ -53,7 +30,7 @@ feedback.
 
 ### Did you find a bug?
 
-The 🤗 Transformers library is robust and reliable thanks to the users who notify us of
+The transformers are robust and reliable thanks to the users who notify us of
 the problems they encounter. So thank you for reporting an issue.
 
 First, we would really appreciate it if you could **make sure the bug was not
@@ -92,7 +69,7 @@ If you are willing to contribute the model yourself, let us know so we can best
 guide you.
 
 We have added a **detailed guide and templates** to guide you in the process of adding a new model. You can find them
-in the [`templates`](https://github.com/huggingface/transformers/tree/main/templates) folder.
+in the [`templates`](https://github.com/huggingface/transformers/tree/master/templates) folder.
 
 ### Do you want a new feature (that is not a model)?
 
@@ -114,7 +91,7 @@ If your issue is well written we're already 80% of the way there by the time you
 post it.
 
 We have added **templates** to guide you in the process of adding a new example script for training or testing the
-models in the library. You can find them in the [`templates`](https://github.com/huggingface/transformers/tree/main/templates)
+models in the library. You can find them in the [`templates`](https://github.com/huggingface/transformers/tree/master/templates)
 folder.
 
 ## Start contributing! (Pull Requests)
@@ -124,11 +101,11 @@ issues to make sure that nobody is already working on the same thing. If you are
 unsure, it is always a good idea to open an issue to get some feedback.
 
 You will need basic `git` proficiency to be able to contribute to
-🤗 Transformers. `git` is not the easiest tool to use but it has the greatest
+`transformers`. `git` is not the easiest tool to use but it has the greatest
 manual. Type `git --help` in a shell and enjoy. If you prefer books, [Pro
 Git](https://git-scm.com/book/en/v2) is a very good reference.
 
-Follow these steps to start contributing ([supported Python versions](https://github.com/huggingface/transformers/blob/main/setup.py#L426)):
+Follow these steps to start contributing:
 
 1. Fork the [repository](https://github.com/huggingface/transformers) by
    clicking on the 'Fork' button on the repository's page. This creates a copy of the code
@@ -148,7 +125,7 @@ Follow these steps to start contributing ([supported Python versions](https://gi
    $ git checkout -b a-descriptive-name-for-my-changes
    ```
 
-   **Do not** work on the `main` branch.
+   **do not** work on the `master` branch.
 
 4. Set up a development environment by running the following command in a virtual environment:
 
@@ -175,26 +152,34 @@ Follow these steps to start contributing ([supported Python versions](https://gi
 5. Develop the features on your branch.
 
    As you work on the features, you should make sure that the test suite
-   passes. You should run the tests impacted by your changes like this:
-
-   ```bash
-   $ pytest tests/<TEST_TO_RUN>.py
-   ```
-
-   You can also run the full suite with the following command, but it takes
-   a beefy machine to produce a result in a decent amount of time now that
-   Transformers has grown a lot. Here is the command for it:
+   passes:
 
    ```bash
    $ make test
    ```
 
-   For more information about tests, check out the
-   [dedicated documentation](https://huggingface.co/docs/transformers/testing)
+   Note, that this command uses `-n auto` pytest flag, therefore, it will start as many parallel `pytest` processes as the number of your computer's CPU-cores, and if you have lots of those and a few GPUs and not a great amount of RAM, it's likely to overload your computer. Therefore, to run the test suite, you may want to consider using this command instead:
 
-   🤗 Transformers relies on `black` and `isort` to format its source code
-   consistently. After you make changes, apply automatic style corrections and code verifications
-   that can't be automated in one go with:
+   ```bash
+   $ python -m pytest -n 3 --dist=loadfile -s -v ./tests/
+   ```
+
+   Adjust the value of `-n` to fit the load your hardware can support.
+
+   `transformers` relies on `black` and `isort` to format its source code
+   consistently. After you make changes, format them with:
+
+   ```bash
+   $ make style
+   ```
+
+   `transformers` also uses `flake8` and a few custom scripts to check for coding mistakes. Quality
+   control runs in CI, however you can also run the same checks with:
+
+   ```bash
+   $ make quality
+   ```
+   You can do the automatic style corrections and code verifications that can't be automated in one go:
 
    ```bash
    $ make fixup
@@ -202,55 +187,16 @@ Follow these steps to start contributing ([supported Python versions](https://gi
 
    This target is also optimized to only work with files modified by the PR you're working on.
 
-   If you prefer to run the checks one after the other, the following command apply the
-   style corrections:
-
-   ```bash
-   $ make style
-   ```
-
-   🤗 Transformers also uses `flake8` and a few custom scripts to check for coding mistakes. Quality
-   control runs in CI, however you can also run the same checks with:
-
-   ```bash
-   $ make quality
-   ```
-
-   Finally we have a lot of scripts that check we didn't forget to update
-   some files when adding a new model, that you can run with
-
-   ```bash
-   $ make repo-consistency
-   ```
-
-   To learn more about those checks and how to fix any issue with them, check out the
-   [documentation](https://huggingface.co/docs/transformers/pr_checks)
-
    If you're modifying documents under `docs/source`, make sure to validate that
    they can still be built. This check also runs in CI. To run a local check
-   make sure you have installed the documentation builder requirements. First you will need to clone the
-   repository containing our tools to build the documentation:
-   
-   ```bash
-   $ pip install git+https://github.com/huggingface/doc-builder
-   ```
-
-   Then, make sure you have all the dependencies to be able to build the doc with:
-   
-   ```bash
-   $ pip install ".[docs]"
-   ```
-
-   Finally run the following command from the root of the repository:
+   make sure you have installed the documentation builder requirements, by
+   running `pip install .[tf,torch,docs]` once from the root of this repository
+   and then run:
 
    ```bash
-   $ doc-builder build transformers docs/source/ --build_dir ~/tmp/test-build
+   $ make docs
    ```
 
-   This will build the documentation in the `~/tmp/test-build` folder where you can inspect the generated
-   Markdown files with your favorite editor. You won't be able to see the final rendering on the website
-   before your PR is merged, we are actively working on adding a tool for this.
-
    Once you're happy with your changes, add changed files using `git add` and
    make a commit with `git commit` to record your changes locally:
 
@@ -267,7 +213,7 @@ Follow these steps to start contributing ([supported Python versions](https://gi
 
    ```bash
    $ git fetch upstream
-   $ git rebase upstream/main
+   $ git rebase upstream/master
    ```
 
    Push the changes to your account using:
@@ -304,21 +250,14 @@ Follow these steps to start contributing ([supported Python versions](https://gi
    - If you are adding a new tokenizer, write tests, and make sure
      `RUN_SLOW=1 python -m pytest tests/test_tokenization_{your_model_name}.py` passes.
    CircleCI does not run the slow tests, but github actions does every night!
-6. All public methods must have informative docstrings that work nicely with sphinx. See `modeling_bert.py` for an
+6. All public methods must have informative docstrings that work nicely with sphinx. See `modeling_ctrl.py` for an
    example.
-7. Due to the rapidly growing repository, it is important to make sure that no files that would significantly weigh down the repository are added. This includes images, videos and other non-text files. We prefer to leverage a hf.co hosted `dataset` like
-   the ones hosted on [`hf-internal-testing`](https://huggingface.co/hf-internal-testing) in which to place these files and reference 
-   them by URL. We recommend putting them in the following dataset: [huggingface/documentation-images](https://huggingface.co/datasets/huggingface/documentation-images).
-   If an external contribution, feel free to add the images to your PR and ask a Hugging Face member to migrate your images
-   to this dataset.
-
-See more about the checks run on a pull request in our [PR guide](pr_checks)
 
 ### Tests
 
 An extensive test suite is included to test the library behavior and several examples. Library tests can be found in
-the [tests folder](https://github.com/huggingface/transformers/tree/main/tests) and examples tests in the
-[examples folder](https://github.com/huggingface/transformers/tree/main/examples).
+the [tests folder](https://github.com/huggingface/transformers/tree/master/tests) and examples tests in the
+[examples folder](https://github.com/huggingface/transformers/tree/master/examples).
 
 We like `pytest` and `pytest-xdist` because it's faster. From the root of the
 repository, here's how to run tests with `pytest` for the library:
@@ -330,7 +269,7 @@ $ python -m pytest -n auto --dist=loadfile -s -v ./tests/
 and for the examples:
 
 ```bash
-$ pip install -r examples/xxx/requirements.txt  # only needed the first time
+$ pip install -r examples/requirements.txt  # only needed the first time
 $ python -m pytest -n auto --dist=loadfile -s -v ./examples/
 ```
 In fact, that's how `make test` and `make test-examples` are implemented (sans the `pip install` line)!
@@ -364,36 +303,17 @@ $ python -m unittest discover -s examples -t examples -v
 
 ### Style guide
 
-For documentation strings, 🤗 Transformers follows the [google style](https://google.github.io/styleguide/pyguide.html).
-Check our [documentation writing guide](https://github.com/huggingface/transformers/tree/main/docs#writing-documentation---specification)
+For documentation strings, `transformers` follows the [google style](https://google.github.io/styleguide/pyguide.html).
+Check our [documentation writing guide](https://github.com/huggingface/transformers/tree/master/docs#writing-documentation---specification)
 for more information.
 
-**This guide was heavily inspired by the awesome [scikit-learn guide to contributing](https://github.com/scikit-learn/scikit-learn/blob/main/CONTRIBUTING.md).**
+#### This guide was heavily inspired by the awesome [scikit-learn guide to contributing](https://github.com/scikit-learn/scikit-learn/blob/master/CONTRIBUTING.md)
+
 
 ### Develop on Windows
 
-On windows, you need to configure git to transform Windows `CRLF` line endings to Linux `LF` line endings:
-
-`git config core.autocrlf input`
-
 One way one can run the make command on Window is to pass by MSYS2:
 
 1. [Download MSYS2](https://www.msys2.org/), we assume to have it installed in C:\msys64
 2. Open the command line C:\msys64\msys2.exe (it should be available from the start menu)
 3. Run in the shell: `pacman -Syu` and install make with `pacman -S make`
-4. Add `C:\msys64\usr\bin` to your PATH environment variable.
-
-You can now use `make` from any terminal (Powershell, cmd.exe, etc) 🎉
-
-### Syncing forked main with upstream (HuggingFace) main
-
-To avoid pinging the upstream repository which adds reference notes to each upstream PR and sends unnecessary notifications to the developers involved in these PRs,
-when syncing the main branch of a forked repository, please, follow these steps:
-1. When possible, avoid syncing with the upstream using a branch and PR on the forked repository. Instead merge directly into the forked main.
-2. If a PR is absolutely necessary, use the following steps after checking out your branch:
-```
-$ git checkout -b your-branch-for-syncing
-$ git pull --squash --no-commit upstream main
-$ git commit -m '<your message without GitHub references>'
-$ git push --set-upstream origin your-branch-for-syncing
-```

ISSUES.md

@@ -1,277 +0,0 @@
-<!---
-Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
--->
-
-# How To Request Support
-
-This is an Open Source Project so please be mindful that like in any other project of this kind there is no obligation to answer all requests for help.
-
-However, we want to encourage you to ask for help whenever you think it's needed! We are happy about every  question we get because it allows us to better understand your needs, possible misunderstandings, and most importantly a way for you to help us make this library better. That being said, this document's main purpose is to provide guidelines at how you can formulate your requests to increase your chances to be understood and to get support.
-
-There are two main venues to receive support: [the forums](https://discuss.huggingface.co/) and [the GitHub issues](https://github.com/huggingface/transformers/issues).
-
-## The Forums
-
-[The user forums](https://discuss.huggingface.co/) are supported by the wide community of the library users and backed up by developers when needed.
-
-If you have a difficulty with deploying this library or some questions, or you'd like to discuss a new feature, please first consider discussing those things at the forums. Only when you feel your subject matter has been crystalized and you still need support from the library developers do proceed to file an [issue](https://github.com/huggingface/transformers/issues).
-
-In particular all "Please explain" questions or objectively very user-specific feature requests belong to the forums. Here are some example of such questions:
-
-* "I would like to use a BertModel within a RL-Agent for a customer support service. How can I use a BertForMaskedLM in my ChatBotModel?"
-
-* "Could you please explain why T5 has no positional embedding matrix under T5Model?"
-
-* "How should I set my generation parameters for translation?"
-
-* "How to train T5 on De->En translation?"
-
-
-## The GitHub Issues
-
-Everything which hints at a bug should be opened as an [issue](https://github.com/huggingface/transformers/issues).
-
-You are not required to read the following guidelines before opening an issue. However, if you notice that your issue doesn't get any replies, chances are that the developers have one or several difficulties with its quality. In this case, reading the following points and adjusting your issue accordingly could help.
-
-1. Before posting an issue, first search for already posted issues, since chances are someone has already asked a similar question before you.
-
-    If you use Google your search query should be:
-
-    ```
-    "huggingface" "transformers" your query
-    ```
-
-    The first two quoted words tell Google to limit the search to the context of the Huggingface Transformers. The remainder is your query - most commonly this would be the error message the software fails with. We will go deeper into details shortly.
-
-    The results of such a query will typically match GitHub issues, Hugging Face forums, StackExchange, and blogs.
-
-    If you find relevant hints, you may choose to continue the discussion there if you have follow up questions.
-
-    If what you found is similar but doesn't quite answer your problem, please, post a new issue and do include links to similar issues or forum discussions you may have found.
-
-    Let's look at some examples:
-
-    The error message, often referred to as an assertion, tells us what went wrong. Here is an example of an assertion:
-
-   ```python
-   Traceback (most recent call last):
-     File "<string>", line 1, in <module>
-     File "/transformers/src/transformers/__init__.py", line 34, in <module>
-       from . import dependency_versions_check
-     File "/transformers/src/transformers/dependency_versions_check.py", line 34, in <module>
-       from .utils import is_tokenizers_available
-     File "/transformers/src/transformers/utils/import_utils.py", line 40, in <module>
-       from tqdm.auto import tqdm
-    ModuleNotFoundError: No module named 'tqdm.auto'
-    ```
-
-   and it typically includes a traceback, so that we can see the full stack of calls the program made before it fails. This gives us the context to know why the program failed.
-
-   Going back to the above example. If you received this error search, look at the very last line of the error which is:
-
-   ```python
-    ModuleNotFoundError: No module named 'tqdm.auto'
-    ```
-
-    And now we can use it to do the searching on your favorite search engine:
-
-    1. first for `"huggingface" "transformers" "ModuleNotFoundError: No module named 'tqdm.auto'"`
-    2. if you don't find relevant results, then search for just `"ModuleNotFoundError: No module named 'tqdm.auto'"`
-    3. and finally if nothing still comes up, then remove the outside quotes: `ModuleNotFoundError: No module named 'tqdm.auto'`
-
-   If the error includes any messages that include bits unique to your filesystem, always remove those in the search query since other users will not have the same filesystem as yours. For example:
-
-   ```bash
-   python -c 'open("/tmp/wrong_path.txt", "r")'
-   Traceback (most recent call last):
-     File "<string>", line 1, in <module>
-   FileNotFoundError: [Errno 2] No such file or directory: '/tmp/wrong_path.txt'
-   ```
-   Here you'd search for just: `"FileNotFoundError: [Errno 2] No such file or directory"`
-
-   If the local information that you removed were inside the error message and you removed them you may need to remove double quotes since your query is no longer exact. So if the error message was something like:
-
-   ```bash
-      ValueError: '/tmp/wrong_path.txt' cannot be found
-   ```
-
-   then you'd search for `"ValueError" "cannot be found"`
-
-   As you search you will notice that when you don't use quotes often the search engines will return a variety of unrelated hits, which may or may not be what you want.
-
-   Experiment with different ways and find which approach gives the most satisfactory results.
-
-2. Keep the issue short, providing the information that you think will aid the developers to understand your situation. Put yourself in the shoes of the person who has never seen your code or knows anything about your custom setup. This mental exercise will help to develop an intuition to what/what not to share"
-
-3. If there is a software failure, always provide the full traceback, for example:
-
-   ```python
-   $ python -c 'import transformers'
-   Traceback (most recent call last):
-     File "<string>", line 1, in <module>
-     File "/transformers/src/transformers/__init__.py", line 34, in <module>
-       from . import dependency_versions_check
-     File "/transformers/src/transformers/dependency_versions_check.py", line 34, in <module>
-       from .utils import is_tokenizers_available
-     File "/transformers/src/transformers/utils/import_utils.py", line 40, in <module>
-       from tqdm.auto import tqdm
-   ModuleNotFoundError: No module named 'tqdm.auto'
-   ```
-
-   As compared to providing just the last line of the error message, e.g.:
-   ```python
-   ModuleNotFoundError: No module named 'tqdm.auto'
-   ```
-   which is not sufficient.
-
-   If your application is running on more than one GPU (e.g. under `DistributedDataParallel`) and typically getting every log and traceback printed multiple times, please make sure that you paste only one copy of it. At times the traceback from parallel processes may get interleaved - so either disentangle these or change the loggers to log only for `local_rank==0` so that only one process logs things.
-
-4. When quoting a traceback, command line instructions and any type of code always enclose it in triple backticks inside the editor window, that is:
-
-   ````
-   ```
-   git clone https://github.com/huggingface/transformers
-   cd transformers
-   pip install .
-   ```
-   ````
-
-   If it's a command line with a long argument list, please consider breaking it down using backslashes and new lines. Here is an example of a good command line quote:
-
-   ```bash
-    cd examples/seq2seq
-    python -m torch.distributed.launch --nproc_per_node=2 ./finetune_trainer.py \
-    --model_name_or_path sshleifer/distill-mbart-en-ro-12-4 --data_dir wmt_en_ro \
-    --output_dir output_dir --overwrite_output_dir \
-    --do_train --n_train 500 --num_train_epochs 1 \
-    --per_device_train_batch_size 1  --freeze_embeds \
-    --src_lang en_XX --tgt_lang ro_RO --task translation \
-    --fp16 --sharded_ddp
-   ```
-
-   If you don't break it up, one has to scroll horizontally which often makes it quite difficult to quickly see what's happening.
-
-   The backslashes allow us to copy the command directly into the console to run it, without needing to edit it.
-
-5. Include only the important information that you think will help the developer to quickly identify the problem.
-
-   For example applications often create huge amounts of logs. Ask yourself whether providing all or parts of the log is useful.
-
-   Pasting a 100-1000 lines of log into the issue is an immediate turn off, since it will take a lot of time to figure out where the pertinent parts of the log are.
-
-   Attaching a full log can be helpful if it's done as an attachment, if it's enclosed in the following html code in the comment editor window:
-
-   ```
-   <details>
-   <summary>Full log</summary>
-   <pre>
-
-   many
-   lines
-   go
-   here
-
-   </pre>
-   </details>
-   ```
-
-   which would result in the following entry, which can be opened if desired, but otherwise takes little space.
-
-   <details>
-   <summary>Full log</summary>
-   <pre>
-   many
-   lines
-   go
-   here
-   </pre>
-   </details>
-
-    You could also provide a link to a pastebin service, but this is less beneficial since those links tend to expire quickly and future readers of your issue might not be able to access that log file anymore and may lack some context.
-
-6. If this is an issue in your code, do try to reduce that code to a minimal example that still demonstrates the problem. Please ask at the forums if you have a hard time figuring how to do that. Please realize that we don't have the luxury of having time to try and understand all of your custom code.
-
-   If you really tried to make a short reproducible code but couldn't figure it out, it might be that having a traceback will give the developer enough information to know what's going on. But if it is not enough and we can't reproduce the problem, we can't really solve it.
-
-   Do not despair if you can't figure it out from the beginning, just share what you can and perhaps someone else will be able to help you at the forums.
-
-   If your setup involves any custom datasets, the best way to help us reproduce the problem is to create a [Google Colab notebook](https://colab.research.google.com/) that demonstrates the issue and once you verify that the issue still exists, include a link to that notebook in the Issue. Just make sure that you don't copy and paste the location bar url of the open notebook - as this is private and we won't be able to open it. Instead, you need to click on `Share` in the right upper corner of the notebook, select `Get Link` and then copy and paste the public link it will give to you.
-
-7. If you forked off some of this project's code or example applications, please, do not ask us to go into your code repository and figure out what you may have done. The code is already very complex and unless there is an easy way to do a diff and it's a small diff, it won't be possible to find someone with time on their hands to make a lengthy investigation. Albeit, you might find someone at the forums who will be generous to do this for you.
-
-8. Before reporting an issue, first, always try to update your environment to the latest official version of this library. We have no resources to go and debug older revisions, which could easily have bugs that have been fixed in the latest released version.
-
-   We understand that this is not always possible, especially when APIs change, in which case file an issue against the highest library version your environment can support.
-
-   Of course, if you upgrade the library, always retest that the problem is still there.
-
-9. Please do not ask us to reproduce an issue with your custom data, since we don't have it. So, either you should use some existing dataset supported by HF datasets or you need to supply a code that generates a small sample on the fly, or some another quick and simple way to get it.
-
-   Please do not send us any non-public domain data that may require a license or a permission to be used.
-
-10. Do not tag multiple developers on the issue unless you know this is expected, either because you asked them and they gave you an explicit permission to tag them or the issue template instructs you to do so.
-
-   The "who to tag for what domain" part of the issue template is there to help users direct their questions to the right developers who are designated maintainers of project's specific domains. They can then decide at their own discretion to tag other developers if they feel it'd help move the issue forward.
-
-   We currently don't have a triage service and we trust your capacity to identify the right domain and thus the persons to tag in your issue. If you are not sure, please use the forums to ask for guidance.
-
-   When in doubt, err on the side of not tagging a given person. If you tag multiple people out of context or permission don't be surprised if you get no response at all. Please remember that every time you tag someone, they get a notification and you're taking their time without their permission. Please be sensitive to that.
-
-   If you got helped by one of the developers in the past please don't tag them in future issues, unless they are listed in the issue template for the domain you are asking about or that developer gave you an explicit permission to tag them in future issues.
-
-   If you see a certain developer doing multiple and/or recent commits into a specific area of the project that you feel is relevant to your issue, it is not a good reason to tag them. Various developers may be fixing things that prevent them from moving forward, but often their work is focused on a totally different domain. And while they may or may not know how to help you with the problem at hand, it would benefit the whole community much more if they focus on the domain of their unique expertise.
-
-11. Use the Edit button. Take your time, and re-read and improve the wording and formatting to make your posts and comments as easy to understand as possible.
-
-    Avoid posting multiple comments in a row, as each comment generates a notification for the developers tagged in that issue. If you happened to post multiple comments in a row, and nobody followed up yet - consider merging those into one or a few comments while editing the combined content to be coherent.
-
-    If you choose to edit your older comments after others posted follow up comments you need to be aware that your modifications might not be noticed, so if it's not a typo fixing, try to write a new comment flagging that something has been changed in the previous comments.
-
-    For example, the very first comment is the most important one. If while the thread unfolds you realize that things aren't as they seemed to you originally you may want to edit the first post to reflect the up-to-date understanding of the issue at hand so that it helps those who read your issue in the future quickly understand what's going on and not need to sift through dozens of comments. It also helps to indicate that the post was edited. So, those reading the thread later can understand why there might be certain discontinuity in the information flow.
-
-    Use bullets and items if you have lists of items and the outcome improves overall readability.
-
-    Use backticks to refer to class and function names, e.g. `BartModel` and `generate` as these stand out and improve the speed of a reader's comprehension.
-
-    Try not use italics and bold text too much as these often make the text more difficult to read.
-
-
-12. If you are cross-referencing a specific comment in a given thread or another issue, always link to that specific comment, rather than using the issue link. If you do the latter it could be quite impossible to find which specific comment you're referring to.
-
-    To get the link to the specific comment do not copy the url from the location bar of your browser, but instead, click the `...` icon in the upper right corner of the comment and then select "Copy Link".
-
-    For example the first link is a link to an issue, and the second to a specific comment in the same issue:
-
-    1. https://github.com/huggingface/transformers/issues/9257
-    2. https://github.com/huggingface/transformers/issues/9257#issuecomment-749945162
-
-
-13. If you are replying to a last comment, it's totally fine to make your reply with just your comment in it. The readers can follow the information flow here.
-
-    But if you're replying to a comment that happened some comments back it's always a good practice to quote just the relevant lines you're replying it. The `>` is used for quoting, or you can always use the menu to do so. For example your editor box will look like:
-
-    ```
-    > How big is your gpu cluster?
-
-    Our cluster is made of 256 gpus.
-    ```
-
-    If you are addressing multiple comments, quote the relevant parts of each before your answer. Some people use the same comment to do multiple replies, others separate them into separate comments. Either way works. The latter approach helps for linking to a specific comment.
-
-In general the best way to figure out what works the best is learn from issues posted by other people - see which issues get great responses and which get little to no response - observe what the posters who received great responses did differently from those who did not.
-
-Thank you for reading this somewhat lengthy document. We would like to conclude that these are not absolute rules, but a friendly advice that will help maximize the chances for us to understand what you are trying to communicate, reproduce the problem then resolve it to your satisfaction and the benefit of the whole community.
-
-If after reading this document there are remaining questions on how and why or there is a need for further elucidation, please, don't hesitate to ask your question in [this thread](https://discuss.huggingface.co/t/how-to-request-support/3128).

LICENSE

@@ -1,4 +1,3 @@
-Copyright 2018- The Hugging Face team. All rights reserved.
 
                                  Apache License
                            Version 2.0, January 2004

Makefile

@@ -1,7 +1,5 @@
-.PHONY: deps_table_update modified_only_fixup extra_style_checks quality style fixup fix-copies test test-examples
+.PHONY: modified_only_fixup extra_quality_checks quality style fixup fix-copies test test-examples docs
 
-# make sure to test the local checkout in scripts and not the pre-installed one (don't use quotes!)
-export PYTHONPATH = src
 
 check_dirs := examples tests src utils
 
@@ -9,75 +7,44 @@ modified_only_fixup:
 	$(eval modified_py_files := $(shell python utils/get_modified_files.py $(check_dirs)))
 	@if test -n "$(modified_py_files)"; then \
 		echo "Checking/fixing $(modified_py_files)"; \
-		black --preview $(modified_py_files); \
+		black $(modified_py_files); \
 		isort $(modified_py_files); \
 		flake8 $(modified_py_files); \
 	else \
 		echo "No library .py files were modified"; \
 	fi
 
-# Update src/transformers/dependency_versions_table.py
+# Check that source code meets quality standards
 
-deps_table_update:
-	@python setup.py deps_table_update
-
-deps_table_check_updated:
-	@md5sum src/transformers/dependency_versions_table.py > md5sum.saved
-	@python setup.py deps_table_update
-	@md5sum -c --quiet md5sum.saved || (printf "\nError: the version dependency table is outdated.\nPlease run 'make fixup' or 'make style' and commit the changes.\n\n" && exit 1)
-	@rm md5sum.saved
-
-# autogenerating code
-
-autogenerate_code: deps_table_update
-
-# Check that the repo is in a good state
-
-repo-consistency:
+extra_quality_checks:
 	python utils/check_copies.py
-	python utils/check_table.py
 	python utils/check_dummies.py
 	python utils/check_repo.py
-	python utils/check_inits.py
-	python utils/check_config_docstrings.py
-	python utils/tests_fetcher.py --sanity_check
+	python utils/style_doc.py src/transformers docs/source --max_len 119
 
 # this target runs checks on all files
-
 quality:
-	black --check --preview $(check_dirs)
+	black --check $(check_dirs)
 	isort --check-only $(check_dirs)
-	python utils/custom_init_isort.py --check_only
-	python utils/sort_auto_mappings.py --check_only
 	flake8 $(check_dirs)
-	doc-builder style src/transformers docs/source --max_len 119 --check_only --path_to_docs docs/source
-	python utils/check_doc_toc.py
+	python utils/style_doc.py src/transformers docs/source --max_len 119 --check_only
+	${MAKE} extra_quality_checks
 
 # Format source code automatically and check is there are any problems left that need manual fixing
 
-extra_style_checks:
-	python utils/custom_init_isort.py
-	python utils/sort_auto_mappings.py
-	doc-builder style src/transformers docs/source --max_len 119 --path_to_docs docs/source
-	python utils/check_doc_toc.py --fix_and_overwrite
-
-# this target runs checks on all files and potentially modifies some of them
-
 style:
-	black --preview $(check_dirs)
+	black $(check_dirs)
 	isort $(check_dirs)
-	${MAKE} autogenerate_code
-	${MAKE} extra_style_checks
+	python utils/style_doc.py src/transformers docs/source --max_len 119
 
 # Super fast fix and check target that only works on relevant modified files since the branch was made
 
-fixup: modified_only_fixup extra_style_checks autogenerate_code repo-consistency
+fixup: modified_only_fixup extra_quality_checks
 
 # Make marked copies of snippets of codes conform to the original
 
 fix-copies:
 	python utils/check_copies.py --fix_and_overwrite
-	python utils/check_table.py --fix_and_overwrite
 	python utils/check_dummies.py --fix_and_overwrite
 
 # Run tests for the library
@@ -88,24 +55,9 @@ test:
 # Run tests for examples
 
 test-examples:
-	python -m pytest -n auto --dist=loadfile -s -v ./examples/pytorch/
+	python -m pytest -n auto --dist=loadfile -s -v ./examples/
 
-# Run tests for SageMaker DLC release
+# Check that docs can build
 
-test-sagemaker: # install sagemaker dependencies in advance with pip install .[sagemaker]
-	TEST_SAGEMAKER=True python -m pytest -n auto  -s -v ./tests/sagemaker
-
-
-# Release stuff
-
-pre-release:
-	python utils/release.py
-
-pre-patch:
-	python utils/release.py --patch
-
-post-release:
-	python utils/release.py --post_release
-
-post-patch:
-	python utils/release.py --post_release --patch
+docs:
+	cd docs && make html SPHINXOPTS="-W"

README.md

@@ -1,163 +1,87 @@
-<!---
-Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
--->
-
 <p align="center">
     <br>
-    <img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers_logo_name.png" width="400"/>
+    <img src="https://raw.githubusercontent.com/huggingface/transformers/master/docs/source/imgs/transformers_logo_name.png" width="400"/>
     <br>
 <p>
 <p align="center">
     <a href="https://circleci.com/gh/huggingface/transformers">
-        <img alt="Build" src="https://img.shields.io/circleci/build/github/huggingface/transformers/main">
+        <img alt="Build" src="https://img.shields.io/circleci/build/github/huggingface/transformers/master">
     </a>
-    <a href="https://github.com/huggingface/transformers/blob/main/LICENSE">
+    <a href="https://github.com/huggingface/transformers/blob/master/LICENSE">
         <img alt="GitHub" src="https://img.shields.io/github/license/huggingface/transformers.svg?color=blue">
     </a>
-    <a href="https://huggingface.co/docs/transformers/index">
-        <img alt="Documentation" src="https://img.shields.io/website/http/huggingface.co/docs/transformers/index.svg?down_color=red&down_message=offline&up_message=online">
+    <a href="https://huggingface.co/transformers/index.html">
+        <img alt="Documentation" src="https://img.shields.io/website/http/huggingface.co/transformers/index.html.svg?down_color=red&down_message=offline&up_message=online">
     </a>
     <a href="https://github.com/huggingface/transformers/releases">
         <img alt="GitHub release" src="https://img.shields.io/github/release/huggingface/transformers.svg">
     </a>
-    <a href="https://github.com/huggingface/transformers/blob/main/CODE_OF_CONDUCT.md">
+    <a href="https://github.com/huggingface/transformers/blob/master/CODE_OF_CONDUCT.md">
         <img alt="Contributor Covenant" src="https://img.shields.io/badge/Contributor%20Covenant-v2.0%20adopted-ff69b4.svg">
     </a>
-    <a href="https://zenodo.org/badge/latestdoi/155220641"><img src="https://zenodo.org/badge/155220641.svg" alt="DOI"></a>
 </p>
 
-<h4 align="center">
-    <p>
-        <b>English</b> |
-        <a href="https://github.com/huggingface/transformers/blob/main/README_zh-hans.md">简体中文</a> |
-        <a href="https://github.com/huggingface/transformers/blob/main/README_zh-hant.md">繁體中文</a> |
-        <a href="https://github.com/huggingface/transformers/blob/main/README_ko.md">한국어</a>
-    <p>
-</h4>
-
 <h3 align="center">
-    <p>State-of-the-art Machine Learning for JAX, PyTorch and TensorFlow</p>
+<p>State-of-the-art Natural Language Processing for PyTorch and TensorFlow 2.0
 </h3>
 
-<h3 align="center">
-    <a href="https://hf.co/course"><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/course_banner.png"></a>
-</h3>
+🤗 Transformers provides thousands of pretrained models to perform tasks on texts such as classification, information extraction, question answering, summarization, translation, text generation, etc in 100+ languages. Its aim is to make cutting-edge NLP easier to use for everyone.
 
-🤗 Transformers provides thousands of pretrained models to perform tasks on different modalities such as text, vision, and audio. 
+🤗 Transformers provides APIs to quickly download and use those pretrained models on a given text, fine-tune them on your own datasets then share them with the community on our [model hub](https://huggingface.co/models). At the same time, each python module defining an architecture can be used as a standalone and modified to enable quick research experiments.
 
-These models can be applied on:
+🤗 Transformers is backed by the two most popular deep learning libraries, [PyTorch](https://pytorch.org/) and [TensorFlow](https://www.tensorflow.org/), with a seamless integration between them, allowing you to train your models with one then load it for inference with the other.
 
-* 📝 Text, for tasks like text classification, information extraction, question answering, summarization, translation, text generation, in over 100 languages. 
-* 🖼️ Images, for tasks like image classification, object detection, and segmentation. 
-* 🗣️ Audio, for tasks like speech recognition and audio classification. 
-
-Transformer models can also perform tasks on **several modalities combined**, such as table question answering, optical character recognition, information extraction from scanned documents, video classification, and visual question answering.
-
-🤗 Transformers provides APIs to quickly download and use those pretrained models on a given text, fine-tune them on your own datasets and then share them with the community on our [model hub](https://huggingface.co/models). At the same time, each python module defining an architecture is fully standalone and can be modified to enable quick research experiments.
-
-🤗 Transformers is backed by the three most popular deep learning libraries — [Jax](https://jax.readthedocs.io/en/latest/), [PyTorch](https://pytorch.org/) and [TensorFlow](https://www.tensorflow.org/) — with a seamless integration between them. It's straightforward to train your models with one before loading them for inference with the other.
+### Recent contributors
+[![](https://sourcerer.io/fame/clmnt/huggingface/transformers/images/0)](https://sourcerer.io/fame/clmnt/huggingface/transformers/links/0)[![](https://sourcerer.io/fame/clmnt/huggingface/transformers/images/1)](https://sourcerer.io/fame/clmnt/huggingface/transformers/links/1)[![](https://sourcerer.io/fame/clmnt/huggingface/transformers/images/2)](https://sourcerer.io/fame/clmnt/huggingface/transformers/links/2)[![](https://sourcerer.io/fame/clmnt/huggingface/transformers/images/3)](https://sourcerer.io/fame/clmnt/huggingface/transformers/links/3)[![](https://sourcerer.io/fame/clmnt/huggingface/transformers/images/4)](https://sourcerer.io/fame/clmnt/huggingface/transformers/links/4)[![](https://sourcerer.io/fame/clmnt/huggingface/transformers/images/5)](https://sourcerer.io/fame/clmnt/huggingface/transformers/links/5)[![](https://sourcerer.io/fame/clmnt/huggingface/transformers/images/6)](https://sourcerer.io/fame/clmnt/huggingface/transformers/links/6)[![](https://sourcerer.io/fame/clmnt/huggingface/transformers/images/7)](https://sourcerer.io/fame/clmnt/huggingface/transformers/links/7)
 
 ## Online demos
 
-You can test most of our models directly on their pages from the [model hub](https://huggingface.co/models). We also offer [private model hosting, versioning, & an inference API](https://huggingface.co/pricing) for public and private models.
+You can test most of our models directly on their pages from the [model hub](https://huggingface.co/models). We also offer an [inference API](https://huggingface.co/pricing) to use those models.
 
 Here are a few examples:
-
- In Natural Language Processing:
 - [Masked word completion with BERT](https://huggingface.co/bert-base-uncased?text=Paris+is+the+%5BMASK%5D+of+France)
 - [Name Entity Recognition with Electra](https://huggingface.co/dbmdz/electra-large-discriminator-finetuned-conll03-english?text=My+name+is+Sarah+and+I+live+in+London+city)
 - [Text generation with GPT-2](https://huggingface.co/gpt2?text=A+long+time+ago%2C+)
-- [Natural Language Inference with RoBERTa](https://huggingface.co/roberta-large-mnli?text=The+dog+was+lost.+Nobody+lost+any+animal)
+- [Natural Langugage Inference with RoBERTa](https://huggingface.co/roberta-large-mnli?text=The+dog+was+lost.+Nobody+lost+any+animal)
 - [Summarization with BART](https://huggingface.co/facebook/bart-large-cnn?text=The+tower+is+324+metres+%281%2C063+ft%29+tall%2C+about+the+same+height+as+an+81-storey+building%2C+and+the+tallest+structure+in+Paris.+Its+base+is+square%2C+measuring+125+metres+%28410+ft%29+on+each+side.+During+its+construction%2C+the+Eiffel+Tower+surpassed+the+Washington+Monument+to+become+the+tallest+man-made+structure+in+the+world%2C+a+title+it+held+for+41+years+until+the+Chrysler+Building+in+New+York+City+was+finished+in+1930.+It+was+the+first+structure+to+reach+a+height+of+300+metres.+Due+to+the+addition+of+a+broadcasting+aerial+at+the+top+of+the+tower+in+1957%2C+it+is+now+taller+than+the+Chrysler+Building+by+5.2+metres+%2817+ft%29.+Excluding+transmitters%2C+the+Eiffel+Tower+is+the+second+tallest+free-standing+structure+in+France+after+the+Millau+Viaduct)
 - [Question answering with DistilBERT](https://huggingface.co/distilbert-base-uncased-distilled-squad?text=Which+name+is+also+used+to+describe+the+Amazon+rainforest+in+English%3F&context=The+Amazon+rainforest+%28Portuguese%3A+Floresta+Amaz%C3%B4nica+or+Amaz%C3%B4nia%3B+Spanish%3A+Selva+Amaz%C3%B3nica%2C+Amazon%C3%ADa+or+usually+Amazonia%3B+French%3A+For%C3%AAt+amazonienne%3B+Dutch%3A+Amazoneregenwoud%29%2C+also+known+in+English+as+Amazonia+or+the+Amazon+Jungle%2C+is+a+moist+broadleaf+forest+that+covers+most+of+the+Amazon+basin+of+South+America.+This+basin+encompasses+7%2C000%2C000+square+kilometres+%282%2C700%2C000+sq+mi%29%2C+of+which+5%2C500%2C000+square+kilometres+%282%2C100%2C000+sq+mi%29+are+covered+by+the+rainforest.+This+region+includes+territory+belonging+to+nine+nations.+The+majority+of+the+forest+is+contained+within+Brazil%2C+with+60%25+of+the+rainforest%2C+followed+by+Peru+with+13%25%2C+Colombia+with+10%25%2C+and+with+minor+amounts+in+Venezuela%2C+Ecuador%2C+Bolivia%2C+Guyana%2C+Suriname+and+French+Guiana.+States+or+departments+in+four+nations+contain+%22Amazonas%22+in+their+names.+The+Amazon+represents+over+half+of+the+planet%27s+remaining+rainforests%2C+and+comprises+the+largest+and+most+biodiverse+tract+of+tropical+rainforest+in+the+world%2C+with+an+estimated+390+billion+individual+trees+divided+into+16%2C000+species)
 - [Translation with T5](https://huggingface.co/t5-base?text=My+name+is+Wolfgang+and+I+live+in+Berlin)
 
-In Computer Vision:
-- [Image classification with ViT](https://huggingface.co/google/vit-base-patch16-224)
-- [Object Detection with DETR](https://huggingface.co/facebook/detr-resnet-50)
-- [Image Segmentation with DETR](https://huggingface.co/facebook/detr-resnet-50-panoptic)
-
-In Audio:
-- [Automatic Speech Recognition with Wav2Vec2](https://huggingface.co/facebook/wav2vec2-base-960h)
-- [Keyword Spotting with Wav2Vec2](https://huggingface.co/superb/wav2vec2-base-superb-ks)
-
 **[Write With Transformer](https://transformer.huggingface.co)**, built by the Hugging Face team, is the official demo of this repo’s text generation capabilities.
 
-## If you are looking for custom support from the Hugging Face team
-
-<a target="_blank" href="https://huggingface.co/support">
-    <img alt="HuggingFace Expert Acceleration Program" src="https://cdn-media.huggingface.co/marketing/transformers/new-support-improved.png" style="max-width: 600px; border: 1px solid #eee; border-radius: 4px; box-shadow: 0 1px 2px 0 rgba(0, 0, 0, 0.05);">
-</a><br>
-
 ## Quick tour
 
-To immediately use a model on a given input (text, image, audio, ...), we provide the `pipeline` API. Pipelines group together a pretrained model with the preprocessing that was used during that model's training. Here is how to quickly use a pipeline to classify positive versus negative texts:
+To immediately use a model on a given text, we provide the `pipeline` API. Pipelines group together a pretrained model with the preprocessing that was used during that model training. Here is how to quickly use a pipeline to classify positive versus negative texts
 
 ```python
 >>> from transformers import pipeline
 
 # Allocate a pipeline for sentiment-analysis
 >>> classifier = pipeline('sentiment-analysis')
->>> classifier('We are very happy to introduce pipeline to the transformers repository.')
-[{'label': 'POSITIVE', 'score': 0.9996980428695679}]
+>>> classifier('We are very happy to include pipeline into the transformers repository.')
+[{'label': 'POSITIVE', 'score': 0.9978193640708923}]
 ```
 
-The second line of code downloads and caches the pretrained model used by the pipeline, while the third evaluates it on the given text. Here the answer is "positive" with a confidence of 99.97%.
+The second line of code downloads and caches the pretrained model used by the pipeline, the third line evaluates it on the given text. Here the answer is "positive" with a confidence of 99.8%.
 
-Many tasks have a pre-trained `pipeline` ready to go, in NLP but also in computer vision and speech. For example, we can easily extract detected objects in an image:
+This is another example of pipeline used for that can extract question answers from some context:
 
 ``` python
->>> import requests
->>> from PIL import Image
 >>> from transformers import pipeline
 
-# Download an image with cute cats
->>> url = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/coco_sample.png"
->>> image_data = requests.get(url, stream=True).raw
->>> image = Image.open(image_data)
+# Allocate a pipeline for question-answering
+>>> question_answerer = pipeline('question-answering')
+>>> question_answerer({
+...     'question': 'What is the name of the repository ?',
+...     'context': 'Pipeline have been included in the huggingface/transformers repository'
+... })
+{'score': 0.5135612454720828, 'start': 35, 'end': 59, 'answer': 'huggingface/transformers'}
 
-# Allocate a pipeline for object detection
->>> object_detector = pipeline('object_detection')
->>> object_detector(image)
-[{'score': 0.9982201457023621,
-  'label': 'remote',
-  'box': {'xmin': 40, 'ymin': 70, 'xmax': 175, 'ymax': 117}},
- {'score': 0.9960021376609802,
-  'label': 'remote',
-  'box': {'xmin': 333, 'ymin': 72, 'xmax': 368, 'ymax': 187}},
- {'score': 0.9954745173454285,
-  'label': 'couch',
-  'box': {'xmin': 0, 'ymin': 1, 'xmax': 639, 'ymax': 473}},
- {'score': 0.9988006353378296,
-  'label': 'cat',
-  'box': {'xmin': 13, 'ymin': 52, 'xmax': 314, 'ymax': 470}},
- {'score': 0.9986783862113953,
-  'label': 'cat',
-  'box': {'xmin': 345, 'ymin': 23, 'xmax': 640, 'ymax': 368}}]
 ```
 
-Here we get a list of objects detected in the image, with a box surrounding the object and a confidence score. Here is the original image on the right, with the predictions displayed on the left:
+On top of the answer, the pretrained model used here returned its confidence score, along with the start position and its end position in the tokenized sentence. You can learn more about the tasks supported by the `pipeline` API in [this tutorial](https://huggingface.co/transformers/task_summary.html).
 
-<h3 align="center">
-    <a><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/coco_sample.png" width="400"></a>
-    <a><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/coco_sample_post_processed.png" width="400"></a>
-</h3>
-
-You can learn more about the tasks supported by the `pipeline` API in [this tutorial](https://huggingface.co/docs/transformers/task_summary).
-
-To download and use any of the pretrained models on your given task, all it takes is three lines of code. Here is the PyTorch version:
+To download and use any of the pretrained models on your given task, you just need to use those three lines of codes (PyTorch version):
 ```python
 >>> from transformers import AutoTokenizer, AutoModel
 
@@ -167,8 +91,7 @@ To download and use any of the pretrained models on your given task, all it take
 >>> inputs = tokenizer("Hello world!", return_tensors="pt")
 >>> outputs = model(**inputs)
 ```
-
-And here is the equivalent code for TensorFlow:
+or for TensorFlow:
 ```python
 >>> from transformers import AutoTokenizer, TFAutoModel
 
@@ -179,14 +102,14 @@ And here is the equivalent code for TensorFlow:
 >>> outputs = model(**inputs)
 ```
 
-The tokenizer is responsible for all the preprocessing the pretrained model expects, and can be called directly on a single string (as in the above examples) or a list. It will output a dictionary that you can use in downstream code or simply directly pass to your model using the ** argument unpacking operator.
+The tokenizer is responsible for all the preprocessing the pretrained model expects, and can be called directly on one (or list) of texts (as we can see on the fourth line of both code examples). It will output a dictionary you can directly pass to your model (which is done on the fifth line).
 
-The model itself is a regular [Pytorch `nn.Module`](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) or a [TensorFlow `tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) (depending on your backend) which you can use normally. [This tutorial](https://huggingface.co/docs/transformers/training) explains how to integrate such a model into a classic PyTorch or TensorFlow training loop, or how to use our `Trainer` API to quickly fine-tune on a new dataset.
+The model itself is a regular [Pytorch `nn.Module`](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) or a [TensorFlow `tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) (depending on your backend) which you can use normally. For instance, [this tutorial](https://huggingface.co/transformers/training.html) explains how to integrate such a model in classic PyTorch or TensorFlow training loop, or how to use our `Trainer` API to quickly fine-tune the on a new dataset.
 
 ## Why should I use transformers?
 
 1. Easy-to-use state-of-the-art models:
-    - High performance on natural language understanding & generation, computer vision, and audio tasks.
+    - High performance on NLU and NLG tasks.
     - Low barrier to entry for educators and practitioners.
     - Few user-facing abstractions with just three classes to learn.
     - A unified API for using all our pretrained models.
@@ -194,44 +117,44 @@ The model itself is a regular [Pytorch `nn.Module`](https://pytorch.org/docs/sta
 1. Lower compute costs, smaller carbon footprint:
     - Researchers can share trained models instead of always retraining.
     - Practitioners can reduce compute time and production costs.
-    - Dozens of architectures with over 20,000 pretrained models, some in more than 100 languages.
+    - Dozens of architectures with over 2,000 pretrained models, some in more than 100 languages.
 
 1. Choose the right framework for every part of a model's lifetime:
     - Train state-of-the-art models in 3 lines of code.
-    - Move a single model between TF2.0/PyTorch/JAX frameworks at will.
-    - Seamlessly pick the right framework for training, evaluation and production.
+    - Move a single model between TF2.0/PyTorch frameworks at will.
+    - Seamlessly pick the right framework for training, evaluation, production.
 
 1. Easily customize a model or an example to your needs:
-    - We provide examples for each architecture to reproduce the results published by its original authors.
-    - Model internals are exposed as consistently as possible.
+    - Examples for each architecture to reproduce the results by the official authors of said architecture.
+    - Expose the models internal as consistently as possible.
     - Model files can be used independently of the library for quick experiments.
 
 ## Why shouldn't I use transformers?
 
-- This library is not a modular toolbox of building blocks for neural nets. The code in the model files is not refactored with additional abstractions on purpose, so that researchers can quickly iterate on each of the models without diving into additional abstractions/files.
+- This library is not a modular toolbox of building blocks for neural nets. The code in the model files is not refactored with additional abstractions on purpose, so that researchers can quickly iterate on each of the models without diving in additional abstractions/files.
 - The training API is not intended to work on any model but is optimized to work with the models provided by the library. For generic machine learning loops, you should use another library.
-- While we strive to present as many use cases as possible, the scripts in our [examples folder](https://github.com/huggingface/transformers/tree/main/examples) are just that: examples. It is expected that they won't work out-of-the box on your specific problem and that you will be required to change a few lines of code to adapt them to your needs.
+- While we strive to present as many use cases as possible, the scripts in our [examples folder](https://github.com/huggingface/transformers/tree/master/examples) are just that: examples. It is expected that they won't work out-of-the box on your specific problem and that you will be required to change a few lines of code to adapt them to your needs.
 
 ## Installation
 
 ### With pip
 
-This repository is tested on Python 3.6+, Flax 0.3.2+, PyTorch 1.3.1+ and TensorFlow 2.3+.
+This repository is tested on Python 3.6+, PyTorch 1.0.0+ (PyTorch 1.3.1+ for [examples](https://github.com/huggingface/transformers/tree/master/examples)) and TensorFlow 2.0.
 
 You should install 🤗 Transformers in a [virtual environment](https://docs.python.org/3/library/venv.html). If you're unfamiliar with Python virtual environments, check out the [user guide](https://packaging.python.org/guides/installing-using-pip-and-virtual-environments/).
 
 First, create a virtual environment with the version of Python you're going to use and activate it.
 
-Then, you will need to install at least one of Flax, PyTorch or TensorFlow.
-Please refer to [TensorFlow installation page](https://www.tensorflow.org/install/), [PyTorch installation page](https://pytorch.org/get-started/locally/#start-locally) and/or [Flax](https://github.com/google/flax#quick-install) and [Jax](https://github.com/google/jax#installation) installation pages regarding the specific install command for your platform.
+Then, you will need to install at least one of TensorFlow 2.0, PyTorch or Flax.
+Please refer to [TensorFlow installation page](https://www.tensorflow.org/install/pip#tensorflow-2.0-rc-is-available), [PyTorch installation page](https://pytorch.org/get-started/locally/#start-locally) regarding the specific install command for your platform and/or [Flax installation page](https://github.com/google/flax#quick-install).
 
-When one of those backends has been installed, 🤗 Transformers can be installed using pip as follows:
+When TensorFlow 2.0 and/or PyTorch has been installed, 🤗 Transformers can be installed using pip as follows:
 
 ```bash
 pip install transformers
 ```
 
-If you'd like to play with the examples or need the bleeding edge of the code and can't wait for a new release, you must [install the library from source](https://huggingface.co/docs/transformers/installation#installing-from-source).
+If you'd like to play with the examples, you must [install the library from source](https://huggingface.co/transformers/installation.html#installing-from-source).
 
 ### With conda
 
@@ -243,182 +166,77 @@ Since Transformers version v4.0.0, we now have a conda channel: `huggingface`.
 conda install -c huggingface transformers
 ```
 
-Follow the installation pages of Flax, PyTorch or TensorFlow to see how to install them with conda.
+Follow the installation pages of TensorFlow, PyTorch or Flax to see how to install them with conda. 
 
-## Model architectures
+## Models architectures
 
-**[All the model checkpoints](https://huggingface.co/models)** provided by 🤗 Transformers are seamlessly integrated from the huggingface.co [model hub](https://huggingface.co) where they are uploaded directly by [users](https://huggingface.co/users) and [organizations](https://huggingface.co/organizations).
+🤗 Transformers currently provides the following architectures (see [here](https://huggingface.co/transformers/model_summary.html) for a high-level summary of each them):
 
-Current number of checkpoints: ![](https://img.shields.io/endpoint?url=https://huggingface.co/api/shields/models&color=brightgreen)
-
-🤗 Transformers currently provides the following architectures (see [here](https://huggingface.co/docs/transformers/model_summary) for a high-level summary of each them):
-
-1. **[ALBERT](https://huggingface.co/docs/transformers/model_doc/albert)** (from Google Research and the Toyota Technological Institute at Chicago) released with the paper [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://arxiv.org/abs/1909.11942), by Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut.
-1. **[BART](https://huggingface.co/docs/transformers/model_doc/bart)** (from Facebook) released with the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/abs/1910.13461) by Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Ves Stoyanov and Luke Zettlemoyer.
-1. **[BARThez](https://huggingface.co/docs/transformers/model_doc/barthez)** (from École polytechnique) released with the paper [BARThez: a Skilled Pretrained French Sequence-to-Sequence Model](https://arxiv.org/abs/2010.12321) by Moussa Kamal Eddine, Antoine J.-P. Tixier, Michalis Vazirgiannis.
-1. **[BARTpho](https://huggingface.co/docs/transformers/model_doc/bartpho)** (from VinAI Research) released with the paper [BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese](https://arxiv.org/abs/2109.09701) by Nguyen Luong Tran, Duong Minh Le and Dat Quoc Nguyen.
-1. **[BEiT](https://huggingface.co/docs/transformers/model_doc/beit)** (from Microsoft) released with the paper [BEiT: BERT Pre-Training of Image Transformers](https://arxiv.org/abs/2106.08254) by Hangbo Bao, Li Dong, Furu Wei.
-1. **[BERT](https://huggingface.co/docs/transformers/model_doc/bert)** (from Google) released with the paper [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://arxiv.org/abs/1810.04805) by Jacob Devlin, Ming-Wei Chang, Kenton Lee and Kristina Toutanova.
-1. **[BERT For Sequence Generation](https://huggingface.co/docs/transformers/model_doc/bert-generation)** (from Google) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
-1. **[BERTweet](https://huggingface.co/docs/transformers/model_doc/bertweet)** (from VinAI Research) released with the paper [BERTweet: A pre-trained language model for English Tweets](https://aclanthology.org/2020.emnlp-demos.2/) by Dat Quoc Nguyen, Thanh Vu and Anh Tuan Nguyen.
-1. **[BigBird-Pegasus](https://huggingface.co/docs/transformers/model_doc/bigbird_pegasus)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
-1. **[BigBird-RoBERTa](https://huggingface.co/docs/transformers/model_doc/big_bird)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
-1. **[Blenderbot](https://huggingface.co/docs/transformers/model_doc/blenderbot)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
-1. **[BlenderbotSmall](https://huggingface.co/docs/transformers/model_doc/blenderbot-small)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
-1. **[BLOOM](https://huggingface.co/docs/transformers/model_doc/bloom)** (from BigScience workshop) released by the [BigSicence Workshop](https://bigscience.huggingface.co/).
-1. **[BORT](https://huggingface.co/docs/transformers/model_doc/bort)** (from Alexa) released with the paper [Optimal Subarchitecture Extraction For BERT](https://arxiv.org/abs/2010.10499) by Adrian de Wynter and Daniel J. Perry.
-1. **[ByT5](https://huggingface.co/docs/transformers/model_doc/byt5)** (from Google Research) released with the paper [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) by Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel.
-1. **[CamemBERT](https://huggingface.co/docs/transformers/model_doc/camembert)** (from Inria/Facebook/Sorbonne) released with the paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) by Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suárez*, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah and Benoît Sagot.
-1. **[CANINE](https://huggingface.co/docs/transformers/model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
-1. **[CLIP](https://huggingface.co/docs/transformers/model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
-1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
-1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
-1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
-1. **[CPM](https://huggingface.co/docs/transformers/model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
-1. **[CTRL](https://huggingface.co/docs/transformers/model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
-1. **[CvT](https://huggingface.co/docs/transformers/model_doc/cvt)** (from Microsoft) released with the paper [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) by Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang.
-1. **[Data2Vec](https://huggingface.co/docs/transformers/model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
-1. **[DeBERTa](https://huggingface.co/docs/transformers/model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
-1. **[DeBERTa-v2](https://huggingface.co/docs/transformers/model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
-1. **[Decision Transformer](https://huggingface.co/docs/transformers/model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
-1. **[DeiT](https://huggingface.co/docs/transformers/model_doc/deit)** (from Facebook) released with the paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) by Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou.
-1. **[DETR](https://huggingface.co/docs/transformers/model_doc/detr)** (from Facebook) released with the paper [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) by Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko.
-1. **[DialoGPT](https://huggingface.co/docs/transformers/model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
-1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation) and a German version of DistilBERT.
-1. **[DiT](https://huggingface.co/docs/transformers/model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
-1. **[DPR](https://huggingface.co/docs/transformers/model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
-1. **[DPT](https://huggingface.co/docs/transformers/master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by René Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
-1. **[ELECTRA](https://huggingface.co/docs/transformers/model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
-1. **[EncoderDecoder](https://huggingface.co/docs/transformers/model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
-1. **[FlauBERT](https://huggingface.co/docs/transformers/model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab.
-1. **[FLAVA](https://huggingface.co/docs/transformers/model_doc/flava)** (from Facebook AI) released with the paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) by Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela.
-1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
-1. **[Funnel Transformer](https://huggingface.co/docs/transformers/model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
-1. **[GLPN](https://huggingface.co/docs/transformers/model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
-1. **[GPT](https://huggingface.co/docs/transformers/model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
-1. **[GPT Neo](https://huggingface.co/docs/transformers/model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
-1. **[GPT NeoX](https://huggingface.co/docs/transformers/model_doc/gpt_neox)** (from EleutherAI) released with the paper [GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://arxiv.org/abs/2204.06745) by Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, Samuel Weinbach
-1. **[GPT-2](https://huggingface.co/docs/transformers/model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
-1. **[GPT-J](https://huggingface.co/docs/transformers/model_doc/gptj)** (from EleutherAI) released in the repository [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
-1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
-1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (from Facebook) released with the paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) by Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
-1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (from Berkeley) released with the paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) by Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
-1. **[ImageGPT](https://huggingface.co/docs/transformers/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
-1. **[LayoutLM](https://huggingface.co/docs/transformers/model_doc/layoutlm)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
-1. **[LayoutLMv2](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) by Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
-1. **[LayoutLMv3](https://huggingface.co/docs/transformers/model_doc/layoutlmv3)** (from Microsoft Research Asia) released with the paper [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) by Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei.
-1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
-1. **[LED](https://huggingface.co/docs/transformers/model_doc/led)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
-1. **[LeViT](https://huggingface.co/docs/transformers/model_doc/levit)** (from Meta AI) released with the paper [LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference](https://arxiv.org/abs/2104.01136) by Ben Graham, Alaaeldin El-Nouby, Hugo Touvron, Pierre Stock, Armand Joulin, Hervé Jégou, Matthijs Douze.
-1. **[Longformer](https://huggingface.co/docs/transformers/model_doc/longformer)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
-1. **[LongT5](https://huggingface.co/docs/transformers/model_doc/longt5)** (from Google AI) released with the paper [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://arxiv.org/abs/2112.07916) by Mandy Guo, Joshua Ainslie, David Uthus, Santiago Ontanon, Jianmo Ni, Yun-Hsuan Sung, Yinfei Yang.
-1. **[LUKE](https://huggingface.co/docs/transformers/model_doc/luke)** (from Studio Ousia) released with the paper [LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention](https://arxiv.org/abs/2010.01057) by Ikuya Yamada, Akari Asai, Hiroyuki Shindo, Hideaki Takeda, Yuji Matsumoto.
-1. **[LXMERT](https://huggingface.co/docs/transformers/model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
-1. **[M-CTC-T](https://huggingface.co/docs/transformers/model_doc/mctct)** (from Facebook) released with the paper [Pseudo-Labeling For Massively Multilingual Speech Recognition](https://arxiv.org/abs/2111.00161) by Loren Lugosch, Tatiana Likhomanenko, Gabriel Synnaeve, and Ronan Collobert.
-1. **[M2M100](https://huggingface.co/docs/transformers/model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
-1. **[MarianMT](https://huggingface.co/docs/transformers/model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jörg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
-1. **[MaskFormer](https://huggingface.co/docs/transformers/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
-1. **[mBART](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
-1. **[mBART-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
-1. **[Megatron-BERT](https://huggingface.co/docs/transformers/model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
-1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
-1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
-1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
-1. **[MobileViT](https://huggingface.co/docs/transformers/model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
-1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
-1. **[MT5](https://huggingface.co/docs/transformers/model_doc/mt5)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
-1. **[MVP](https://huggingface.co/docs/transformers/model_doc/mvp)** (from RUC AI Box) released with the paper [MVP: Multi-task Supervised Pre-training for Natural Language Generation](https://arxiv.org/abs/2206.12131) by Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen.
-1. **[Nezha](https://huggingface.co/docs/transformers/model_doc/nezha)** (from Huawei Noah’s Ark Lab) released with the paper [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://arxiv.org/abs/1909.00204) by Junqiu Wei, Xiaozhe Ren, Xiaoguang Li, Wenyong Huang, Yi Liao, Yasheng Wang, Jiashu Lin, Xin Jiang, Xiao Chen and Qun Liu.
-1. **[NLLB](https://huggingface.co/docs/transformers/model_doc/nllb)** (from Meta) released with the paper [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) by the NLLB team.
-1. **[Nyströmformer](https://huggingface.co/docs/transformers/model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
-1. **[OPT](https://huggingface.co/docs/transformers/master/model_doc/opt)** (from Meta AI) released with the paper [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) by Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al.
-1. **[OWL-ViT](https://huggingface.co/docs/transformers/model_doc/owlvit)** (from Google AI) released with the paper [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) by Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby.
-1. **[Pegasus](https://huggingface.co/docs/transformers/model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
-1. **[Perceiver IO](https://huggingface.co/docs/transformers/model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira.
-1. **[PhoBERT](https://huggingface.co/docs/transformers/model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
-1. **[PLBart](https://huggingface.co/docs/transformers/model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
-1. **[PoolFormer](https://huggingface.co/docs/transformers/model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
-1. **[ProphetNet](https://huggingface.co/docs/transformers/model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
-1. **[QDQBert](https://huggingface.co/docs/transformers/model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
-1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
-1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
-1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
-1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Platforms) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
-1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
-1. **[ResNet](https://huggingface.co/docs/transformers/model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun.
-1. **[RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta)** (from Facebook), released together with the paper [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
-1. **[RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/abs/2104.09864) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
-1. **[SegFormer](https://huggingface.co/docs/transformers/model_doc/segformer)** (from NVIDIA) released with the paper [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) by Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo.
-1. **[SEW](https://huggingface.co/docs/transformers/model_doc/sew)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
-1. **[SEW-D](https://huggingface.co/docs/transformers/model_doc/sew_d)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
-1. **[SpeechToTextTransformer](https://huggingface.co/docs/transformers/model_doc/speech_to_text)** (from Facebook), released together with the paper [fairseq S2T: Fast Speech-to-Text Modeling with fairseq](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Dmytro Okhonko, Juan Pino.
-1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (from Facebook), released together with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
-1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (from Tel Aviv University), released together with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
-1. **[SqueezeBERT](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
-1. **[Swin Transformer](https://huggingface.co/docs/transformers/model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
-1. **[Swin Transformer V2](https://huggingface.co/docs/transformers/main/model_doc/swinv2)** (from Microsoft) released with the paper [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) by Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo.
-1. **[T5](https://huggingface.co/docs/transformers/model_doc/t5)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
-1. **[T5v1.1](https://huggingface.co/docs/transformers/model_doc/t5v1.1)** (from Google AI) released in the repository [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
-1. **[TAPAS](https://huggingface.co/docs/transformers/model_doc/tapas)** (from Google AI) released with the paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) by Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos.
-1. **[TAPEX](https://huggingface.co/docs/transformers/model_doc/tapex)** (from Microsoft Research) released with the paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) by Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
-1. **[Trajectory Transformer](https://huggingface.co/docs/transformers/model_doc/trajectory_transformers)** (from the University of California at Berkeley) released with the paper [Offline Reinforcement Learning as One Big Sequence Modeling Problem](https://arxiv.org/abs/2106.02039) by Michael Janner, Qiyang Li, Sergey Levine
-1. **[Transformer-XL](https://huggingface.co/docs/transformers/model_doc/transfo-xl)** (from Google/CMU) released with the paper [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) by Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov.
-1. **[TrOCR](https://huggingface.co/docs/transformers/model_doc/trocr)** (from Microsoft), released together with the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
-1. **[UL2](https://huggingface.co/docs/transformers/model_doc/ul2)** (from Google Research) released with the paper [Unifying Language Learning Paradigms](https://arxiv.org/abs/2205.05131v1) by Yi Tay, Mostafa Dehghani, Vinh Q. Tran, Xavier Garcia, Dara Bahri, Tal Schuster, Huaixiu Steven Zheng, Neil Houlsby, Donald Metzler
-1. **[UniSpeech](https://huggingface.co/docs/transformers/model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
-1. **[UniSpeechSat](https://huggingface.co/docs/transformers/model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
-1. **[VAN](https://huggingface.co/docs/transformers/model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/abs/2202.09741) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
-1. **[VideoMAE](https://huggingface.co/docs/transformers/main/model_doc/videomae)** (from Multimedia Computing Group, Nanjing University) released with the paper [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://arxiv.org/abs/2203.12602) by Zhan Tong, Yibing Song, Jue Wang, Limin Wang.
-1. **[ViLT](https://huggingface.co/docs/transformers/model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
-1. **[Vision Transformer (ViT)](https://huggingface.co/docs/transformers/model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby.
-1. **[VisualBERT](https://huggingface.co/docs/transformers/model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
-1. **[ViTMAE](https://huggingface.co/docs/transformers/model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
-1. **[Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
-1. **[Wav2Vec2-Conformer](https://huggingface.co/docs/transformers/model_doc/wav2vec2-conformer)** (from Facebook AI) released with the paper [FAIRSEQ S2T: Fast Speech-to-Text Modeling with FAIRSEQ](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Sravya Popuri, Dmytro Okhonko, Juan Pino.
-1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/transformers/model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
-1. **[WavLM](https://huggingface.co/docs/transformers/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
-1. **[XGLM](https://huggingface.co/docs/transformers/model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li.
-1. **[XLM](https://huggingface.co/docs/transformers/model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
-1. **[XLM-ProphetNet](https://huggingface.co/docs/transformers/model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
-1. **[XLM-RoBERTa](https://huggingface.co/docs/transformers/model_doc/xlm-roberta)** (from Facebook AI), released together with the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov.
-1. **[XLM-RoBERTa-XL](https://huggingface.co/docs/transformers/model_doc/xlm-roberta-xl)** (from Facebook AI), released together with the paper [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) by Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau.
-1. **[XLNet](https://huggingface.co/docs/transformers/model_doc/xlnet)** (from Google/CMU) released with the paper [​XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
-1. **[XLS-R](https://huggingface.co/docs/transformers/model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
-1. **[XLSR-Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
-1. **[YOLOS](https://huggingface.co/docs/transformers/model_doc/yolos)** (from Huazhong University of Science & Technology) released with the paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) by Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
-1. **[YOSO](https://huggingface.co/docs/transformers/model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) Once: Linear Cost Self-Attention Via Bernoulli Sampling](https://arxiv.org/abs/2111.09714) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
+1. **[ALBERT](https://huggingface.co/transformers/model_doc/albert.html)** (from Google Research and the Toyota Technological Institute at Chicago) released with the paper [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://arxiv.org/abs/1909.11942), by Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut.
+1. **[BART](https://huggingface.co/transformers/model_doc/bart.html)** (from Facebook) released with the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/pdf/1910.13461.pdf) by Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Ves Stoyanov and Luke Zettlemoyer.
+1. **[BERT](https://huggingface.co/transformers/model_doc/bert.html)** (from Google) released with the paper [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://arxiv.org/abs/1810.04805) by Jacob Devlin, Ming-Wei Chang, Kenton Lee and Kristina Toutanova.
+1. **[BERT For Sequence Generation](https://huggingface.co/transformers/model_doc/bertgeneration.html)** (from Google) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
+1. **[Blenderbot](https://huggingface.co/transformers/model_doc/blenderbot.html)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
+1. **[CamemBERT](https://huggingface.co/transformers/model_doc/camembert.html)** (from Inria/Facebook/Sorbonne) released with the paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) by Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suárez*, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah and Benoît Sagot.
+1. **[CTRL](https://huggingface.co/transformers/model_doc/ctrl.html)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
+1. **[DeBERTa](https://huggingface.co/transformers/model_doc/deberta.html)** (from Microsoft Research) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
+1. **[DialoGPT](https://huggingface.co/transformers/model_doc/dialogpt.html)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
+1. **[DistilBERT](https://huggingface.co/transformers/model_doc/distilbert.html)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/master/examples/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/master/examples/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/master/examples/distillation) and a German version of DistilBERT.
+1. **[DPR](https://huggingface.co/transformers/model_doc/dpr.html)** (from Facebook) released with the paper [Dense Passage Retrieval
+for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon
+Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
+1. **[ELECTRA](https://huggingface.co/transformers/model_doc/electra.html)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
+1. **[FlauBERT](https://huggingface.co/transformers/model_doc/flaubert.html)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab.
+1. **[Funnel Transformer](https://huggingface.co/transformers/model_doc/funnel.html)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
+1. **[GPT](https://huggingface.co/transformers/model_doc/gpt.html)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
+1. **[GPT-2](https://huggingface.co/transformers/model_doc/gpt2.html)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
+1. **[LayoutLM](https://huggingface.co/transformers/model_doc/layoutlm.html)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
+1. **[Longformer](https://huggingface.co/transformers/model_doc/longformer.html)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
+1. **[LXMERT](https://huggingface.co/transformers/model_doc/lxmert.html)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
+1. **[MarianMT](https://huggingface.co/transformers/model_doc/marian.html)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jörg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
+1. **[MBart](https://huggingface.co/transformers/model_doc/mbart.html)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
+1. **[MT5](https://huggingface.co/transformers/model_doc/mt5.html)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
+1. **[Pegasus](https://huggingface.co/transformers/model_doc/pegasus.html)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777)> by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
+1. **[ProphetNet](https://huggingface.co/transformers/model_doc/prophetnet.html)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
+1. **[Reformer](https://huggingface.co/transformers/model_doc/reformer.html)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
+1. **[RoBERTa](https://huggingface.co/transformers/model_doc/roberta.html)** (from Facebook), released together with the paper a [Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
+ultilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/master/examples/distillation) and a German version of DistilBERT.
+1. **[SqueezeBert](https://huggingface.co/transformers/model_doc/squeezebert.html)** released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
+1. **[T5](https://huggingface.co/transformers/model_doc/t5.html)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
+1. **[Transformer-XL](https://huggingface.co/transformers/model_doc/transformerxl.html)** (from Google/CMU) released with the paper [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) by Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov.
+1. **[XLM](https://huggingface.co/transformers/model_doc/xlm.html)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
+1. **[XLM-ProphetNet](https://huggingface.co/transformers/model_doc/xlmprophetnet.html)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
+1. **[XLM-RoBERTa](https://huggingface.co/transformers/model_doc/xlmroberta.html)** (from Facebook AI), released together with the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov.
+1. **[XLNet](https://huggingface.co/transformers/model_doc/xlnet.html)** (from Google/CMU) released with the paper [​XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
+1. **[Other community models](https://huggingface.co/models)**, contributed by the [community](https://huggingface.co/users).
 1. Want to contribute a new model? We have added a **detailed guide and templates** to guide you in the process of adding a new model. You can find them in the [`templates`](./templates) folder of the repository. Be sure to check the [contributing guidelines](./CONTRIBUTING.md) and contact the maintainers or open an issue to collect feedbacks before starting your PR.
 
-To check if each model has an implementation in Flax, PyTorch or TensorFlow, or has an associated tokenizer backed by the 🤗 Tokenizers library, refer to [this table](https://huggingface.co/docs/transformers/index#supported-frameworks).
+To cehck if each model has an implementation in PyTorch/TensorFlow/Flax or has an associated tokenizer backed by the 🤗 Tokenizers library, refer to [this table](https://huggingface.co/transformers/index.html#bigtable)
 
-These implementations have been tested on several datasets (see the example scripts) and should match the performance of the original implementations. You can find more details on performance in the Examples section of the [documentation](https://huggingface.co/docs/transformers/examples).
+These implementations have been tested on several datasets (see the example scripts) and should match the performances of the original implementations. You can find more details on the performances in the Examples section of the [documentation](https://huggingface.co/transformers/examples.html).
 
 
 ## Learn more
 
 | Section | Description |
 |-|-|
-| [Documentation](https://huggingface.co/docs/transformers/) | Full API documentation and tutorials |
-| [Task summary](https://huggingface.co/docs/transformers/task_summary) | Tasks supported by 🤗 Transformers |
-| [Preprocessing tutorial](https://huggingface.co/docs/transformers/preprocessing) | Using the `Tokenizer` class to prepare data for the models |
-| [Training and fine-tuning](https://huggingface.co/docs/transformers/training) | Using the models provided by 🤗 Transformers in a PyTorch/TensorFlow training loop and the `Trainer` API |
-| [Quick tour: Fine-tuning/usage scripts](https://github.com/huggingface/transformers/tree/main/examples) | Example scripts for fine-tuning models on a wide range of tasks |
-| [Model sharing and uploading](https://huggingface.co/docs/transformers/model_sharing) | Upload and share your fine-tuned models with the community |
-| [Migration](https://huggingface.co/docs/transformers/migration) | Migrate to 🤗 Transformers from `pytorch-transformers` or `pytorch-pretrained-bert` |
+| [Documentation](https://huggingface.co/transformers/) | Full API documentation and tutorials |
+| [Task summary](https://huggingface.co/transformers/task_summary.html) | Tasks supported by 🤗 Transformers |
+| [Preprocessing tutorial](https://huggingface.co/transformers/preprocessing.html) | Using the `Tokenizer` class to prepare data for the models |
+| [Training and fine-tuning](https://huggingface.co/transformers/training.html) | Using the models provided by 🤗 Transformers in a PyTorch/TensorFlow training loop and the `Trainer` API |
+| [Quick tour: Fine-tuning/usage scripts](https://github.com/huggingface/transformers/tree/master/examples) | Example scripts for fine-tuning models on a wide range of tasks |
+| [Model sharing and uploading](https://huggingface.co/transformers/model_sharing.html) | Upload and share your fine-tuned models with the community |
+| [Migration](https://huggingface.co/transformers/migration.html) | Migrate to 🤗 Transformers from `pytorch-transformers` or `pytorch-pretrained-bert` |
 
 ## Citation
 
-We now have a [paper](https://www.aclweb.org/anthology/2020.emnlp-demos.6/) you can cite for the 🤗 Transformers library:
+We now have a [paper](https://arxiv.org/abs/1910.03771) you can cite for the 🤗 Transformers library:
 ```bibtex
-@inproceedings{wolf-etal-2020-transformers,
-    title = "Transformers: State-of-the-Art Natural Language Processing",
-    author = "Thomas Wolf and Lysandre Debut and Victor Sanh and Julien Chaumond and Clement Delangue and Anthony Moi and Pierric Cistac and Tim Rault and Rémi Louf and Morgan Funtowicz and Joe Davison and Sam Shleifer and Patrick von Platen and Clara Ma and Yacine Jernite and Julien Plu and Canwen Xu and Teven Le Scao and Sylvain Gugger and Mariama Drame and Quentin Lhoest and Alexander M. Rush",
-    booktitle = "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations",
-    month = oct,
-    year = "2020",
-    address = "Online",
-    publisher = "Association for Computational Linguistics",
-    url = "https://www.aclweb.org/anthology/2020.emnlp-demos.6",
-    pages = "38--45"
+@article{Wolf2019HuggingFacesTS,
+  title={HuggingFace's Transformers: State-of-the-art Natural Language Processing},
+  author={Thomas Wolf and Lysandre Debut and Victor Sanh and Julien Chaumond and Clement Delangue and Anthony Moi and Pierric Cistac and Tim Rault and Rémi Louf and Morgan Funtowicz and Joe Davison and Sam Shleifer and Patrick von Platen and Clara Ma and Yacine Jernite and Julien Plu and Canwen Xu and Teven Le Scao and Sylvain Gugger and Mariama Drame and Quentin Lhoest and Alexander M. Rush},
+  journal={ArXiv},
+  year={2019},
+  volume={abs/1910.03771}
 }
 ```

README_ko.md

@@ -1,379 +0,0 @@
-<!---
-Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
--->
-
-<p align="center">
-    <br>
-    <img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers_logo_name.png" width="400"/>
-    <br>
-<p>
-<p align="center">
-    <a href="https://circleci.com/gh/huggingface/transformers">
-        <img alt="Build" src="https://img.shields.io/circleci/build/github/huggingface/transformers/main">
-    </a>
-    <a href="https://github.com/huggingface/transformers/blob/main/LICENSE">
-        <img alt="GitHub" src="https://img.shields.io/github/license/huggingface/transformers.svg?color=blue">
-    </a>
-    <a href="https://huggingface.co/docs/transformers/index">
-        <img alt="Documentation" src="https://img.shields.io/website/http/huggingface.co/docs/transformers/index.svg?down_color=red&down_message=offline&up_message=online">
-    </a>
-    <a href="https://github.com/huggingface/transformers/releases">
-        <img alt="GitHub release" src="https://img.shields.io/github/release/huggingface/transformers.svg">
-    </a>
-    <a href="https://github.com/huggingface/transformers/blob/main/CODE_OF_CONDUCT.md">
-        <img alt="Contributor Covenant" src="https://img.shields.io/badge/Contributor%20Covenant-v2.0%20adopted-ff69b4.svg">
-    </a>
-    <a href="https://zenodo.org/badge/latestdoi/155220641"><img src="https://zenodo.org/badge/155220641.svg" alt="DOI"></a>
-</p>
-
-<h4 align="center">
-    <p>
-        <a href="https://github.com/huggingface/transformers/">English</a> |
-        <a href="https://github.com/huggingface/transformers/blob/main/README_zh-hans.md">简体中文</a> |
-        <a href="https://github.com/huggingface/transformers/blob/main/README_zh-hant.md">繁體中文</a> |
-        <b>한국어</b>
-    <p>
-</h4>
-
-<h3 align="center">
-    <p> Jax, Pytorch, TensorFlow를 위한 최첨단 자연어처리</p>
-</h3>
-
-<h3 align="center">
-    <a href="https://hf.co/course"><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/course_banner.png"></a>
-</h3>
-
-🤗 Transformers는 분류, 정보 추출, 질문 답변, 요약, 번역, 문장 생성 등을 100개 이상의 언어로 수행할 수 있는 수천개의 사전학습된 모델을 제공합니다. 우리의 목표는 모두가 최첨단의 NLP 기술을 쉽게 사용하는 것입니다.
-
-🤗 Transformers는 이러한 사전학습 모델을 빠르게 다운로드해 특정 텍스트에 사용하고, 원하는 데이터로 fine-tuning해 커뮤니티나 우리의 [모델 허브](https://huggingface.co/models)에 공유할 수 있도록 API를 제공합니다. 또한, 모델 구조를 정의하는 각 파이썬 모듈은 완전히 독립적이여서 연구 실험을 위해 손쉽게 수정할 수 있습니다.
-
-🤗 Transformers는 가장 유명한 3개의 딥러닝 라이브러리를 지원합니다. 이들은 서로 완벽히 연동됩니다 — [Jax](https://jax.readthedocs.io/en/latest/), [PyTorch](https://pytorch.org/), [TensorFlow](https://www.tensorflow.org/). 간단하게 이 라이브러리 중 하나로 모델을 학습하고, 또 다른 라이브러리로 추론을 위해 모델을 불러올 수 있습니다. 
-
-## 온라인 데모
-
-대부분의 모델을 [모델 허브](https://huggingface.co/models) 페이지에서 바로 테스트해볼 수 있습니다. 공개 및 비공개 모델을 위한 [비공개 모델 호스팅, 버전 관리, 추론 API](https://huggingface.co/pricing)도 제공합니다.
-
-예시:
-- [BERT로 마스킹된 단어 완성하기](https://huggingface.co/bert-base-uncased?text=Paris+is+the+%5BMASK%5D+of+France)
-- [Electra를 이용한 개체명 인식](https://huggingface.co/dbmdz/electra-large-discriminator-finetuned-conll03-english?text=My+name+is+Sarah+and+I+live+in+London+city)
-- [GPT-2로 텍스트 생성하기](https://huggingface.co/gpt2?text=A+long+time+ago%2C+)
-- [RoBERTa로 자연어 추론하기](https://huggingface.co/roberta-large-mnli?text=The+dog+was+lost.+Nobody+lost+any+animal)
-- [BART를 이용한 요약](https://huggingface.co/facebook/bart-large-cnn?text=The+tower+is+324+metres+%281%2C063+ft%29+tall%2C+about+the+same+height+as+an+81-storey+building%2C+and+the+tallest+structure+in+Paris.+Its+base+is+square%2C+measuring+125+metres+%28410+ft%29+on+each+side.+During+its+construction%2C+the+Eiffel+Tower+surpassed+the+Washington+Monument+to+become+the+tallest+man-made+structure+in+the+world%2C+a+title+it+held+for+41+years+until+the+Chrysler+Building+in+New+York+City+was+finished+in+1930.+It+was+the+first+structure+to+reach+a+height+of+300+metres.+Due+to+the+addition+of+a+broadcasting+aerial+at+the+top+of+the+tower+in+1957%2C+it+is+now+taller+than+the+Chrysler+Building+by+5.2+metres+%2817+ft%29.+Excluding+transmitters%2C+the+Eiffel+Tower+is+the+second+tallest+free-standing+structure+in+France+after+the+Millau+Viaduct)
-- [DistilBERT를 이용한 질문 답변](https://huggingface.co/distilbert-base-uncased-distilled-squad?text=Which+name+is+also+used+to+describe+the+Amazon+rainforest+in+English%3F&context=The+Amazon+rainforest+%28Portuguese%3A+Floresta+Amaz%C3%B4nica+or+Amaz%C3%B4nia%3B+Spanish%3A+Selva+Amaz%C3%B3nica%2C+Amazon%C3%ADa+or+usually+Amazonia%3B+French%3A+For%C3%AAt+amazonienne%3B+Dutch%3A+Amazoneregenwoud%29%2C+also+known+in+English+as+Amazonia+or+the+Amazon+Jungle%2C+is+a+moist+broadleaf+forest+that+covers+most+of+the+Amazon+basin+of+South+America.+This+basin+encompasses+7%2C000%2C000+square+kilometres+%282%2C700%2C000+sq+mi%29%2C+of+which+5%2C500%2C000+square+kilometres+%282%2C100%2C000+sq+mi%29+are+covered+by+the+rainforest.+This+region+includes+territory+belonging+to+nine+nations.+The+majority+of+the+forest+is+contained+within+Brazil%2C+with+60%25+of+the+rainforest%2C+followed+by+Peru+with+13%25%2C+Colombia+with+10%25%2C+and+with+minor+amounts+in+Venezuela%2C+Ecuador%2C+Bolivia%2C+Guyana%2C+Suriname+and+French+Guiana.+States+or+departments+in+four+nations+contain+%22Amazonas%22+in+their+names.+The+Amazon+represents+over+half+of+the+planet%27s+remaining+rainforests%2C+and+comprises+the+largest+and+most+biodiverse+tract+of+tropical+rainforest+in+the+world%2C+with+an+estimated+390+billion+individual+trees+divided+into+16%2C000+species)
-- [T5로 번역하기](https://huggingface.co/t5-base?text=My+name+is+Wolfgang+and+I+live+in+Berlin)
-
-**[Transformer와 글쓰기](https://transformer.huggingface.co)** 는 이 저장소의 텍스트 생성 능력에 관한 Hugging Face 팀의 공식 데모입니다. 
-
-## Hugging Face 팀의 커스텀 지원을 원한다면
-
-<a target="_blank" href="https://huggingface.co/support">
-    <img alt="HuggingFace Expert Acceleration Program" src="https://huggingface.co/front/thumbnails/support.png" style="max-width: 600px; border: 1px solid #eee; border-radius: 4px; box-shadow: 0 1px 2px 0 rgba(0, 0, 0, 0.05);">
-</a><br>
-
-## 퀵 투어
-
-원하는 텍스트에 바로 모델을 사용할 수 있도록, 우리는 `pipeline` API를 제공합니다. Pipeline은 사전학습 모델과 그 모델을 학습할 때 적용한 전처리 방식을 하나로 합칩니다. 다음은 긍정적인 텍스트와 부정적인 텍스트를 분류하기 위해 pipeline을 사용한 간단한 예시입니다:
-
-```python
->>> from transformers import pipeline
-
-# Allocate a pipeline for sentiment-analysis
->>> classifier = pipeline('sentiment-analysis')
->>> classifier('We are very happy to introduce pipeline to the transformers repository.')
-[{'label': 'POSITIVE', 'score': 0.9996980428695679}]
-```
-
-코드의 두번째 줄은 pipeline이 사용하는 사전학습 모델을 다운로드하고 캐시로 저장합니다. 세번째 줄에선 그 모델이 주어진 텍스트를 평가합니다. 여기서 모델은 99.97%의 확률로 텍스트가 긍정적이라고 평가했습니다.
-
-많은 NLP 과제들을 `pipeline`으로 바로 수행할 수 있습니다. 예를 들어, 질문과 문맥이 주어지면 손쉽게 답변을 추출할 수 있습니다:
-
-``` python
->>> from transformers import pipeline
-
-# Allocate a pipeline for question-answering
->>> question_answerer = pipeline('question-answering')
->>> question_answerer({
-...     'question': 'What is the name of the repository ?',
-...     'context': 'Pipeline has been included in the huggingface/transformers repository'
-... })
-{'score': 0.30970096588134766, 'start': 34, 'end': 58, 'answer': 'huggingface/transformers'}
-
-```
-
-답변뿐만 아니라, 여기에 사용된 사전학습 모델은 확신도와 토크나이즈된 문장 속 답변의 시작점, 끝점까지 반환합니다. [이 튜토리얼](https://huggingface.co/docs/transformers/task_summary)에서 `pipeline` API가 지원하는 다양한 과제를 확인할 수 있습니다.
-
-코드 3줄로 원하는 과제에 맞게 사전학습 모델을 다운로드 받고 사용할 수 있습니다. 다음은 PyTorch 버전입니다:
-```python
->>> from transformers import AutoTokenizer, AutoModel
-
->>> tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
->>> model = AutoModel.from_pretrained("bert-base-uncased")
-
->>> inputs = tokenizer("Hello world!", return_tensors="pt")
->>> outputs = model(**inputs)
-```
-다음은 TensorFlow 버전입니다:
-```python
->>> from transformers import AutoTokenizer, TFAutoModel
-
->>> tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
->>> model = TFAutoModel.from_pretrained("bert-base-uncased")
-
->>> inputs = tokenizer("Hello world!", return_tensors="tf")
->>> outputs = model(**inputs)
-```
-
-토크나이저는 사전학습 모델의 모든 전처리를 책임집니다. 그리고 (위의 예시처럼) 1개의 스트링이나 리스트도 처리할 수 있습니다. 토크나이저는 딕셔너리를 반환하는데, 이는 다운스트림 코드에 사용하거나 언패킹 연산자 ** 를 이용해 모델에 바로 전달할 수도 있습니다.
-
-모델 자체는 일반적으로 사용되는 [Pytorch `nn.Module`](https://pytorch.org/docs/stable/nn.html#torch.nn.Module)나 [TensorFlow `tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model)입니다. [이 튜토리얼](https://huggingface.co/transformers/training.html)은 이러한 모델을 표준적인 PyTorch나 TensorFlow 학습 과정에서 사용하는 방법, 또는 새로운 데이터로 fine-tune하기 위해 `Trainer` API를 사용하는 방법을 설명해줍니다.
-
-## 왜 transformers를 사용해야 할까요?
-
-1. 손쉽게 사용할 수 있는 최첨단 모델:
-    - NLU와 NLG 과제에서 뛰어난 성능을 보입니다.
-    - 교육자 실무자에게 진입 장벽이 낮습니다.
-    - 3개의 클래스만 배우면 바로 사용할 수 있습니다.
-    - 하나의 API로 모든 사전학습 모델을 사용할 수 있습니다.
-
-1. 더 적은 계산 비용, 더 적은 탄소 발자국:
-    - 연구자들은 모델을 계속 다시 학습시키는 대신 학습된 모델을 공유할 수 있습니다.
-    - 실무자들은 학습에 필요한 시간과 비용을 절약할 수 있습니다.
-    - 수십개의 모델 구조, 2,000개 이상의 사전학습 모델, 100개 이상의 언어로 학습된 모델 등.
-
-1. 모델의 각 생애주기에 적합한 프레임워크:
-    - 코드 3줄로 최첨단 모델을 학습하세요.
-    - 자유롭게 모델을 TF2.0나 PyTorch 프레임워크로 변환하세요.
-    - 학습, 평가, 공개 등 각 단계에 맞는 프레임워크를 원하는대로 선택하세요.
-
-1. 필요한 대로 모델이나 예시를 커스터마이즈하세요:
-    - 우리는 저자가 공개한 결과를 재현하기 위해 각 모델 구조의 예시를 제공합니다.
-    - 모델 내부 구조는 가능한 일관적으로 공개되어 있습니다.
-    - 빠른 실험을 위해 모델 파일은 라이브러리와 독립적으로 사용될 수 있습니다.
-
-## 왜 transformers를 사용하지 말아야 할까요?
-
-- 이 라이브러리는 신경망 블록을 만들기 위한 모듈이 아닙니다. 연구자들이 여러 파일을 살펴보지 않고 바로 각 모델을 사용할 수 있도록, 모델 파일 코드의 추상화 수준을 적정하게 유지했습니다.
-- 학습 API는 모든 모델에 적용할 수 있도록 만들어지진 않았지만, 라이브러리가 제공하는 모델들에 적용할 수 있도록 최적화되었습니다. 일반적인 머신 러닝을 위해선, 다른 라이브러리를 사용하세요.
-- 가능한 많은 사용 예시를 보여드리고 싶어서, [예시 폴더](https://github.com/huggingface/transformers/tree/main/examples)의 스크립트를 준비했습니다. 이 스크립트들을 수정 없이 특정한 문제에 바로 적용하지 못할 수 있습니다. 필요에 맞게 일부 코드를 수정해야 할 수 있습니다.
-
-## 설치
-
-### pip로 설치하기
-
-이 저장소는 Python 3.6+, Flax 0.3.2+, PyTorch 1.3.1+, TensorFlow 2.3+에서 테스트 되었습니다.
-
-[가상 환경](https://docs.python.org/3/library/venv.html)에 🤗 Transformers를 설치하세요. Python 가상 환경에 익숙하지 않다면, [사용자 가이드](https://packaging.python.org/guides/installing-using-pip-and-virtual-environments/)를 확인하세요.
-
-우선, 사용할 Python 버전으로 가상 환경을 만들고 실행하세요.
-
-그 다음, Flax, PyTorch, TensorFlow 중 적어도 하나는 설치해야 합니다.
-플랫폼에 맞는 설치 명령어를 확인하기 위해 [TensorFlow 설치 페이지](https://www.tensorflow.org/install/), [PyTorch 설치 페이지](https://pytorch.org/get-started/locally/#start-locally), [Flax 설치 페이지](https://github.com/google/flax#quick-install)를 확인하세요.
-
-이들 중 적어도 하나가 설치되었다면, 🤗 Transformers는 다음과 같이 pip을 이용해 설치할 수 있습니다:
-
-```bash
-pip install transformers
-```
-
-예시들을 체험해보고 싶거나, 최최최첨단 코드를 원하거나, 새로운 버전이 나올 때까지 기다릴 수 없다면 [라이브러리를 소스에서 바로 설치](https://huggingface.co/docs/transformers/installation#installing-from-source)하셔야 합니다.
-
-### conda로 설치하기
-
-Transformers 버전 v4.0.0부터, conda 채널이 생겼습니다: `huggingface`.
-
-🤗 Transformers는 다음과 같이 conda로 설치할 수 있습니다:
-
-```shell script
-conda install -c huggingface transformers
-```
-
-Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는 방법을 확인하세요.
-
-## 모델 구조
-
-**🤗 Transformers가 제공하는 [모든 모델 체크포인트](https://huggingface.co/models)** 는 huggingface.co [모델 허브](https://huggingface.co)에 완벽히 연동되어 있습니다. [개인](https://huggingface.co/users)과 [기관](https://huggingface.co/organizations)이 모델 허브에 직접 업로드할 수 있습니다.
-
-현재 사용 가능한 모델 체크포인트의 개수: ![](https://img.shields.io/endpoint?url=https://huggingface.co/api/shields/models&color=brightgreen)
-
-🤗 Transformers는 다음 모델들을 제공합니다 (각 모델의 요약은 [여기](https://huggingface.co/docs/transformers/model_summary)서 확인하세요):
-
-1. **[ALBERT](https://huggingface.co/docs/transformers/model_doc/albert)** (from Google Research and the Toyota Technological Institute at Chicago) released with the paper [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://arxiv.org/abs/1909.11942), by Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut.
-1. **[BART](https://huggingface.co/docs/transformers/model_doc/bart)** (from Facebook) released with the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/pdf/1910.13461.pdf) by Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Ves Stoyanov and Luke Zettlemoyer.
-1. **[BARThez](https://huggingface.co/docs/transformers/model_doc/barthez)** (from École polytechnique) released with the paper [BARThez: a Skilled Pretrained French Sequence-to-Sequence Model](https://arxiv.org/abs/2010.12321) by Moussa Kamal Eddine, Antoine J.-P. Tixier, Michalis Vazirgiannis.
-1. **[BARTpho](https://huggingface.co/docs/transformers/model_doc/bartpho)** (from VinAI Research) released with the paper [BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese](https://arxiv.org/abs/2109.09701) by Nguyen Luong Tran, Duong Minh Le and Dat Quoc Nguyen.
-1. **[BEiT](https://huggingface.co/docs/transformers/model_doc/beit)** (from Microsoft) released with the paper [BEiT: BERT Pre-Training of Image Transformers](https://arxiv.org/abs/2106.08254) by Hangbo Bao, Li Dong, Furu Wei.
-1. **[BERT](https://huggingface.co/docs/transformers/model_doc/bert)** (from Google) released with the paper [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://arxiv.org/abs/1810.04805) by Jacob Devlin, Ming-Wei Chang, Kenton Lee and Kristina Toutanova.
-1. **[BERT For Sequence Generation](https://huggingface.co/docs/transformers/model_doc/bert-generation)** (from Google) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
-1. **[BERTweet](https://huggingface.co/docs/transformers/model_doc/bertweet)** (from VinAI Research) released with the paper [BERTweet: A pre-trained language model for English Tweets](https://aclanthology.org/2020.emnlp-demos.2/) by Dat Quoc Nguyen, Thanh Vu and Anh Tuan Nguyen.
-1. **[BigBird-Pegasus](https://huggingface.co/docs/transformers/model_doc/bigbird_pegasus)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
-1. **[BigBird-RoBERTa](https://huggingface.co/docs/transformers/model_doc/big_bird)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
-1. **[Blenderbot](https://huggingface.co/docs/transformers/model_doc/blenderbot)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
-1. **[BlenderbotSmall](https://huggingface.co/docs/transformers/model_doc/blenderbot-small)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
-1. **[BLOOM](https://huggingface.co/docs/transformers/model_doc/bloom)** (from BigScience workshop) released by the [BigSicence Workshop](https://bigscience.huggingface.co/).
-1. **[BORT](https://huggingface.co/docs/transformers/model_doc/bort)** (from Alexa) released with the paper [Optimal Subarchitecture Extraction For BERT](https://arxiv.org/abs/2010.10499) by Adrian de Wynter and Daniel J. Perry.
-1. **[ByT5](https://huggingface.co/docs/transformers/model_doc/byt5)** (from Google Research) released with the paper [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) by Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel.
-1. **[CamemBERT](https://huggingface.co/docs/transformers/model_doc/camembert)** (from Inria/Facebook/Sorbonne) released with the paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) by Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suárez*, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah and Benoît Sagot.
-1. **[CANINE](https://huggingface.co/docs/transformers/model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
-1. **[CLIP](https://huggingface.co/docs/transformers/model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
-1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
-1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
-1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
-1. **[CPM](https://huggingface.co/docs/transformers/model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
-1. **[CTRL](https://huggingface.co/docs/transformers/model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
-1. **[CvT](https://huggingface.co/docs/transformers/model_doc/cvt)** (from Microsoft) released with the paper [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) by Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang.
-1. **[Data2Vec](https://huggingface.co/docs/transformers/model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
-1. **[DeBERTa](https://huggingface.co/docs/transformers/model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
-1. **[DeBERTa-v2](https://huggingface.co/docs/transformers/model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
-1. **[Decision Transformer](https://huggingface.co/docs/transformers/model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
-1. **[DeiT](https://huggingface.co/docs/transformers/model_doc/deit)** (from Facebook) released with the paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) by Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou.
-1. **[DETR](https://huggingface.co/docs/transformers/model_doc/detr)** (from Facebook) released with the paper [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) by Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko.
-1. **[DialoGPT](https://huggingface.co/docs/transformers/model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
-1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/distillation) and a German version of DistilBERT.
-1. **[DiT](https://huggingface.co/docs/transformers/model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
-1. **[DPR](https://huggingface.co/docs/transformers/model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
-1. **[DPT](https://huggingface.co/docs/transformers/master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by René Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
-1. **[ELECTRA](https://huggingface.co/docs/transformers/model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
-1. **[EncoderDecoder](https://huggingface.co/docs/transformers/model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
-1. **[FlauBERT](https://huggingface.co/docs/transformers/model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab.
-1. **[FLAVA](https://huggingface.co/docs/transformers/model_doc/flava)** (from Facebook AI) released with the paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) by Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela.
-1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
-1. **[Funnel Transformer](https://huggingface.co/docs/transformers/model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
-1. **[GLPN](https://huggingface.co/docs/transformers/model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
-1. **[GPT](https://huggingface.co/docs/transformers/model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
-1. **[GPT Neo](https://huggingface.co/docs/transformers/model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
-1. **[GPT NeoX](https://huggingface.co/docs/transformers/model_doc/gpt_neox)** (from EleutherAI) released with the paper [GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://arxiv.org/abs/2204.06745) by Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, Samuel Weinbach 
-1. **[GPT-2](https://huggingface.co/docs/transformers/model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
-1. **[GPT-J](https://huggingface.co/docs/transformers/model_doc/gptj)** (from EleutherAI) released in the repository [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
-1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
-1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (from Facebook) released with the paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) by Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
-1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (from Berkeley) released with the paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) by Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
-1. **[ImageGPT](https://huggingface.co/docs/transformers/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
-1. **[LayoutLM](https://huggingface.co/docs/transformers/model_doc/layoutlm)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
-1. **[LayoutLMv2](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) by Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
-1. **[LayoutLMv3](https://huggingface.co/docs/transformers/model_doc/layoutlmv3)** (from Microsoft Research Asia) released with the paper [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) by Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei.
-1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
-1. **[LED](https://huggingface.co/docs/transformers/model_doc/led)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
-1. **[LeViT](https://huggingface.co/docs/transformers/model_doc/levit)** (from Meta AI) released with the paper [LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference](https://arxiv.org/abs/2104.01136) by Ben Graham, Alaaeldin El-Nouby, Hugo Touvron, Pierre Stock, Armand Joulin, Hervé Jégou, Matthijs Douze.
-1. **[Longformer](https://huggingface.co/docs/transformers/model_doc/longformer)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
-1. **[LongT5](https://huggingface.co/docs/transformers/model_doc/longt5)** (from Google AI) released with the paper [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://arxiv.org/abs/2112.07916) by Mandy Guo, Joshua Ainslie, David Uthus, Santiago Ontanon, Jianmo Ni, Yun-Hsuan Sung, Yinfei Yang.
-1. **[LUKE](https://huggingface.co/docs/transformers/model_doc/luke)** (from Studio Ousia) released with the paper [LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention](https://arxiv.org/abs/2010.01057) by Ikuya Yamada, Akari Asai, Hiroyuki Shindo, Hideaki Takeda, Yuji Matsumoto.
-1. **[LXMERT](https://huggingface.co/docs/transformers/model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
-1. **[M-CTC-T](https://huggingface.co/docs/transformers/model_doc/mctct)** (from Facebook) released with the paper [Pseudo-Labeling For Massively Multilingual Speech Recognition](https://arxiv.org/abs/2111.00161) by Loren Lugosch, Tatiana Likhomanenko, Gabriel Synnaeve, and Ronan Collobert.
-1. **[M2M100](https://huggingface.co/docs/transformers/model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
-1. **[MarianMT](https://huggingface.co/docs/transformers/model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jörg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
-1. **[MaskFormer](https://huggingface.co/docs/transformers/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
-1. **[mBART](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
-1. **[mBART-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
-1. **[Megatron-BERT](https://huggingface.co/docs/transformers/model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
-1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
-1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
-1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
-1. **[MobileViT](https://huggingface.co/docs/transformers/model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
-1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
-1. **[MT5](https://huggingface.co/docs/transformers/model_doc/mt5)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
-1. **[MVP](https://huggingface.co/docs/transformers/model_doc/mvp)** (from RUC AI Box) released with the paper [MVP: Multi-task Supervised Pre-training for Natural Language Generation](https://arxiv.org/abs/2206.12131) by Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen.
-1. **[Nezha](https://huggingface.co/docs/transformers/model_doc/nezha)** (from Huawei Noah’s Ark Lab) released with the paper [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://arxiv.org/abs/1909.00204) by Junqiu Wei, Xiaozhe Ren, Xiaoguang Li, Wenyong Huang, Yi Liao, Yasheng Wang, Jiashu Lin, Xin Jiang, Xiao Chen and Qun Liu.
-1. **[NLLB](https://huggingface.co/docs/transformers/model_doc/nllb)** (from Meta) released with the paper [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) by the NLLB team.
-1. **[Nyströmformer](https://huggingface.co/docs/transformers/model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
-1. **[OPT](https://huggingface.co/docs/transformers/master/model_doc/opt)** (from Meta AI) released with the paper [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) by Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al.
-1. **[OWL-ViT](https://huggingface.co/docs/transformers/model_doc/owlvit)** (from Google AI) released with the paper [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) by Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby.
-1. **[Pegasus](https://huggingface.co/docs/transformers/model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
-1. **[Perceiver IO](https://huggingface.co/docs/transformers/model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira.
-1. **[PhoBERT](https://huggingface.co/docs/transformers/model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
-1. **[PLBart](https://huggingface.co/docs/transformers/model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
-1. **[PoolFormer](https://huggingface.co/docs/transformers/model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
-1. **[ProphetNet](https://huggingface.co/docs/transformers/model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
-1. **[QDQBert](https://huggingface.co/docs/transformers/model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
-1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
-1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
-1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
-1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Research) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár. 
-1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/pdf/2010.12821.pdf) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
-1. **[ResNet](https://huggingface.co/docs/transformers/model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun. 
-1. **[RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta)** (from Facebook), released together with the paper a [Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
-1. **[RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper a [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/pdf/2104.09864v1.pdf) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
-1. **[SegFormer](https://huggingface.co/docs/transformers/model_doc/segformer)** (from NVIDIA) released with the paper [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) by Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo.
-1. **[SEW](https://huggingface.co/docs/transformers/model_doc/sew)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
-1. **[SEW-D](https://huggingface.co/docs/transformers/model_doc/sew_d)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
-1. **[SpeechToTextTransformer](https://huggingface.co/docs/transformers/model_doc/speech_to_text)** (from Facebook), released together with the paper [fairseq S2T: Fast Speech-to-Text Modeling with fairseq](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Dmytro Okhonko, Juan Pino.
-1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (from Facebook), released together with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
-1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (from Tel Aviv University), released together with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
-1. **[SqueezeBERT](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
-1. **[Swin Transformer](https://huggingface.co/docs/transformers/model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
-1. **[Swin Transformer V2](https://huggingface.co/docs/transformers/main/model_doc/swinv2)** (from Microsoft) released with the paper [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) by Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo.
-1. **[T5](https://huggingface.co/docs/transformers/model_doc/t5)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
-1. **[T5v1.1](https://huggingface.co/docs/transformers/model_doc/t5v1.1)** (from Google AI) released in the repository [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
-1. **[TAPAS](https://huggingface.co/docs/transformers/model_doc/tapas)** (from Google AI) released with the paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) by Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos.
-1. **[TAPEX](https://huggingface.co/docs/transformers/model_doc/tapex)** (from Microsoft Research) released with the paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) by Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
-1. **[Trajectory Transformer](https://huggingface.co/docs/transformers/model_doc/trajectory_transformers)** (from the University of California at Berkeley) released with the paper [Offline Reinforcement Learning as One Big Sequence Modeling Problem](https://arxiv.org/abs/2106.02039) by Michael Janner, Qiyang Li, Sergey Levine 
-1. **[Transformer-XL](https://huggingface.co/docs/transformers/model_doc/transfo-xl)** (from Google/CMU) released with the paper [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) by Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov.
-1. **[TrOCR](https://huggingface.co/docs/transformers/model_doc/trocr)** (from Microsoft), released together with the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
-1. **[UL2](https://huggingface.co/docs/transformers/model_doc/ul2)** (from Google Research) released with the paper [Unifying Language Learning Paradigms](https://arxiv.org/abs/2205.05131v1) by Yi Tay, Mostafa Dehghani, Vinh Q. Tran, Xavier Garcia, Dara Bahri, Tal Schuster, Huaixiu Steven Zheng, Neil Houlsby, Donald Metzler 
-1. **[UniSpeech](https://huggingface.co/docs/transformers/model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
-1. **[UniSpeechSat](https://huggingface.co/docs/transformers/model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
-1. **[VAN](https://huggingface.co/docs/transformers/model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/pdf/2202.09741.pdf) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
-1. **[VideoMAE](https://huggingface.co/docs/transformers/main/model_doc/videomae)** (from Multimedia Computing Group, Nanjing University) released with the paper [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://arxiv.org/abs/2203.12602) by Zhan Tong, Yibing Song, Jue Wang, Limin Wang.
-1. **[ViLT](https://huggingface.co/docs/transformers/model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
-1. **[Vision Transformer (ViT)](https://huggingface.co/docs/transformers/model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby.
-1. **[VisualBERT](https://huggingface.co/docs/transformers/model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
-1. **[ViTMAE](https://huggingface.co/docs/transformers/model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
-1. **[Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
-1. **[Wav2Vec2-Conformer](https://huggingface.co/docs/transformers/model_doc/wav2vec2-conformer)** (from Facebook AI) released with the paper [FAIRSEQ S2T: Fast Speech-to-Text Modeling with FAIRSEQ](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Sravya Popuri, Dmytro Okhonko, Juan Pino.
-1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/transformers/model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
-1. **[WavLM](https://huggingface.co/docs/transformers/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
-1. **[XGLM](https://huggingface.co/docs/transformers/model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li. 
-1. **[XLM](https://huggingface.co/docs/transformers/model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
-1. **[XLM-ProphetNet](https://huggingface.co/docs/transformers/model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
-1. **[XLM-RoBERTa](https://huggingface.co/docs/transformers/model_doc/xlm-roberta)** (from Facebook AI), released together with the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov.
-1. **[XLM-RoBERTa-XL](https://huggingface.co/docs/transformers/model_doc/xlm-roberta-xl)** (from Facebook AI) released with the paper [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) by Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau.
-1. **[XLNet](https://huggingface.co/docs/transformers/model_doc/xlnet)** (from Google/CMU) released with the paper [​XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
-1. **[XLS-R](https://huggingface.co/docs/transformers/model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
-1. **[XLSR-Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
-1. **[YOLOS](https://huggingface.co/docs/transformers/model_doc/yolos)** (from Huazhong University of Science & Technology) released with the paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) by Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
-1. **[YOSO](https://huggingface.co/docs/transformers/model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
-1. 새로운 모델을 올리고 싶나요? 우리가 **상세한 가이드와 템플릿** 으로 새로운 모델을 올리도록 도와드릴게요. 가이드와 템플릿은 이 저장소의 [`templates`](./templates) 폴더에서 확인하실 수 있습니다. [컨트리뷰션 가이드라인](./CONTRIBUTING.md)을 꼭 확인해주시고, PR을 올리기 전에 메인테이너에게 연락하거나 이슈를 오픈해 피드백을 받으시길 바랍니다. 
-
-각 모델이 Flax, PyTorch, TensorFlow으로 구현되었는지 또는 🤗 Tokenizers 라이브러리가 지원하는 토크나이저를 사용하는지 확인하려면, [이 표](https://huggingface.co/docs/transformers/index#supported-frameworks)를 확인하세요.
-
-이 구현은 여러 데이터로 검증되었고 (예시 스크립트를 참고하세요) 오리지널 구현의 성능과 같아야 합니다. [도큐먼트](https://huggingface.co/docs/transformers/examples)의 Examples 섹션에서 성능에 대한 자세한 설명을 확인할 수 있습니다.
-
-## 더 알아보기
-
-| 섹션 | 설명 |
-|-|-|
-| [도큐먼트](https://huggingface.co/transformers/) | 전체 API 도큐먼트와 튜토리얼 |
-| [과제 요약](https://huggingface.co/docs/transformers/task_summary) | 🤗 Transformers가 지원하는 과제들 |
-| [전처리 튜토리얼](https://huggingface.co/docs/transformers/preprocessing) | `Tokenizer` 클래스를 이용해 모델을 위한 데이터 준비하기 |
-| [학습과 fine-tuning](https://huggingface.co/docs/transformers/training) | 🤗 Transformers가 제공하는 모델 PyTorch/TensorFlow 학습 과정과 `Trainer` API에서 사용하기 |
-| [퀵 투어: Fine-tuning/사용 스크립트](https://github.com/huggingface/transformers/tree/main/examples) | 다양한 과제에서 모델 fine-tuning하는 예시 스크립트 |
-| [모델 공유 및 업로드](https://huggingface.co/docs/transformers/model_sharing) | 커뮤니티에 fine-tune된 모델을 업로드 및 공유하기 |
-| [마이그레이션](https://huggingface.co/docs/transformers/migration) | `pytorch-transformers`나 `pytorch-pretrained-bert`에서 🤗 Transformers로 이동하기|
-
-## 인용
-
-🤗 Transformers 라이브러리를 인용하고 싶다면, 이 [논문](https://www.aclweb.org/anthology/2020.emnlp-demos.6/)을 인용해 주세요:
-```bibtex
-@inproceedings{wolf-etal-2020-transformers,
-    title = "Transformers: State-of-the-Art Natural Language Processing",
-    author = "Thomas Wolf and Lysandre Debut and Victor Sanh and Julien Chaumond and Clement Delangue and Anthony Moi and Pierric Cistac and Tim Rault and Rémi Louf and Morgan Funtowicz and Joe Davison and Sam Shleifer and Patrick von Platen and Clara Ma and Yacine Jernite and Julien Plu and Canwen Xu and Teven Le Scao and Sylvain Gugger and Mariama Drame and Quentin Lhoest and Alexander M. Rush",
-    booktitle = "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations",
-    month = oct,
-    year = "2020",
-    address = "Online",
-    publisher = "Association for Computational Linguistics",
-    url = "https://www.aclweb.org/anthology/2020.emnlp-demos.6",
-    pages = "38--45"
-}
-```

README_zh-hans.md

@@ -1,404 +0,0 @@
-<!---
-Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
--->
-
-<!---
-A useful guide for English-Chinese translation of Hugging Face documentation
-- Add space around English words and numbers when they appear between Chinese characters. E.g., 共 100 多种语言; 使用 transformers 库。
-- Use square quotes, e.g.,「引用」
-
-Dictionary
-
-Hugging Face: 抱抱脸
-token: 词符（并用括号标注原英文）
-tokenize: 词符化（并用括号标注原英文）
-tokenizer: 词符化器（并用括号标注原英文）
-transformer: transformer（不翻译）
-pipeline: 流水线
-API: API (不翻译）
-inference: 推理
-Trainer: 训练器。当作为类名出现时不翻译。
-pretrained/pretrain: 预训练
-finetune: 微调
-community: 社区
-example: 当特指仓库中 example 目录时翻译为「用例」
-Python data structures (e.g., list, set, dict): 翻译为列表，集合，词典，并用括号标注原英文
-NLP/Natural Language Processing: 以 NLP 出现时不翻译，以 Natural Language Processing 出现时翻译为自然语言处理
-checkpoint: 检查点
--->
-
-<p align="center">
-    <br>
-    <img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers_logo_name.png" width="400"/>
-    <br>
-<p>
-<p align="center">
-    <a href="https://circleci.com/gh/huggingface/transformers">
-        <img alt="Build" src="https://img.shields.io/circleci/build/github/huggingface/transformers/main">
-    </a>
-    <a href="https://github.com/huggingface/transformers/blob/main/LICENSE">
-        <img alt="GitHub" src="https://img.shields.io/github/license/huggingface/transformers.svg?color=blue">
-    </a>
-    <a href="https://huggingface.co/docs/transformers/index">
-        <img alt="Documentation" src="https://img.shields.io/website/http/huggingface.co/docs/transformers/index.svg?down_color=red&down_message=offline&up_message=online">
-    </a>
-    <a href="https://github.com/huggingface/transformers/releases">
-        <img alt="GitHub release" src="https://img.shields.io/github/release/huggingface/transformers.svg">
-    </a>
-    <a href="https://github.com/huggingface/transformers/blob/main/CODE_OF_CONDUCT.md">
-        <img alt="Contributor Covenant" src="https://img.shields.io/badge/Contributor%20Covenant-v2.0%20adopted-ff69b4.svg">
-    </a>
-    <a href="https://zenodo.org/badge/latestdoi/155220641"><img src="https://zenodo.org/badge/155220641.svg" alt="DOI"></a>
-</p>
-
-<h4 align="center">
-    <p>
-        <a href="https://github.com/huggingface/transformers/">English</a> |
-        <b>简体中文</b> |
-        <a href="https://github.com/huggingface/transformers/blob/main/README_zh-hant.md">繁體中文</a> |
-        <a href="https://github.com/huggingface/transformers/blob/main/README_ko.md">한국어</a>
-    <p>
-</h4>
-
-<h3 align="center">
-    <p>为 Jax、PyTorch 和 TensorFlow 打造的先进的自然语言处理</p>
-</h3>
-
-<h3 align="center">
-    <a href="https://hf.co/course"><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/course_banner.png"></a>
-</h3>
-
-🤗 Transformers 提供了数以千计的预训练模型，支持 100 多种语言的文本分类、信息抽取、问答、摘要、翻译、文本生成。它的宗旨让最先进的 NLP 技术人人易用。
-
-🤗 Transformers 提供了便于快速下载和使用的API，让你可以把预训练模型用在给定文本、在你的数据集上微调然后通过 [model hub](https://huggingface.co/models) 与社区共享。同时，每个定义的 Python 模块均完全独立，方便修改和快速研究实验。
-
-🤗 Transformers 支持三个最热门的深度学习库： [Jax](https://jax.readthedocs.io/en/latest/), [PyTorch](https://pytorch.org/) and [TensorFlow](https://www.tensorflow.org/) — 并与之无缝整合。你可以直接使用一个框架训练你的模型然后用另一个加载和推理。
-
-## 在线演示
-
-你可以直接在模型页面上测试大多数 [model hub](https://huggingface.co/models) 上的模型。 我们也提供了 [私有模型托管、模型版本管理以及推理API](https://huggingface.co/pricing)。
-
-这里是一些例子：
-- [用 BERT 做掩码填词](https://huggingface.co/bert-base-uncased?text=Paris+is+the+%5BMASK%5D+of+France)
-- [用 Electra 做命名实体识别](https://huggingface.co/dbmdz/electra-large-discriminator-finetuned-conll03-english?text=My+name+is+Sarah+and+I+live+in+London+city)
-- [用 GPT-2 做文本生成](https://huggingface.co/gpt2?text=A+long+time+ago%2C+)
-- [用 RoBERTa 做自然语言推理](https://huggingface.co/roberta-large-mnli?text=The+dog+was+lost.+Nobody+lost+any+animal)
-- [用 BART 做文本摘要](https://huggingface.co/facebook/bart-large-cnn?text=The+tower+is+324+metres+%281%2C063+ft%29+tall%2C+about+the+same+height+as+an+81-storey+building%2C+and+the+tallest+structure+in+Paris.+Its+base+is+square%2C+measuring+125+metres+%28410+ft%29+on+each+side.+During+its+construction%2C+the+Eiffel+Tower+surpassed+the+Washington+Monument+to+become+the+tallest+man-made+structure+in+the+world%2C+a+title+it+held+for+41+years+until+the+Chrysler+Building+in+New+York+City+was+finished+in+1930.+It+was+the+first+structure+to+reach+a+height+of+300+metres.+Due+to+the+addition+of+a+broadcasting+aerial+at+the+top+of+the+tower+in+1957%2C+it+is+now+taller+than+the+Chrysler+Building+by+5.2+metres+%2817+ft%29.+Excluding+transmitters%2C+the+Eiffel+Tower+is+the+second+tallest+free-standing+structure+in+France+after+the+Millau+Viaduct)
-- [用 DistilBERT 做问答](https://huggingface.co/distilbert-base-uncased-distilled-squad?text=Which+name+is+also+used+to+describe+the+Amazon+rainforest+in+English%3F&context=The+Amazon+rainforest+%28Portuguese%3A+Floresta+Amaz%C3%B4nica+or+Amaz%C3%B4nia%3B+Spanish%3A+Selva+Amaz%C3%B3nica%2C+Amazon%C3%ADa+or+usually+Amazonia%3B+French%3A+For%C3%AAt+amazonienne%3B+Dutch%3A+Amazoneregenwoud%29%2C+also+known+in+English+as+Amazonia+or+the+Amazon+Jungle%2C+is+a+moist+broadleaf+forest+that+covers+most+of+the+Amazon+basin+of+South+America.+This+basin+encompasses+7%2C000%2C000+square+kilometres+%282%2C700%2C000+sq+mi%29%2C+of+which+5%2C500%2C000+square+kilometres+%282%2C100%2C000+sq+mi%29+are+covered+by+the+rainforest.+This+region+includes+territory+belonging+to+nine+nations.+The+majority+of+the+forest+is+contained+within+Brazil%2C+with+60%25+of+the+rainforest%2C+followed+by+Peru+with+13%25%2C+Colombia+with+10%25%2C+and+with+minor+amounts+in+Venezuela%2C+Ecuador%2C+Bolivia%2C+Guyana%2C+Suriname+and+French+Guiana.+States+or+departments+in+four+nations+contain+%22Amazonas%22+in+their+names.+The+Amazon+represents+over+half+of+the+planet%27s+remaining+rainforests%2C+and+comprises+the+largest+and+most+biodiverse+tract+of+tropical+rainforest+in+the+world%2C+with+an+estimated+390+billion+individual+trees+divided+into+16%2C000+species)
-- [用 T5 做翻译](https://huggingface.co/t5-base?text=My+name+is+Wolfgang+and+I+live+in+Berlin)
-
-**[Write With Transformer](https://transformer.huggingface.co)**，由抱抱脸团队打造，是一个文本生成的官方 demo。
-
-## 如果你在寻找由抱抱脸团队提供的定制化支持服务
-
-<a target="_blank" href="https://huggingface.co/support">
-    <img alt="HuggingFace Expert Acceleration Program" src="https://huggingface.co/front/thumbnails/support.png" style="max-width: 600px; border: 1px solid #eee; border-radius: 4px; box-shadow: 0 1px 2px 0 rgba(0, 0, 0, 0.05);">
-</a><br>
-
-## 快速上手
-
-我们为快速使用模型提供了 `pipeline` （流水线）API。流水线聚合了预训练模型和对应的文本预处理。下面是一个快速使用流水线去判断正负面情绪的例子：
-
-```python
->>> from transformers import pipeline
-
-# 使用情绪分析流水线
->>> classifier = pipeline('sentiment-analysis')
->>> classifier('We are very happy to introduce pipeline to the transformers repository.')
-[{'label': 'POSITIVE', 'score': 0.9996980428695679}]
-```
-
-第二行代码下载并缓存了流水线使用的预训练模型，而第三行代码则在给定的文本上进行了评估。这里的答案“正面” (positive) 具有 99 的置信度。
-
-许多的 NLP 任务都有开箱即用的预训练流水线。比如说，我们可以轻松的从给定文本中抽取问题答案：
-
-``` python
->>> from transformers import pipeline
-
-# 使用问答流水线
->>> question_answerer = pipeline('question-answering')
->>> question_answerer({
-...     'question': 'What is the name of the repository ?',
-...     'context': 'Pipeline has been included in the huggingface/transformers repository'
-... })
-{'score': 0.30970096588134766, 'start': 34, 'end': 58, 'answer': 'huggingface/transformers'}
-
-```
-
-除了给出答案，预训练模型还给出了对应的置信度分数、答案在词符化 (tokenized) 后的文本中开始和结束的位置。你可以从[这个教程](https://huggingface.co/docs/transformers/task_summary)了解更多流水线API支持的任务。
-
-要在你的任务上下载和使用任意预训练模型也很简单，只需三行代码。这里是 PyTorch 版的示例：
-```python
->>> from transformers import AutoTokenizer, AutoModel
-
->>> tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
->>> model = AutoModel.from_pretrained("bert-base-uncased")
-
->>> inputs = tokenizer("Hello world!", return_tensors="pt")
->>> outputs = model(**inputs)
-```
-这里是等效的 TensorFlow 代码：
-```python
->>> from transformers import AutoTokenizer, TFAutoModel
-
->>> tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
->>> model = TFAutoModel.from_pretrained("bert-base-uncased")
-
->>> inputs = tokenizer("Hello world!", return_tensors="tf")
->>> outputs = model(**inputs)
-```
-
-词符化器 (tokenizer) 为所有的预训练模型提供了预处理，并可以直接对单个字符串进行调用（比如上面的例子）或对列表 (list) 调用。它会输出一个你可以在下游代码里使用或直接通过 `**` 解包表达式传给模型的词典 (dict)。
-
-模型本身是一个常规的 [Pytorch `nn.Module`](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) 或 [TensorFlow `tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model)（取决于你的后端），可以常规方式使用。 [这个教程](https://huggingface.co/transformers/training.html)解释了如何将这样的模型整合到经典的 PyTorch 或 TensorFlow 训练循环中，或是如何使用我们的 `Trainer` 训练器）API 来在一个新的数据集上快速微调。
-
-## 为什么要用 transformers？
-
-1. 便于使用的先进模型：
-    - NLU 和 NLG 上表现优越
-    - 对教学和实践友好且低门槛
-    - 高级抽象，只需了解三个类
-    - 对所有模型统一的API
-
-1. 更低计算开销，更少的碳排放：
-    - 研究人员可以分享亿训练的模型而非次次从头开始训练
-    - 工程师可以减少计算用时和生产环境开销
-    - 数十种模型架构、两千多个预训练模型、100多种语言支持
-
-1. 对于模型生命周期的每一个部分都面面俱到：
-    - 训练先进的模型，只需 3 行代码
-    - 模型在不同深度学习框架间任意转移，随你心意
-    - 为训练、评估和生产选择最适合的框架，衔接无缝
-
-1. 为你的需求轻松定制专属模型和用例：
-    - 我们为每种模型架构提供了多个用例来复现原论文结果
-    - 模型内部结构保持透明一致
-    - 模型文件可单独使用，方便魔改和快速实验
-
-## 什么情况下我不该用 transformers？
-
-- 本库并不是模块化的神经网络工具箱。模型文件中的代码特意呈若璞玉，未经额外抽象封装，以便研究人员快速迭代魔改而不致溺于抽象和文件跳转之中。
-- `Trainer` API 并非兼容任何模型，只为本库之模型优化。若是在寻找适用于通用机器学习的训练循环实现，请另觅他库。
-- 尽管我们已尽力而为，[examples 目录](https://github.com/huggingface/transformers/tree/main/examples)中的脚本也仅为用例而已。对于你的特定问题，它们并不一定开箱即用，可能需要改几行代码以适之。
-
-## 安装
-
-### 使用 pip
-
-这个仓库已在 Python 3.6+、Flax 0.3.2+、PyTorch 1.3.1+ 和 TensorFlow 2.3+ 下经过测试。
-
-你可以在[虚拟环境](https://docs.python.org/3/library/venv.html)中安装 🤗 Transformers。如果你还不熟悉 Python 的虚拟环境，请阅此[用户说明](https://packaging.python.org/guides/installing-using-pip-and-virtual-environments/)。
-
-首先，用你打算使用的版本的 Python 创建一个虚拟环境并激活。
-
-然后，你需要安装 Flax、PyTorch 或 TensorFlow 其中之一。关于在你使用的平台上安装这些框架，请参阅 [TensorFlow 安装页](https://www.tensorflow.org/install/), [PyTorch 安装页](https://pytorch.org/get-started/locally/#start-locally) 或 [Flax 安装页](https://github.com/google/flax#quick-install)。
-
-当这些后端之一安装成功后， 🤗 Transformers 可依此安装：
-
-```bash
-pip install transformers
-```
-
-如果你想要试试用例或者想在正式发布前使用最新的开发中代码，你得[从源代码安装](https://huggingface.co/docs/transformers/installation#installing-from-source)。
-
-### 使用 conda
-
-自 Transformers 4.0.0 版始，我们有了一个 conda 频道： `huggingface`。
-
-🤗 Transformers 可以通过 conda 依此安装：
-
-```shell script
-conda install -c huggingface transformers
-```
-
-要通过 conda 安装 Flax、PyTorch 或 TensorFlow 其中之一，请参阅它们各自安装页的说明。
-
-## 模型架构
-
-🤗 Transformers 支持的[**所有的模型检查点**](https://huggingface.co/models)由[用户](https://huggingface.co/users)和[组织](https://huggingface.co/organizations)上传，均与 huggingface.co [model hub](https://huggingface.co) 无缝整合。
-
-目前的检查点数量： ![](https://img.shields.io/endpoint?url=https://huggingface.co/api/shields/models&color=brightgreen)
-
-🤗 Transformers 目前支持如下的架构（模型概述请阅[这里](https://huggingface.co/docs/transformers/model_summary)）：
-
-1. **[ALBERT](https://huggingface.co/docs/transformers/model_doc/albert)** (来自 Google Research and the Toyota Technological Institute at Chicago) 伴随论文 [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://arxiv.org/abs/1909.11942), 由 Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut 发布。
-1. **[BART](https://huggingface.co/docs/transformers/model_doc/bart)** (来自 Facebook) 伴随论文 [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/pdf/1910.13461.pdf) 由 Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Ves Stoyanov and Luke Zettlemoyer 发布。
-1. **[BARThez](https://huggingface.co/docs/transformers/model_doc/barthez)** (来自 École polytechnique) 伴随论文 [BARThez: a Skilled Pretrained French Sequence-to-Sequence Model](https://arxiv.org/abs/2010.12321) 由 Moussa Kamal Eddine, Antoine J.-P. Tixier, Michalis Vazirgiannis 发布。
-1. **[BARTpho](https://huggingface.co/docs/transformers/model_doc/bartpho)** (来自 VinAI Research) 伴随论文 [BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese](https://arxiv.org/abs/2109.09701) 由 Nguyen Luong Tran, Duong Minh Le and Dat Quoc Nguyen 发布。
-1. **[BEiT](https://huggingface.co/docs/transformers/model_doc/beit)** (来自 Microsoft) 伴随论文 [BEiT: BERT Pre-Training of Image Transformers](https://arxiv.org/abs/2106.08254) 由 Hangbo Bao, Li Dong, Furu Wei 发布。
-1. **[BERT](https://huggingface.co/docs/transformers/model_doc/bert)** (来自 Google) 伴随论文 [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://arxiv.org/abs/1810.04805) 由 Jacob Devlin, Ming-Wei Chang, Kenton Lee and Kristina Toutanova 发布。
-1. **[BERT For Sequence Generation](https://huggingface.co/docs/transformers/model_doc/bert-generation)** (来自 Google) 伴随论文 [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) 由 Sascha Rothe, Shashi Narayan, Aliaksei Severyn 发布。
-1. **[BERTweet](https://huggingface.co/docs/transformers/model_doc/bertweet)** (来自 VinAI Research) 伴随论文 [BERTweet: A pre-trained language model for English Tweets](https://aclanthology.org/2020.emnlp-demos.2/) 由 Dat Quoc Nguyen, Thanh Vu and Anh Tuan Nguyen 发布。
-1. **[BigBird-Pegasus](https://huggingface.co/docs/transformers/model_doc/bigbird_pegasus)** (来自 Google Research) 伴随论文 [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) 由 Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed 发布。
-1. **[BigBird-RoBERTa](https://huggingface.co/docs/transformers/model_doc/big_bird)** (来自 Google Research) 伴随论文 [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) 由 Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed 发布。
-1. **[Blenderbot](https://huggingface.co/docs/transformers/model_doc/blenderbot)** (来自 Facebook) 伴随论文 [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) 由 Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston 发布。
-1. **[BlenderbotSmall](https://huggingface.co/docs/transformers/model_doc/blenderbot-small)** (来自 Facebook) 伴随论文 [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) 由 Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston 发布。
-1. **[BLOOM](https://huggingface.co/docs/transformers/model_doc/bloom)** (from BigScience workshop) released by the [BigSicence Workshop](https://bigscience.huggingface.co/). 
-1. **[BORT](https://huggingface.co/docs/transformers/model_doc/bort)** (来自 Alexa) 伴随论文 [Optimal Subarchitecture Extraction For BERT](https://arxiv.org/abs/2010.10499) 由 Adrian de Wynter and Daniel J. Perry 发布。
-1. **[ByT5](https://huggingface.co/docs/transformers/model_doc/byt5)** (来自 Google Research) 伴随论文 [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) 由 Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel 发布。
-1. **[CamemBERT](https://huggingface.co/docs/transformers/model_doc/camembert)** (来自 Inria/Facebook/Sorbonne) 伴随论文 [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) 由 Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suárez*, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah and Benoît Sagot 发布。
-1. **[CANINE](https://huggingface.co/docs/transformers/model_doc/canine)** (来自 Google Research) 伴随论文 [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) 由 Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting 发布。
-1. **[CLIP](https://huggingface.co/docs/transformers/model_doc/clip)** (来自 OpenAI) 伴随论文 [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) 由 Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever 发布。
-1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (来自 Salesforce) 伴随论文 [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) 由 Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong 发布。
-1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (来自 YituTech) 伴随论文 [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) 由 Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan 发布。
-1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (来自 Facebook AI) 伴随论文 [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) 由 Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie 发布。
-1. **[CPM](https://huggingface.co/docs/transformers/model_doc/cpm)** (来自 Tsinghua University) 伴随论文 [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) 由 Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun 发布。
-1. **[CTRL](https://huggingface.co/docs/transformers/model_doc/ctrl)** (来自 Salesforce) 伴随论文 [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) 由 Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher 发布。
-1. **[CvT](https://huggingface.co/docs/transformers/model_doc/cvt)** (来自 Microsoft) 伴随论文 [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) 由 Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang 发布。
-1. **[Data2Vec](https://huggingface.co/docs/transformers/model_doc/data2vec)** (来自 Facebook) 伴随论文 [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) 由 Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli 发布。
-1. **[DeBERTa](https://huggingface.co/docs/transformers/model_doc/deberta)** (来自 Microsoft) 伴随论文 [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) 由 Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen 发布。
-1. **[DeBERTa-v2](https://huggingface.co/docs/transformers/model_doc/deberta-v2)** (来自 Microsoft) 伴随论文 [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) 由 Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen 发布。
-1. **[Decision Transformer](https://huggingface.co/docs/transformers/model_doc/decision_transformer)** (来自 Berkeley/Facebook/Google) 伴随论文 [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) 由 Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch 发布。
-1. **[DeiT](https://huggingface.co/docs/transformers/model_doc/deit)** (来自 Facebook) 伴随论文 [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) 由 Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou 发布。
-1. **[DETR](https://huggingface.co/docs/transformers/model_doc/detr)** (来自 Facebook) 伴随论文 [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) 由 Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko 发布。
-1. **[DialoGPT](https://huggingface.co/docs/transformers/model_doc/dialogpt)** (来自 Microsoft Research) 伴随论文 [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) 由 Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan 发布。
-1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (来自 HuggingFace), 伴随论文 [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) 由 Victor Sanh, Lysandre Debut and Thomas Wolf 发布。 同样的方法也应用于压缩 GPT-2 到 [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/distillation), RoBERTa 到 [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/distillation), Multilingual BERT 到 [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/distillation) 和德语版 DistilBERT。
-1. **[DiT](https://huggingface.co/docs/transformers/model_doc/dit)** (来自 Microsoft Research) 伴随论文 [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) 由 Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei 发布。
-1. **[DPR](https://huggingface.co/docs/transformers/model_doc/dpr)** (来自 Facebook) 伴随论文 [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) 由 Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih 发布。
-1. **[DPT](https://huggingface.co/docs/transformers/master/model_doc/dpt)** (来自 Intel Labs) 伴随论文 [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) 由 René Ranftl, Alexey Bochkovskiy, Vladlen Koltun 发布。
-1. **[ELECTRA](https://huggingface.co/docs/transformers/model_doc/electra)** (来自 Google Research/Stanford University) 伴随论文 [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) 由 Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning 发布。
-1. **[EncoderDecoder](https://huggingface.co/docs/transformers/model_doc/encoder-decoder)** (来自 Google Research) 伴随论文 [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) 由 Sascha Rothe, Shashi Narayan, Aliaksei Severyn 发布。
-1. **[FlauBERT](https://huggingface.co/docs/transformers/model_doc/flaubert)** (来自 CNRS) 伴随论文 [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) 由 Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab 发布。
-1. **[FLAVA](https://huggingface.co/docs/transformers/model_doc/flava)** (来自 Facebook AI) 伴随论文 [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) 由 Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela 发布。
-1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (来自 Google Research) 伴随论文 [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) 由 James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon 发布。
-1. **[Funnel Transformer](https://huggingface.co/docs/transformers/model_doc/funnel)** (来自 CMU/Google Brain) 伴随论文 [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) 由 Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le 发布。
-1. **[GLPN](https://huggingface.co/docs/transformers/model_doc/glpn)** (来自 KAIST) 伴随论文 [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) 由 Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim 发布。
-1. **[GPT](https://huggingface.co/docs/transformers/model_doc/openai-gpt)** (来自 OpenAI) 伴随论文 [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) 由 Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever 发布。
-1. **[GPT Neo](https://huggingface.co/docs/transformers/model_doc/gpt_neo)** (来自 EleutherAI) 随仓库 [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) 发布。作者为 Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy 发布。
-1. **[GPT NeoX](https://huggingface.co/docs/transformers/model_doc/gpt_neox)** (from EleutherAI) released with the paper [GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://arxiv.org/abs/2204.06745) by Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, Samuel Weinbach 
-1. **[GPT-2](https://huggingface.co/docs/transformers/model_doc/gpt2)** (来自 OpenAI) 伴随论文 [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) 由 Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever** 发布。
-1. **[GPT-J](https://huggingface.co/docs/transformers/model_doc/gptj)** (来自 EleutherAI) 伴随论文 [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) 由 Ben Wang and Aran Komatsuzaki 发布。
-1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (来自 UCSD, NVIDIA) 伴随论文 [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) 由 Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang 发布。
-1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (来自 Facebook) 伴随论文 [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) 由 Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed 发布。
-1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (来自 Berkeley) 伴随论文 [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) 由 Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer 发布。
-1. **[ImageGPT](https://huggingface.co/docs/transformers/model_doc/imagegpt)** (来自 OpenAI) 伴随论文 [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) 由 Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever 发布。
-1. **[LayoutLM](https://huggingface.co/docs/transformers/model_doc/layoutlm)** (来自 Microsoft Research Asia) 伴随论文 [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) 由 Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou 发布。
-1. **[LayoutLMv2](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (来自 Microsoft Research Asia) 伴随论文 [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) 由 Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou 发布。
-1. **[LayoutLMv3](https://huggingface.co/docs/transformers/model_doc/layoutlmv3)** (来自 Microsoft Research Asia) 伴随论文 [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) 由 Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei 发布。
-1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (来自 Microsoft Research Asia) 伴随论文 [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) 由 Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei 发布。
-1. **[LED](https://huggingface.co/docs/transformers/model_doc/led)** (来自 AllenAI) 伴随论文 [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) 由 Iz Beltagy, Matthew E. Peters, Arman Cohan 发布。
-1. **[LeViT](https://huggingface.co/docs/transformers/model_doc/levit)** (来自 Meta AI) 伴随论文 [LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference](https://arxiv.org/abs/2104.01136) 由 Ben Graham, Alaaeldin El-Nouby, Hugo Touvron, Pierre Stock, Armand Joulin, Hervé Jégou, Matthijs Douze 发布。
-1. **[Longformer](https://huggingface.co/docs/transformers/model_doc/longformer)** (来自 AllenAI) 伴随论文 [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) 由 Iz Beltagy, Matthew E. Peters, Arman Cohan 发布。
-1. **[LongT5](https://huggingface.co/docs/transformers/model_doc/longt5)** (来自 Google AI) released 伴随论文 [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://arxiv.org/abs/2112.07916) 由 Mandy Guo, Joshua Ainslie, David Uthus, Santiago Ontanon, Jianmo Ni, Yun-Hsuan Sung, Yinfei Yang 发布。
-1. **[LUKE](https://huggingface.co/docs/transformers/model_doc/luke)** (来自 Studio Ousia) 伴随论文 [LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention](https://arxiv.org/abs/2010.01057) 由 Ikuya Yamada, Akari Asai, Hiroyuki Shindo, Hideaki Takeda, Yuji Matsumoto 发布。
-1. **[LXMERT](https://huggingface.co/docs/transformers/model_doc/lxmert)** (来自 UNC Chapel Hill) 伴随论文 [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) 由 Hao Tan and Mohit Bansal 发布。
-1. **[M-CTC-T](https://huggingface.co/docs/transformers/model_doc/mctct)** (来自 Facebook) 伴随论文 [Pseudo-Labeling For Massively Multilingual Speech Recognition](https://arxiv.org/abs/2111.00161) 由 Loren Lugosch, Tatiana Likhomanenko, Gabriel Synnaeve, and Ronan Collobert 发布。
-1. **[M2M100](https://huggingface.co/docs/transformers/model_doc/m2m_100)** (来自 Facebook) 伴随论文 [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) 由 Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin 发布。
-1. **[MarianMT](https://huggingface.co/docs/transformers/model_doc/marian)** 用 [OPUS](http://opus.nlpl.eu/) 数据训练的机器翻译模型由 Jörg Tiedemann 发布。[Marian Framework](https://marian-nmt.github.io/) 由微软翻译团队开发。
-1. **[MaskFormer](https://huggingface.co/docs/transformers/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov 
-1. **[mBART](https://huggingface.co/docs/transformers/model_doc/mbart)** (来自 Facebook) 伴随论文 [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) 由 Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer 发布。
-1. **[mBART-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (来自 Facebook) 伴随论文 [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) 由 Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan 发布。
-1. **[Megatron-BERT](https://huggingface.co/docs/transformers/model_doc/megatron-bert)** (来自 NVIDIA) 伴随论文 [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) 由 Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro 发布。
-1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (来自 NVIDIA) 伴随论文 [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) 由 Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro 发布。
-1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (来自 Studio Ousia) 伴随论文 [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) 由 Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka 发布。
-1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (来自 CMU/Google Brain) 伴随论文 [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) 由 Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou 发布。
-1. **[MobileViT](https://huggingface.co/docs/transformers/model_doc/mobilevit)** (来自 Apple) 伴随论文 [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) 由 Sachin Mehta and Mohammad Rastegari 发布。
-1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (来自 Microsoft Research) 伴随论文 [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) 由 Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu 发布。
-1. **[MT5](https://huggingface.co/docs/transformers/model_doc/mt5)** (来自 Google AI) 伴随论文 [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) 由 Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel 发布。
-1. **[MVP](https://huggingface.co/docs/transformers/model_doc/mvp)** (来自 中国人民大学 AI Box) 伴随论文 [MVP: Multi-task Supervised Pre-training for Natural Language Generation](https://arxiv.org/abs/2206.12131) 由 Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen 发布。 
-1. **[Nezha](https://huggingface.co/docs/transformers/model_doc/nezha)** (来自华为诺亚方舟实验室) 伴随论文 [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://arxiv.org/abs/1909.00204) 由 Junqiu Wei, Xiaozhe Ren, Xiaoguang Li, Wenyong Huang, Yi Liao, Yasheng Wang, Jiashu Lin, Xin Jiang, Xiao Chen and Qun Liu 发布。
-1. **[NLLB](https://huggingface.co/docs/transformers/model_doc/nllb)** (来自 Meta) 伴随论文 [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) 由 the NLLB team 发布。
-1. **[Nyströmformer](https://huggingface.co/docs/transformers/model_doc/nystromformer)** (来自 the University of Wisconsin - Madison) 伴随论文 [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) 由 Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh 发布。
-1. **[OPT](https://huggingface.co/docs/transformers/master/model_doc/opt)** (来自 Meta AI) 伴随论文 [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) 由 Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al 发布。
-1. **[OWL-ViT](https://huggingface.co/docs/transformers/model_doc/owlvit)** (来自 Google AI) 伴随论文 [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) 由 Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby 发布。
-1. **[Pegasus](https://huggingface.co/docs/transformers/model_doc/pegasus)** (来自 Google) 伴随论文 [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) 由 Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu 发布。
-1. **[Perceiver IO](https://huggingface.co/docs/transformers/model_doc/perceiver)** (来自 Deepmind) 伴随论文 [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) 由 Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira 发布。
-1. **[PhoBERT](https://huggingface.co/docs/transformers/model_doc/phobert)** (来自 VinAI Research) 伴随论文 [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) 由 Dat Quoc Nguyen and Anh Tuan Nguyen 发布。
-1. **[PLBart](https://huggingface.co/docs/transformers/model_doc/plbart)** (来自 UCLA NLP) 伴随论文 [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) 由 Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang 发布。
-1. **[PoolFormer](https://huggingface.co/docs/transformers/model_doc/poolformer)** (来自 Sea AI Labs) 伴随论文 [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) 由 Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng 发布。
-1. **[ProphetNet](https://huggingface.co/docs/transformers/model_doc/prophetnet)** (来自 Microsoft Research) 伴随论文 [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) 由 Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou 发布。
-1. **[QDQBert](https://huggingface.co/docs/transformers/model_doc/qdqbert)** (来自 NVIDIA) 伴随论文 [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) 由 Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius 发布。
-1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (来自 Facebook) 伴随论文 [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) 由 Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela 发布。
-1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (来自 Google Research) 伴随论文 [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) 由 Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang 发布。
-1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (来自 Google Research) 伴随论文 [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) 由 Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya 发布。
-1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Research) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár. 
-1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (来自 Google Research) 伴随论文 [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/pdf/2010.12821.pdf) 由 Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder 发布。
-1. **[ResNet](https://huggingface.co/docs/transformers/model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun. 
-1. **[RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta)** (来自 Facebook), 伴随论文 [Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) 由 Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov 发布。
-1. **[RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer)** (来自 ZhuiyiTechnology), 伴随论文 [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/pdf/2104.09864v1.pdf) 由 Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu 发布。
-1. **[SegFormer](https://huggingface.co/docs/transformers/model_doc/segformer)** (来自 NVIDIA) 伴随论文 [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) 由 Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo 发布。
-1. **[SEW](https://huggingface.co/docs/transformers/model_doc/sew)** (来自 ASAPP) 伴随论文 [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) 由 Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi 发布。
-1. **[SEW-D](https://huggingface.co/docs/transformers/model_doc/sew_d)** (来自 ASAPP) 伴随论文 [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) 由 Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi 发布。
-1. **[SpeechToTextTransformer](https://huggingface.co/docs/transformers/model_doc/speech_to_text)** (来自 Facebook), 伴随论文 [fairseq S2T: Fast Speech-to-Text Modeling with fairseq](https://arxiv.org/abs/2010.05171) 由 Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Dmytro Okhonko, Juan Pino 发布。
-1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (来自 Facebook) 伴随论文 [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) 由 Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau 发布。
-1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (来自 Tel Aviv University) 伴随论文 [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) 由 Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy 发布。
-1. **[SqueezeBERT](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (来自 Berkeley) 伴随论文 [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) 由 Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer 发布。
-1. **[Swin Transformer](https://huggingface.co/docs/transformers/model_doc/swin)** (来自 Microsoft) 伴随论文 [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) 由 Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo 发布。
-1. **[Swin Transformer V2](https://huggingface.co/docs/transformers/main/model_doc/swinv2)** (来自 Microsoft) 伴随论文 [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) 由 Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo 发布。
-1. **[T5](https://huggingface.co/docs/transformers/model_doc/t5)** (来自 Google AI) 伴随论文 [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) 由 Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu 发布。
-1. **[T5v1.1](https://huggingface.co/docs/transformers/model_doc/t5v1.1)** (来自 Google AI) 伴随论文 [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) 由 Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu 发布。
-1. **[TAPAS](https://huggingface.co/docs/transformers/model_doc/tapas)** (来自 Google AI) 伴随论文 [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) 由 Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos 发布。
-1. **[TAPEX](https://huggingface.co/docs/transformers/model_doc/tapex)** (来自 Microsoft Research) 伴随论文 [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) 由 Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou 发布。
-1. **[Trajectory Transformer](https://huggingface.co/docs/transformers/model_doc/trajectory_transformers)** (from the University of California at Berkeley) released with the paper [Offline Reinforcement Learning as One Big Sequence Modeling Problem](https://arxiv.org/abs/2106.02039) by Michael Janner, Qiyang Li, Sergey Levine 
-1. **[Transformer-XL](https://huggingface.co/docs/transformers/model_doc/transfo-xl)** (来自 Google/CMU) 伴随论文 [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) 由 Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov 发布。
-1. **[TrOCR](https://huggingface.co/docs/transformers/model_doc/trocr)** (来自 Microsoft) 伴随论文 [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) 由 Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei 发布。
-1. **[UL2](https://huggingface.co/docs/transformers/model_doc/ul2)** (from Google Research) released with the paper [Unifying Language Learning Paradigms](https://arxiv.org/abs/2205.05131v1) by Yi Tay, Mostafa Dehghani, Vinh Q. Tran, Xavier Garcia, Dara Bahri, Tal Schuster, Huaixiu Steven Zheng, Neil Houlsby, Donald Metzler 
-1. **[UniSpeech](https://huggingface.co/docs/transformers/model_doc/unispeech)** (来自 Microsoft Research) 伴随论文 [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) 由 Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang 发布。
-1. **[UniSpeechSat](https://huggingface.co/docs/transformers/model_doc/unispeech-sat)** (来自 Microsoft Research) 伴随论文 [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) 由 Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu 发布。
-1. **[VAN](https://huggingface.co/docs/transformers/model_doc/van)** (来自 Tsinghua University and Nankai University) 伴随论文 [Visual Attention Network](https://arxiv.org/pdf/2202.09741.pdf) 由 Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu 发布。
-1. **[VideoMAE](https://huggingface.co/docs/transformers/main/model_doc/videomae)** (来自 Multimedia Computing Group, Nanjing University) 伴随论文 [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://arxiv.org/abs/2203.12602) 由 Zhan Tong, Yibing Song, Jue Wang, Limin Wang 发布。
-1. **[ViLT](https://huggingface.co/docs/transformers/model_doc/vilt)** (来自 NAVER AI Lab/Kakao Enterprise/Kakao Brain) 伴随论文 [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) 由 Wonjae Kim, Bokyung Son, Ildoo Kim 发布。
-1. **[Vision Transformer (ViT)](https://huggingface.co/docs/transformers/model_doc/vit)** (来自 Google AI) 伴随论文 [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) 由 Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby 发布。
-1. **[VisualBERT](https://huggingface.co/docs/transformers/model_doc/visual_bert)** (来自 UCLA NLP) 伴随论文 [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) 由 Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang 发布。
-1. **[ViTMAE](https://huggingface.co/docs/transformers/model_doc/vit_mae)** (来自 Meta AI) 伴随论文 [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) 由 Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick 发布。
-1. **[Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/wav2vec2)** (来自 Facebook AI) 伴随论文 [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) 由 Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli 发布。
-1. **[Wav2Vec2-Conformer](https://huggingface.co/docs/transformers/model_doc/wav2vec2-conformer)** (来自 Facebook AI) 伴随论文 [FAIRSEQ S2T: Fast Speech-to-Text Modeling with FAIRSEQ](https://arxiv.org/abs/2010.05171) 由 Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Sravya Popuri, Dmytro Okhonko, Juan Pino 发布。
-1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/transformers/model_doc/wav2vec2_phoneme)** (来自 Facebook AI) 伴随论文 [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) 由 Qiantong Xu, Alexei Baevski, Michael Auli 发布。
-1. **[WavLM](https://huggingface.co/docs/transformers/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
-1. **[XGLM](https://huggingface.co/docs/transformers/model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li. 
-1. **[XLM](https://huggingface.co/docs/transformers/model_doc/xlm)** (来自 Facebook) 伴随论文 [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) 由 Guillaume Lample and Alexis Conneau 发布。
-1. **[XLM-ProphetNet](https://huggingface.co/docs/transformers/model_doc/xlm-prophetnet)** (来自 Microsoft Research) 伴随论文 [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) 由 Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou 发布。
-1. **[XLM-RoBERTa](https://huggingface.co/docs/transformers/model_doc/xlm-roberta)** (来自 Facebook AI), 伴随论文 [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) 由 Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov 发布。
-1. **[XLM-RoBERTa-XL](https://huggingface.co/docs/transformers/model_doc/xlm-roberta-xl)** (来自 Facebook AI) 伴随论文 [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) 由 Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau 发布。
-1. **[XLNet](https://huggingface.co/docs/transformers/model_doc/xlnet)** (来自 Google/CMU) 伴随论文 [XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) 由 Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le 发布。
-1. **[XLS-R](https://huggingface.co/docs/transformers/model_doc/xls_r)** (来自 Facebook AI) 伴随论文 [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) 由 Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli 发布。
-1. **[XLSR-Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/xlsr_wav2vec2)** (来自 Facebook AI) 伴随论文 [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) 由 Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli 发布。
-1. **[YOLOS](https://huggingface.co/docs/transformers/model_doc/yolos)** (来自 Huazhong University of Science & Technology) 伴随论文 [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) 由 Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu 发布。
-1. **[YOSO](https://huggingface.co/docs/transformers/model_doc/yoso)** (来自 the University of Wisconsin - Madison) 伴随论文 [You Only Sample (Almost) 由 Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh 发布。
-1. 想要贡献新的模型？我们这里有一份**详细指引和模板**来引导你添加新的模型。你可以在 [`templates`](./templates) 目录中找到他们。记得查看 [贡献指南](./CONTRIBUTING.md) 并在开始写 PR 前联系维护人员或开一个新的 issue 来获得反馈。
-
-要检查某个模型是否已有 Flax、PyTorch 或 TensorFlow 的实现，或其是否在 🤗 Tokenizers 库中有对应词符化器（tokenizer），敬请参阅[此表](https://huggingface.co/docs/transformers/index#supported-frameworks)。
-
-这些实现均已于多个数据集测试（请参看用例脚本）并应于原版实现表现相当。你可以在用例文档的[此节](https://huggingface.co/docs/transformers/examples)中了解表现的细节。
-
-
-## 了解更多
-
-| 章节 | 描述 |
-|-|-|
-| [文档](https://huggingface.co/transformers/) | 完整的 API 文档和教程 |
-| [任务总结](https://huggingface.co/docs/transformers/task_summary) | 🤗 Transformers 支持的任务 |
-| [预处理教程](https://huggingface.co/docs/transformers/preprocessing) | 使用 `Tokenizer` 来为模型准备数据 |
-| [训练和微调](https://huggingface.co/docs/transformers/training) | 在 PyTorch/TensorFlow 的训练循环或 `Trainer` API 中使用 🤗 Transformers 提供的模型 |
-| [快速上手：微调和用例脚本](https://github.com/huggingface/transformers/tree/main/examples) | 为各种任务提供的用例脚本 |
-| [模型分享和上传](https://huggingface.co/docs/transformers/model_sharing) | 和社区上传和分享你微调的模型 |
-| [迁移](https://huggingface.co/docs/transformers/migration) | 从 `pytorch-transformers` 或 `pytorch-pretrained-bert` 迁移到 🤗 Transformers |
-
-## 引用
-
-我们已将此库的[论文](https://www.aclweb.org/anthology/2020.emnlp-demos.6/)正式发表，如果你使用了 🤗 Transformers 库，请引用:
-```bibtex
-@inproceedings{wolf-etal-2020-transformers,
-    title = "Transformers: State-of-the-Art Natural Language Processing",
-    author = "Thomas Wolf and Lysandre Debut and Victor Sanh and Julien Chaumond and Clement Delangue and Anthony Moi and Pierric Cistac and Tim Rault and Rémi Louf and Morgan Funtowicz and Joe Davison and Sam Shleifer and Patrick von Platen and Clara Ma and Yacine Jernite and Julien Plu and Canwen Xu and Teven Le Scao and Sylvain Gugger and Mariama Drame and Quentin Lhoest and Alexander M. Rush",
-    booktitle = "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations",
-    month = oct,
-    year = "2020",
-    address = "Online",
-    publisher = "Association for Computational Linguistics",
-    url = "https://www.aclweb.org/anthology/2020.emnlp-demos.6",
-    pages = "38--45"
-}
-```

README_zh-hant.md

@@ -1,416 +0,0 @@
-<!---
-Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
--->
-
-<!---
-A useful guide for English-Traditional Chinese translation of Hugging Face documentation
-- Add space around English words and numbers when they appear between Chinese characters. E.g., 共 100 多種語言; 使用 transformers 函式庫。
-- Use square quotes, e.g.,「引用」
-- Some of terms in the file can be found at National Academy for Educational Research (https://terms.naer.edu.tw/), an official website providing bilingual translations between English and Traditional Chinese.
-
-Dictionary
-
-API: API (不翻譯）
-add: 加入
-checkpoint: 檢查點
-code: 程式碼
-community: 社群
-confidence: 信賴度
-dataset: 資料集
-documentation: 文件
-example: 基本翻譯為「範例」，或依語意翻為「例子」
-finetune: 微調
-Hugging Face: Hugging Face（不翻譯）
-implementation: 實作
-inference: 推論
-library: 函式庫
-module: 模組
-NLP/Natural Language Processing: 以 NLP 出現時不翻譯，以 Natural Language Processing 出現時翻譯為自然語言處理
-online demos: 線上Demo
-pipeline: pipeline（不翻譯）
-pretrained/pretrain: 預訓練
-Python data structures (e.g., list, set, dict): 翻譯為串列，集合，字典，並用括號標註原英文
-repository: repository（不翻譯）
-summary: 概覽
-token-: token-（不翻譯）
-Trainer: Trainer（不翻譯）
-transformer: transformer（不翻譯）
-tutorial: 教學
-user: 使用者
--->
-
-<p align="center">
-    <br>
-    <img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers_logo_name.png" width="400"/>
-    <br>
-<p>
-<p align="center">
-    <a href="https://circleci.com/gh/huggingface/transformers">
-        <img alt="Build" src="https://img.shields.io/circleci/build/github/huggingface/transformers/main">
-    </a>
-    <a href="https://github.com/huggingface/transformers/blob/main/LICENSE">
-        <img alt="GitHub" src="https://img.shields.io/github/license/huggingface/transformers.svg?color=blue">
-    </a>
-    <a href="https://huggingface.co/docs/transformers/index">
-        <img alt="Documentation" src="https://img.shields.io/website/http/huggingface.co/docs/transformers/index.svg?down_color=red&down_message=offline&up_message=online">
-    </a>
-    <a href="https://github.com/huggingface/transformers/releases">
-        <img alt="GitHub release" src="https://img.shields.io/github/release/huggingface/transformers.svg">
-    </a>
-    <a href="https://github.com/huggingface/transformers/blob/main/CODE_OF_CONDUCT.md">
-        <img alt="Contributor Covenant" src="https://img.shields.io/badge/Contributor%20Covenant-v2.0%20adopted-ff69b4.svg">
-    </a>
-    <a href="https://zenodo.org/badge/latestdoi/155220641"><img src="https://zenodo.org/badge/155220641.svg" alt="DOI"></a>
-</p>
-
-<h4 align="center">
-    <p>
-        <a href="https://github.com/huggingface/transformers/">English</a> |
-        <a href="https://github.com/huggingface/transformers/blob/main/README_zh-hans.md">简体中文</a> |
-        <b>繁體中文</b> |
-        <a href="https://github.com/huggingface/transformers/blob/main/README_ko.md">한국어</a>
-    <p>
-</h4>
-
-<h3 align="center">
-    <p>為 Jax、PyTorch 以及 TensorFlow 打造的先進自然語言處理函式庫</p>
-</h3>
-
-<h3 align="center">
-    <a href="https://hf.co/course"><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/course_banner.png"></a>
-</h3>
-
-🤗 Transformers 提供了數以千計的預訓練模型，支援 100 多種語言的文本分類、資訊擷取、問答、摘要、翻譯、文本生成。它的宗旨是讓最先進的 NLP 技術人人易用。
-
-🤗 Transformers 提供了便於快速下載和使用的API，讓你可以將預訓練模型用在給定文本、在你的資料集上微調然後經由 [model hub](https://huggingface.co/models) 與社群共享。同時，每個定義的 Python 模組架構均完全獨立，方便修改和快速研究實驗。
-
-🤗 Transformers 支援三個最熱門的深度學習函式庫： [Jax](https://jax.readthedocs.io/en/latest/), [PyTorch](https://pytorch.org/) 以及 [TensorFlow](https://www.tensorflow.org/) — 並與之完美整合。你可以直接使用其中一個框架訓練你的模型，然後用另一個載入和推論。
-
-## 線上Demo
-
-你可以直接在 [model hub](https://huggingface.co/models) 上測試大多數的模型。我們也提供了 [私有模型託管、模型版本管理以及推論API](https://huggingface.co/pricing)。
-
-這裡是一些範例：
-- [用 BERT 做遮蓋填詞](https://huggingface.co/bert-base-uncased?text=Paris+is+the+%5BMASK%5D+of+France)
-- [用 Electra 做專有名詞辨識](https://huggingface.co/dbmdz/electra-large-discriminator-finetuned-conll03-english?text=My+name+is+Sarah+and+I+live+in+London+city)
-- [用 GPT-2 做文本生成](https://huggingface.co/gpt2?text=A+long+time+ago%2C+)
-- [用 RoBERTa 做自然語言推論](https://huggingface.co/roberta-large-mnli?text=The+dog+was+lost.+Nobody+lost+any+animal)
-- [用 BART 做文本摘要](https://huggingface.co/facebook/bart-large-cnn?text=The+tower+is+324+metres+%281%2C063+ft%29+tall%2C+about+the+same+height+as+an+81-storey+building%2C+and+the+tallest+structure+in+Paris.+Its+base+is+square%2C+measuring+125+metres+%28410+ft%29+on+each+side.+During+its+construction%2C+the+Eiffel+Tower+surpassed+the+Washington+Monument+to+become+the+tallest+man-made+structure+in+the+world%2C+a+title+it+held+for+41+years+until+the+Chrysler+Building+in+New+York+City+was+finished+in+1930.+It+was+the+first+structure+to+reach+a+height+of+300+metres.+Due+to+the+addition+of+a+broadcasting+aerial+at+the+top+of+the+tower+in+1957%2C+it+is+now+taller+than+the+Chrysler+Building+by+5.2+metres+%2817+ft%29.+Excluding+transmitters%2C+the+Eiffel+Tower+is+the+second+tallest+free-standing+structure+in+France+after+the+Millau+Viaduct)
-- [用 DistilBERT 做問答](https://huggingface.co/distilbert-base-uncased-distilled-squad?text=Which+name+is+also+used+to+describe+the+Amazon+rainforest+in+English%3F&context=The+Amazon+rainforest+%28Portuguese%3A+Floresta+Amaz%C3%B4nica+or+Amaz%C3%B4nia%3B+Spanish%3A+Selva+Amaz%C3%B3nica%2C+Amazon%C3%ADa+or+usually+Amazonia%3B+French%3A+For%C3%AAt+amazonienne%3B+Dutch%3A+Amazoneregenwoud%29%2C+also+known+in+English+as+Amazonia+or+the+Amazon+Jungle%2C+is+a+moist+broadleaf+forest+that+covers+most+of+the+Amazon+basin+of+South+America.+This+basin+encompasses+7%2C000%2C000+square+kilometres+%282%2C700%2C000+sq+mi%29%2C+of+which+5%2C500%2C000+square+kilometres+%282%2C100%2C000+sq+mi%29+are+covered+by+the+rainforest.+This+region+includes+territory+belonging+to+nine+nations.+The+majority+of+the+forest+is+contained+within+Brazil%2C+with+60%25+of+the+rainforest%2C+followed+by+Peru+with+13%25%2C+Colombia+with+10%25%2C+and+with+minor+amounts+in+Venezuela%2C+Ecuador%2C+Bolivia%2C+Guyana%2C+Suriname+and+French+Guiana.+States+or+departments+in+four+nations+contain+%22Amazonas%22+in+their+names.+The+Amazon+represents+over+half+of+the+planet%27s+remaining+rainforests%2C+and+comprises+the+largest+and+most+biodiverse+tract+of+tropical+rainforest+in+the+world%2C+with+an+estimated+390+billion+individual+trees+divided+into+16%2C000+species)
-- [用 T5 做翻譯](https://huggingface.co/t5-base?text=My+name+is+Wolfgang+and+I+live+in+Berlin)
-
-**[Write With Transformer](https://transformer.huggingface.co)**，由 Hugging Face 團隊所打造，是一個文本生成的官方 demo。
-
-## 如果你在尋找由 Hugging Face 團隊所提供的客製化支援服務
-
-<a target="_blank" href="https://huggingface.co/support">
-    <img alt="HuggingFace Expert Acceleration Program" src="https://huggingface.co/front/thumbnails/support.png" style="max-width: 600px; border: 1px solid #eee; border-radius: 4px; box-shadow: 0 1px 2px 0 rgba(0, 0, 0, 0.05);">
-</a><br>
-
-## 快速上手
-
-我們為快速使用模型提供了 `pipeline` API。 Pipeline 包含了預訓練模型和對應的文本預處理。下面是一個快速使用 pipeline 去判斷正負面情緒的例子：
-
-```python
->>> from transformers import pipeline
-
-# 使用情緒分析 pipeline
->>> classifier = pipeline('sentiment-analysis')
->>> classifier('We are very happy to introduce pipeline to the transformers repository.')
-[{'label': 'POSITIVE', 'score': 0.9996980428695679}]
-```
-
-第二行程式碼下載並快取 pipeline 使用的預訓練模型，而第三行程式碼則在給定的文本上進行了評估。這裡的答案“正面” (positive) 具有 99.97% 的信賴度。
-
-許多的 NLP 任務都有隨選即用的預訓練 `pipeline`。例如，我們可以輕鬆地從給定文本中擷取問題答案：
-
-``` python
->>> from transformers import pipeline
-
-# 使用問答 pipeline
->>> question_answerer = pipeline('question-answering')
->>> question_answerer({
-...     'question': 'What is the name of the repository ?',
-...     'context': 'Pipeline has been included in the huggingface/transformers repository'
-... })
-{'score': 0.30970096588134766, 'start': 34, 'end': 58, 'answer': 'huggingface/transformers'}
-
-```
-
-除了提供問題解答，預訓練模型還提供了對應的信賴度分數以及解答在 tokenized 後的文本中開始和結束的位置。你可以從[這個教學](https://huggingface.co/docs/transformers/task_summary)了解更多 `pipeline` API支援的任務。
-
-要在你的任務中下載和使用任何預訓練模型很簡單，只需三行程式碼。這裡是 PyTorch 版的範例：
-```python
->>> from transformers import AutoTokenizer, AutoModel
-
->>> tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
->>> model = AutoModel.from_pretrained("bert-base-uncased")
-
->>> inputs = tokenizer("Hello world!", return_tensors="pt")
->>> outputs = model(**inputs)
-```
-這裡是對應的 TensorFlow 程式碼：
-```python
->>> from transformers import AutoTokenizer, TFAutoModel
-
->>> tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
->>> model = TFAutoModel.from_pretrained("bert-base-uncased")
-
->>> inputs = tokenizer("Hello world!", return_tensors="tf")
->>> outputs = model(**inputs)
-```
-
-Tokenizer 為所有的預訓練模型提供了預處理，並可以直接轉換單一字串（比如上面的例子）或串列 (list)。它會輸出一個的字典 (dict) 讓你可以在下游程式碼裡使用或直接藉由 `**` 運算式傳給模型。
-
-模型本身是一個常規的 [Pytorch `nn.Module`](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) 或 [TensorFlow `tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model)（取決於你的後端），可依常規方式使用。 [這個教學](https://huggingface.co/transformers/training.html)解釋了如何將這樣的模型整合到一般的 PyTorch 或 TensorFlow 訓練迴圈中，或是如何使用我們的 `Trainer` API 在一個新的資料集上快速進行微調。
-
-## 為什麼要用 transformers？
-
-1. 便於使用的先進模型：
-    - NLU 和 NLG 上性能卓越
-    - 對教學和實作友好且低門檻
-    - 高度抽象，使用者只須學習 3 個類別
-    - 對所有模型使用的制式化API
-
-1. 更低的運算成本，更少的碳排放：
-    - 研究人員可以分享預訓練的模型而非從頭開始訓練
-    - 工程師可以減少計算時間以及生產成本
-    - 數十種模型架構、兩千多個預訓練模型、100多種語言支援
-
-1. 對於模型生命週期的每一個部分都面面俱到：
-    - 訓練先進的模型，只需 3 行程式碼
-    - 模型可以在不同深度學習框架之間任意轉換
-    - 為訓練、評估和生產選擇最適合的框架，並完美銜接
-
-1. 為你的需求輕鬆客製化專屬模型和範例：
-    - 我們為每種模型架構提供了多個範例來重現原論文結果
-    - 一致的模型內部架構
-    - 模型檔案可單獨使用，便於修改和快速實驗
-
-## 什麼情況下我不該用 transformers？
-
-- 本函式庫並不是模組化的神經網絡工具箱。模型文件中的程式碼並未做額外的抽象封裝，以便研究人員快速地翻閱及修改程式碼，而不會深陷複雜的類別包裝之中。
-- `Trainer` API 並非相容任何模型，它只為本函式庫中的模型最佳化。對於一般的機器學習用途，請使用其他函式庫。
-- 儘管我們已盡力而為，[examples 目錄](https://github.com/huggingface/transformers/tree/main/examples)中的腳本也僅為範例而已。對於特定問題，它們並不一定隨選即用，可能需要修改幾行程式碼以符合需求。
-
-## 安裝
-
-### 使用 pip
-
-這個 Repository 已在 Python 3.6+、Flax 0.3.2+、PyTorch 1.3.1+ 和 TensorFlow 2.3+ 下經過測試。
-
-你可以在[虛擬環境](https://docs.python.org/3/library/venv.html)中安裝 🤗 Transformers。如果你還不熟悉 Python 的虛擬環境，請閱此[使用者指引](https://packaging.python.org/guides/installing-using-pip-and-virtual-environments/)。
-
-首先，用你打算使用的版本的 Python 創建一個虛擬環境並進入。
-
-然後，你需要安裝 Flax、PyTorch 或 TensorFlow 其中之一。對於該如何在你使用的平台上安裝這些框架，請參閱 [TensorFlow 安裝頁面](https://www.tensorflow.org/install/), [PyTorch 安裝頁面](https://pytorch.org/get-started/locally/#start-locally) 或 [Flax 安裝頁面](https://github.com/google/flax#quick-install)。
-
-當其中一個後端安裝成功後，🤗 Transformers 可依此安裝：
-
-```bash
-pip install transformers
-```
-
-如果你想要試試範例或者想在正式發布前使用最新開發中的程式碼，你必須[從原始碼安裝](https://huggingface.co/docs/transformers/installation#installing-from-source)。
-
-### 使用 conda
-
-自 Transformers 4.0.0 版始，我們有了一個 conda channel： `huggingface`。
-
-🤗 Transformers 可以藉由 conda 依此安裝：
-
-```shell script
-conda install -c huggingface transformers
-```
-
-要藉由 conda 安裝 Flax、PyTorch 或 TensorFlow 其中之一，請參閱它們各自安裝頁面的說明。
-
-## 模型架構
-
-**🤗 Transformers 支援的[所有的模型檢查點](https://huggingface.co/models)**，由[使用者](https://huggingface.co/users)和[組織](https://huggingface.co/organizations)上傳，均與 huggingface.co [model hub](https://huggingface.co) 完美結合。
-
-目前的檢查點數量： ![](https://img.shields.io/endpoint?url=https://huggingface.co/api/shields/models&color=brightgreen)
-
-🤗 Transformers 目前支援以下的架構（模型概覽請參閱[這裡](https://huggingface.co/docs/transformers/model_summary)）：
-
-1. **[ALBERT](https://huggingface.co/docs/transformers/model_doc/albert)** (from Google Research and the Toyota Technological Institute at Chicago) released with the paper [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://arxiv.org/abs/1909.11942), by Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut.
-1. **[BART](https://huggingface.co/docs/transformers/model_doc/bart)** (from Facebook) released with the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/pdf/1910.13461.pdf) by Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Ves Stoyanov and Luke Zettlemoyer.
-1. **[BARThez](https://huggingface.co/docs/transformers/model_doc/barthez)** (from École polytechnique) released with the paper [BARThez: a Skilled Pretrained French Sequence-to-Sequence Model](https://arxiv.org/abs/2010.12321) by Moussa Kamal Eddine, Antoine J.-P. Tixier, Michalis Vazirgiannis.
-1. **[BARTpho](https://huggingface.co/docs/transformers/model_doc/bartpho)** (from VinAI Research) released with the paper [BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese](https://arxiv.org/abs/2109.09701) by Nguyen Luong Tran, Duong Minh Le and Dat Quoc Nguyen.
-1. **[BEiT](https://huggingface.co/docs/transformers/model_doc/beit)** (from Microsoft) released with the paper [BEiT: BERT Pre-Training of Image Transformers](https://arxiv.org/abs/2106.08254) by Hangbo Bao, Li Dong, Furu Wei.
-1. **[BERT](https://huggingface.co/docs/transformers/model_doc/bert)** (from Google) released with the paper [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://arxiv.org/abs/1810.04805) by Jacob Devlin, Ming-Wei Chang, Kenton Lee and Kristina Toutanova.
-1. **[BERT For Sequence Generation](https://huggingface.co/docs/transformers/model_doc/bert-generation)** (from Google) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
-1. **[BERTweet](https://huggingface.co/docs/transformers/model_doc/bertweet)** (from VinAI Research) released with the paper [BERTweet: A pre-trained language model for English Tweets](https://aclanthology.org/2020.emnlp-demos.2/) by Dat Quoc Nguyen, Thanh Vu and Anh Tuan Nguyen.
-1. **[BigBird-Pegasus](https://huggingface.co/docs/transformers/model_doc/bigbird_pegasus)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
-1. **[BigBird-RoBERTa](https://huggingface.co/docs/transformers/model_doc/big_bird)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
-1. **[Blenderbot](https://huggingface.co/docs/transformers/model_doc/blenderbot)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
-1. **[BlenderbotSmall](https://huggingface.co/docs/transformers/model_doc/blenderbot-small)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
-1. **[BLOOM](https://huggingface.co/docs/transformers/model_doc/bloom)** (from BigScience workshop) released by the [BigSicence Workshop](https://bigscience.huggingface.co/). 
-1. **[BORT](https://huggingface.co/docs/transformers/model_doc/bort)** (from Alexa) released with the paper [Optimal Subarchitecture Extraction For BERT](https://arxiv.org/abs/2010.10499) by Adrian de Wynter and Daniel J. Perry.
-1. **[ByT5](https://huggingface.co/docs/transformers/model_doc/byt5)** (from Google Research) released with the paper [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) by Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel.
-1. **[CamemBERT](https://huggingface.co/docs/transformers/model_doc/camembert)** (from Inria/Facebook/Sorbonne) released with the paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) by Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suárez*, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah and Benoît Sagot.
-1. **[CANINE](https://huggingface.co/docs/transformers/model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
-1. **[CLIP](https://huggingface.co/docs/transformers/model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
-1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
-1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
-1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
-1. **[CPM](https://huggingface.co/docs/transformers/model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
-1. **[CTRL](https://huggingface.co/docs/transformers/model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
-1. **[CvT](https://huggingface.co/docs/transformers/model_doc/cvt)** (from Microsoft) released with the paper [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) by Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang.
-1. **[Data2Vec](https://huggingface.co/docs/transformers/model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
-1. **[DeBERTa](https://huggingface.co/docs/transformers/model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
-1. **[DeBERTa-v2](https://huggingface.co/docs/transformers/model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
-1. **[Decision Transformer](https://huggingface.co/docs/transformers/model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
-1. **[DeiT](https://huggingface.co/docs/transformers/model_doc/deit)** (from Facebook) released with the paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) by Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou.
-1. **[DETR](https://huggingface.co/docs/transformers/model_doc/detr)** (from Facebook) released with the paper [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) by Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko.
-1. **[DialoGPT](https://huggingface.co/docs/transformers/model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
-1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/distillation) and a German version of DistilBERT.
-1. **[DiT](https://huggingface.co/docs/transformers/model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
-1. **[DPR](https://huggingface.co/docs/transformers/model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
-1. **[DPT](https://huggingface.co/docs/transformers/master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by René Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
-1. **[ELECTRA](https://huggingface.co/docs/transformers/model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
-1. **[EncoderDecoder](https://huggingface.co/docs/transformers/model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
-1. **[FlauBERT](https://huggingface.co/docs/transformers/model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab.
-1. **[FLAVA](https://huggingface.co/docs/transformers/model_doc/flava)** (from Facebook AI) released with the paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) by Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela.
-1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
-1. **[Funnel Transformer](https://huggingface.co/docs/transformers/model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
-1. **[GLPN](https://huggingface.co/docs/transformers/model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
-1. **[GPT](https://huggingface.co/docs/transformers/model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
-1. **[GPT Neo](https://huggingface.co/docs/transformers/model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
-1. **[GPT NeoX](https://huggingface.co/docs/transformers/model_doc/gpt_neox)** (from EleutherAI) released with the paper [GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://arxiv.org/abs/2204.06745) by Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, Samuel Weinbach 
-1. **[GPT-2](https://huggingface.co/docs/transformers/model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
-1. **[GPT-J](https://huggingface.co/docs/transformers/model_doc/gptj)** (from EleutherAI) released with the paper [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
-1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
-1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (from Facebook) released with the paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) by Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
-1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (from Berkeley) released with the paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) by Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
-1. **[ImageGPT](https://huggingface.co/docs/transformers/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
-1. **[LayoutLM](https://huggingface.co/docs/transformers/model_doc/layoutlm)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
-1. **[LayoutLMv2](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) by Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
-1. **[LayoutLMv3](https://huggingface.co/docs/transformers/model_doc/layoutlmv3)** (from Microsoft Research Asia) released with the paper [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) by Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei.
-1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
-1. **[LED](https://huggingface.co/docs/transformers/model_doc/led)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
-1. **[LeViT](https://huggingface.co/docs/transformers/model_doc/levit)** (from Meta AI) released with the paper [LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference](https://arxiv.org/abs/2104.01136) by Ben Graham, Alaaeldin El-Nouby, Hugo Touvron, Pierre Stock, Armand Joulin, Hervé Jégou, Matthijs Douze.
-1. **[Longformer](https://huggingface.co/docs/transformers/model_doc/longformer)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
-1. **[LongT5](https://huggingface.co/docs/transformers/model_doc/longt5)** (from Google AI) released with the paper [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://arxiv.org/abs/2112.07916) by Mandy Guo, Joshua Ainslie, David Uthus, Santiago Ontanon, Jianmo Ni, Yun-Hsuan Sung, Yinfei Yang.
-1. **[LUKE](https://huggingface.co/docs/transformers/model_doc/luke)** (from Studio Ousia) released with the paper [LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention](https://arxiv.org/abs/2010.01057) by Ikuya Yamada, Akari Asai, Hiroyuki Shindo, Hideaki Takeda, Yuji Matsumoto.
-1. **[LXMERT](https://huggingface.co/docs/transformers/model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
-1. **[M-CTC-T](https://huggingface.co/docs/transformers/model_doc/mctct)** (from Facebook) released with the paper [Pseudo-Labeling For Massively Multilingual Speech Recognition](https://arxiv.org/abs/2111.00161) by Loren Lugosch, Tatiana Likhomanenko, Gabriel Synnaeve, and Ronan Collobert.
-1. **[M2M100](https://huggingface.co/docs/transformers/model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
-1. **[MarianMT](https://huggingface.co/docs/transformers/model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jörg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
-1. **[MaskFormer](https://huggingface.co/docs/transformers/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov 
-1. **[mBART](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
-1. **[mBART-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
-1. **[Megatron-BERT](https://huggingface.co/docs/transformers/model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
-1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
-1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
-1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
-1. **[MobileViT](https://huggingface.co/docs/transformers/model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
-1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
-1. **[MT5](https://huggingface.co/docs/transformers/model_doc/mt5)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
-1. **[MVP](https://huggingface.co/docs/transformers/model_doc/mvp)** (from RUC AI Box) released with the paper [MVP: Multi-task Supervised Pre-training for Natural Language Generation](https://arxiv.org/abs/2206.12131) by Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen.
-1. **[Nezha](https://huggingface.co/docs/transformers/model_doc/nezha)** (from Huawei Noah’s Ark Lab) released with the paper [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://arxiv.org/abs/1909.00204) by Junqiu Wei, Xiaozhe Ren, Xiaoguang Li, Wenyong Huang, Yi Liao, Yasheng Wang, Jiashu Lin, Xin Jiang, Xiao Chen and Qun Liu.
-1. **[NLLB](https://huggingface.co/docs/transformers/model_doc/nllb)** (from Meta) released with the paper [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) by the NLLB team.
-1. **[Nyströmformer](https://huggingface.co/docs/transformers/model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
-1. **[OPT](https://huggingface.co/docs/transformers/master/model_doc/opt)** (from Meta AI) released with the paper [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) by Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al.
-1. **[OWL-ViT](https://huggingface.co/docs/transformers/model_doc/owlvit)** (from Google AI) released with the paper [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) by Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby.
-1. **[Pegasus](https://huggingface.co/docs/transformers/model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
-1. **[Perceiver IO](https://huggingface.co/docs/transformers/model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira.
-1. **[PhoBERT](https://huggingface.co/docs/transformers/model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
-1. **[PLBart](https://huggingface.co/docs/transformers/model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
-1. **[PoolFormer](https://huggingface.co/docs/transformers/model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
-1. **[ProphetNet](https://huggingface.co/docs/transformers/model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
-1. **[QDQBert](https://huggingface.co/docs/transformers/model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
-1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
-1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
-1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
-1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Research) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár. 
-1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/pdf/2010.12821.pdf) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
-1. **[ResNet](https://huggingface.co/docs/transformers/model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun. 
-1. **[RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta)** (from Facebook), released together with the paper a [Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
-1. **[RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper a [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/pdf/2104.09864v1.pdf) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
-1. **[SegFormer](https://huggingface.co/docs/transformers/model_doc/segformer)** (from NVIDIA) released with the paper [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) by Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo.
-1. **[SEW](https://huggingface.co/docs/transformers/model_doc/sew)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
-1. **[SEW-D](https://huggingface.co/docs/transformers/model_doc/sew_d)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
-1. **[SpeechToTextTransformer](https://huggingface.co/docs/transformers/model_doc/speech_to_text)** (from Facebook), released together with the paper [fairseq S2T: Fast Speech-to-Text Modeling with fairseq](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Dmytro Okhonko, Juan Pino.
-1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (from Facebook) released with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
-1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (from Tel Aviv University) released with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
-1. **[SqueezeBERT](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
-1. **[Swin Transformer](https://huggingface.co/docs/transformers/model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
-1. **[Swin Transformer V2](https://huggingface.co/docs/transformers/main/model_doc/swinv2)** (from Microsoft) released with the paper [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) by Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo.
-1. **[T5](https://huggingface.co/docs/transformers/model_doc/t5)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
-1. **[T5v1.1](https://huggingface.co/docs/transformers/model_doc/t5v1.1)** (from Google AI) released with the paper [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
-1. **[TAPAS](https://huggingface.co/docs/transformers/model_doc/tapas)** (from Google AI) released with the paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) by Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos.
-1. **[TAPEX](https://huggingface.co/docs/transformers/model_doc/tapex)** (from Microsoft Research) released with the paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) by Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
-1. **[Trajectory Transformer](https://huggingface.co/docs/transformers/model_doc/trajectory_transformers)** (from the University of California at Berkeley) released with the paper [Offline Reinforcement Learning as One Big Sequence Modeling Problem](https://arxiv.org/abs/2106.02039) by Michael Janner, Qiyang Li, Sergey Levine 
-1. **[Transformer-XL](https://huggingface.co/docs/transformers/model_doc/transfo-xl)** (from Google/CMU) released with the paper [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) by Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov.
-1. **[TrOCR](https://huggingface.co/docs/transformers/model_doc/trocr)** (from Microsoft) released with the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
-1. **[UL2](https://huggingface.co/docs/transformers/model_doc/ul2)** (from Google Research) released with the paper [Unifying Language Learning Paradigms](https://arxiv.org/abs/2205.05131v1) by Yi Tay, Mostafa Dehghani, Vinh Q. Tran, Xavier Garcia, Dara Bahri, Tal Schuster, Huaixiu Steven Zheng, Neil Houlsby, Donald Metzler 
-1. **[UniSpeech](https://huggingface.co/docs/transformers/model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
-1. **[UniSpeechSat](https://huggingface.co/docs/transformers/model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
-1. **[VAN](https://huggingface.co/docs/transformers/model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/pdf/2202.09741.pdf) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
-1. **[VideoMAE](https://huggingface.co/docs/transformers/main/model_doc/videomae)** (from Multimedia Computing Group, Nanjing University) released with the paper [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://arxiv.org/abs/2203.12602) by Zhan Tong, Yibing Song, Jue Wang, Limin Wang.
-1. **[ViLT](https://huggingface.co/docs/transformers/model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
-1. **[Vision Transformer (ViT)](https://huggingface.co/docs/transformers/model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby. 
-1. **[VisualBERT](https://huggingface.co/docs/transformers/model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
-1. **[ViTMAE](https://huggingface.co/docs/transformers/model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
-1. **[Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
-1. **[Wav2Vec2-Conformer](https://huggingface.co/docs/transformers/model_doc/wav2vec2-conformer)** (from Facebook AI) released with the paper [FAIRSEQ S2T: Fast Speech-to-Text Modeling with FAIRSEQ](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Sravya Popuri, Dmytro Okhonko, Juan Pino.
-1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/transformers/model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
-1. **[WavLM](https://huggingface.co/docs/transformers/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
-1. **[XGLM](https://huggingface.co/docs/transformers/model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li. 
-1. **[XLM](https://huggingface.co/docs/transformers/model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
-1. **[XLM-ProphetNet](https://huggingface.co/docs/transformers/model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
-1. **[XLM-RoBERTa](https://huggingface.co/docs/transformers/model_doc/xlm-roberta)** (from Facebook AI), released together with the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov.
-1. **[XLM-RoBERTa-XL](https://huggingface.co/docs/transformers/model_doc/xlm-roberta-xl)** (from Facebook AI) released with the paper [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) by Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau.
-1. **[XLNet](https://huggingface.co/docs/transformers/model_doc/xlnet)** (from Google/CMU) released with the paper [​XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
-1. **[XLS-R](https://huggingface.co/docs/transformers/model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
-1. **[XLSR-Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
-1. **[YOLOS](https://huggingface.co/docs/transformers/model_doc/yolos)** (from Huazhong University of Science & Technology) released with the paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) by Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
-1. **[YOSO](https://huggingface.co/docs/transformers/model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
-1. 想要貢獻新的模型？我們這裡有一份**詳細指引和模板**來引導你加入新的模型。你可以在 [`templates`](./templates) 目錄中找到它們。記得查看[貢獻指引](./CONTRIBUTING.md)並在開始寫 PR 前聯繫維護人員或開一個新的 issue 來獲得 feedbacks。
-
-要檢查某個模型是否已有 Flax、PyTorch 或 TensorFlow 的實作，或其是否在🤗 Tokenizers 函式庫中有對應的 tokenizer，敬請參閱[此表](https://huggingface.co/docs/transformers/index#supported-frameworks)。
-
-這些實作均已於多個資料集測試（請參閱範例腳本）並應與原版實作表現相當。你可以在範例文件的[此節](https://huggingface.co/docs/transformers/examples)中了解實作的細節。
-
-
-## 了解更多
-
-| 章節 | 描述 |
-|-|-|
-| [文件](https://huggingface.co/transformers/) | 完整的 API 文件和教學 |
-| [任務概覽](https://huggingface.co/docs/transformers/task_summary) | 🤗 Transformers 支援的任務 |
-| [預處理教學](https://huggingface.co/docs/transformers/preprocessing) | 使用 `Tokenizer` 來為模型準備資料 |
-| [訓練和微調](https://huggingface.co/docs/transformers/training) | 使用 PyTorch/TensorFlow 的內建的訓練方式或於 `Trainer` API 中使用 🤗 Transformers 提供的模型 |
-| [快速上手：微調和範例腳本](https://github.com/huggingface/transformers/tree/main/examples) | 為各種任務提供的範例腳本 |
-| [模型分享和上傳](https://huggingface.co/docs/transformers/model_sharing) | 上傳並與社群分享你微調的模型 |
-| [遷移](https://huggingface.co/docs/transformers/migration) | 從 `pytorch-transformers` 或 `pytorch-pretrained-bert` 遷移到 🤗 Transformers |
-
-## 引用
-
-我們已將此函式庫的[論文](https://www.aclweb.org/anthology/2020.emnlp-demos.6/)正式發表。如果你使用了 🤗 Transformers 函式庫，可以引用：
-```bibtex
-@inproceedings{wolf-etal-2020-transformers,
-    title = "Transformers: State-of-the-Art Natural Language Processing",
-    author = "Thomas Wolf and Lysandre Debut and Victor Sanh and Julien Chaumond and Clement Delangue and Anthony Moi and Pierric Cistac and Tim Rault and Rémi Louf and Morgan Funtowicz and Joe Davison and Sam Shleifer and Patrick von Platen and Clara Ma and Yacine Jernite and Julien Plu and Canwen Xu and Teven Le Scao and Sylvain Gugger and Mariama Drame and Quentin Lhoest and Alexander M. Rush",
-    booktitle = "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations",
-    month = oct,
-    year = "2020",
-    address = "Online",
-    publisher = "Association for Computational Linguistics",
-    url = "https://www.aclweb.org/anthology/2020.emnlp-demos.6",
-    pages = "38--45"
-}
-```

conftest.py

@@ -1,78 +0,0 @@
-# Copyright 2020 The HuggingFace Team. All rights reserved.
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#     http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-# tests directory-specific settings - this file is run automatically
-# by pytest before any tests are run
-
-import doctest
-import sys
-import warnings
-from os.path import abspath, dirname, join
-
-
-# allow having multiple repository checkouts and not needing to remember to rerun
-# 'pip install -e .[dev]' when switching between checkouts and running tests.
-git_repo_path = abspath(join(dirname(__file__), "src"))
-sys.path.insert(1, git_repo_path)
-
-# silence FutureWarning warnings in tests since often we can't act on them until
-# they become normal warnings - i.e. the tests still need to test the current functionality
-warnings.simplefilter(action="ignore", category=FutureWarning)
-
-
-def pytest_configure(config):
-    config.addinivalue_line("markers", "is_pipeline_test: mark test to run only when pipeline are tested")
-    config.addinivalue_line(
-        "markers", "is_pt_tf_cross_test: mark test to run only when PT and TF interactions are tested"
-    )
-    config.addinivalue_line(
-        "markers", "is_pt_flax_cross_test: mark test to run only when PT and FLAX interactions are tested"
-    )
-    config.addinivalue_line("markers", "is_staging_test: mark test to run only in the staging environment")
-
-
-def pytest_addoption(parser):
-    from transformers.testing_utils import pytest_addoption_shared
-
-    pytest_addoption_shared(parser)
-
-
-def pytest_terminal_summary(terminalreporter):
-    from transformers.testing_utils import pytest_terminal_summary_main
-
-    make_reports = terminalreporter.config.getoption("--make-reports")
-    if make_reports:
-        pytest_terminal_summary_main(terminalreporter, id=make_reports)
-
-
-def pytest_sessionfinish(session, exitstatus):
-    # If no tests are collected, pytest exists with code 5, which makes the CI fail.
-    if exitstatus == 5:
-        session.exitstatus = 0
-
-
-# Doctest custom flag to ignore output.
-IGNORE_RESULT = doctest.register_optionflag('IGNORE_RESULT')
-
-OutputChecker = doctest.OutputChecker
-
-
-class CustomOutputChecker(OutputChecker):
-    def check_output(self, want, got, optionflags):
-        if IGNORE_RESULT & optionflags:
-            return True
-        return OutputChecker.check_output(self, want, got, optionflags)
-
-
-doctest.OutputChecker = CustomOutputChecker

docker/transformers-all-latest-gpu/Dockerfile

@@ -1,50 +0,0 @@
-FROM nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04
-LABEL maintainer="Hugging Face"
-
-ARG DEBIAN_FRONTEND=noninteractive
-
-# Use login shell to read variables from `~/.profile` (to pass dynamic created variables between RUN commands)
-SHELL ["sh", "-lc"]
-
-# The following `ARG` are mainly used to specify the versions explicitly & directly in this docker file, and not meant
-# to be used as arguments for docker build (so far).
-
-ARG PYTORCH='1.12.0'
-# (not always a valid torch version)
-ARG INTEL_TORCH_EXT='1.11.0'
-# Example: `cu102`, `cu113`, etc.
-ARG CUDA='cu113'
-
-RUN apt update
-RUN apt install -y git libsndfile1-dev tesseract-ocr espeak-ng python3 python3-pip ffmpeg git-lfs
-RUN git lfs install
-RUN python3 -m pip install --no-cache-dir --upgrade pip
-
-ARG REF=main
-RUN git clone https://github.com/huggingface/transformers && cd transformers && git checkout $REF
-RUN python3 -m pip install --no-cache-dir -e ./transformers[dev,onnxruntime]
-
-# TODO: Handle these in a python utility script
-RUN [ ${#PYTORCH} -gt 0 -a "$PYTORCH" != "pre" ] && VERSION='torch=='$PYTORCH'.*' ||  VERSION='torch'; echo "export VERSION='$VERSION'" >> ~/.profile
-RUN echo torch=$VERSION
-# `torchvision` and `torchaudio` should be installed along with `torch`, especially for nightly build.
-# Currently, let's just use their latest releases (when `torch` is installed with a release version)
-# TODO: We might need to specify proper versions that work with a specific torch version (especially for past CI).
-RUN [ "$PYTORCH" != "pre" ] && python3 -m pip install --no-cache-dir -U $VERSION torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/$CUDA || python3 -m pip install --no-cache-dir -U --pre torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/nightly/$CUDA
-
-RUN python3 -m pip install --no-cache-dir -U tensorflow
-RUN python3 -m pip uninstall -y flax jax
-
-# Use installed torch version for `torch-scatter` to avid to deal with PYTORCH='pre'.
-# If torch is nightly version, the link is likely to be invalid, but the installation falls back to the latest torch-scatter
-RUN python3 -m pip install --no-cache-dir torch-scatter -f https://data.pyg.org/whl/torch-$(python3 -c "from torch import version; print(version.__version__.split('+')[0])")+$CUDA.html
-RUN python3 -m pip install --no-cache-dir intel_extension_for_pytorch==$INTEL_TORCH_EXT+cpu -f https://software.intel.com/ipex-whl-stable
-
-RUN python3 -m pip install --no-cache-dir git+https://github.com/facebookresearch/detectron2.git pytesseract https://github.com/kpu/kenlm/archive/master.zip
-RUN python3 -m pip install -U "itsdangerous<2.1.0"
-
-RUN python3 -m pip install --no-cache-dir git+https://github.com/huggingface/accelerate@main#egg=accelerate
-
-# When installing in editable mode, `transformers` is not recognized as a package.
-# this line must be added in order for python to be aware of transformers.
-RUN cd transformers && python3 setup.py develop

docker/transformers-doc-builder/Dockerfile

@@ -1,20 +0,0 @@
-FROM python:3.8
-LABEL maintainer="Hugging Face"
-
-RUN apt update
-RUN git clone https://github.com/huggingface/transformers
-
-RUN python3 -m pip install --no-cache-dir --upgrade pip && python3 -m pip install --no-cache-dir git+https://github.com/huggingface/doc-builder ./transformers[dev]
-RUN apt-get -y update && apt-get install -y libsndfile1-dev && apt install -y tesseract-ocr
-
-# Torch needs to be installed before deepspeed
-RUN python3 -m pip install --no-cache-dir ./transformers[deepspeed]
-
-RUN python3 -m pip install --no-cache-dir torch-scatter -f https://data.pyg.org/whl/torch-$(python -c "from torch import version; print(version.__version__.split('+')[0])")+cpu.html
-RUN python3 -m pip install --no-cache-dir torchvision git+https://github.com/facebookresearch/detectron2.git pytesseract https://github.com/kpu/kenlm/archive/master.zip
-RUN python3 -m pip install --no-cache-dir pytorch-quantization --extra-index-url https://pypi.ngc.nvidia.com
-RUN python3 -m pip install -U "itsdangerous<2.1.0"
-
-# Test if the image could successfully build the doc. before publishing the image
-RUN doc-builder build transformers transformers/docs/source/en --build_dir doc-build-dev --notebook_dir notebooks/transformers_doc --clean
-RUN rm -rf doc-build-dev

docker/transformers-past-gpu/Dockerfile

@@ -1,43 +0,0 @@
-ARG BASE_DOCKER_IMAGE="nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04"
-FROM $BASE_DOCKER_IMAGE
-LABEL maintainer="Hugging Face"
-
-ARG DEBIAN_FRONTEND=noninteractive
-
-# Use login shell to read variables from `~/.profile` (to pass dynamic created variables between RUN commands)
-SHELL ["sh", "-lc"]
-
-RUN apt update
-RUN apt install -y git libsndfile1-dev tesseract-ocr espeak-ng python3 python3-pip ffmpeg git-lfs
-RUN git lfs install
-RUN python3 -m pip install --no-cache-dir --upgrade pip
-
-ARG REF=main
-RUN git clone https://github.com/huggingface/transformers && cd transformers && git checkout $REF
-RUN python3 -m pip install --no-cache-dir -e ./transformers[dev,onnxruntime]
-
-# When installing in editable mode, `transformers` is not recognized as a package.
-# this line must be added in order for python to be aware of transformers.
-RUN cd transformers && python3 setup.py develop
-
-ARG FRAMEWORK
-ARG VERSION
-
-# Remove all frameworks
-# (`accelerate` requires `torch`, and this causes import issues for TF-only testing)
-RUN python3 -m pip uninstall -y torch torchvision torchaudio accelerate tensorflow jax flax
-
-# Get the libraries and their versions to install, and write installation command to `~/.profile`.
-RUN python3 ./transformers/utils/past_ci_versions.py --framework $FRAMEWORK --version $VERSION
-
-# Install the target framework
-RUN echo "INSTALL_CMD = $INSTALL_CMD"
-RUN $INSTALL_CMD
-
-# Having installation problems for torch-scatter with torch <= 1.6. Disable so we have the same set of tests.
-# (This part will be removed once the logic of using `past_ci_versions.py` is used in other Dockerfile files.)
-# # Use installed torch version for `torch-scatter`.
-# # (The env. variable $CUDA is defined in `past_ci_versions.py`)
-# RUN [ "$FRAMEWORK" = "pytorch" ] && python3 -m pip install --no-cache-dir torch-scatter -f https://data.pyg.org/whl/torch-$(python3 -c "from torch import version; print(version.__version__.split('+')[0])")+$CUDA.html || echo "torch-scatter not to be installed"
-
-RUN python3 -m pip install -U "itsdangerous<2.1.0"

docker/transformers-pytorch-deepspeed-latest-gpu/Dockerfile

@@ -1,35 +0,0 @@
-FROM nvcr.io/nvidia/pytorch:21.03-py3
-LABEL maintainer="Hugging Face"
-
-ARG DEBIAN_FRONTEND=noninteractive
-
-ARG PYTORCH='1.12.0'
-# Example: `cu102`, `cu113`, etc.
-ARG CUDA='cu113'
-
-RUN apt -y update
-RUN apt install -y libaio-dev
-RUN python3 -m pip install --no-cache-dir --upgrade pip
-
-ARG REF=main
-RUN git clone https://github.com/huggingface/transformers && cd transformers && git checkout $REF
-
-# Install latest release PyTorch
-# (PyTorch must be installed before pre-compiling any DeepSpeed c++/cuda ops.)
-# (https://www.deepspeed.ai/tutorials/advanced-install/#pre-install-deepspeed-ops)
-RUN python3 -m pip install --no-cache-dir -U torch==$PYTORCH torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/$CUDA
-
-RUN python3 -m pip install --no-cache-dir ./transformers[deepspeed-testing]
-
-# Pre-build **latest** DeepSpeed, so it would be ready for testing (otherwise, the 1st deepspeed test will timeout)
-RUN python3 -m pip uninstall -y deepspeed
-# This has to be run (again) inside the GPU VMs running the tests.
-# The installation works here, but some tests fail, if we don't pre-build deepspeed again in the VMs running the tests.
-# TODO: Find out why test fail.
-RUN DS_BUILD_CPU_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install deepspeed --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check 2>&1
-
-# When installing in editable mode, `transformers` is not recognized as a package.
-# this line must be added in order for python to be aware of transformers.
-RUN cd transformers && python3 setup.py develop
-
-RUN python3 -c "from deepspeed.launcher.runner import main"

docker/transformers-pytorch-deepspeed-nightly-gpu/Dockerfile

@@ -1,35 +0,0 @@
-FROM nvcr.io/nvidia/pytorch:21.03-py3
-LABEL maintainer="Hugging Face"
-
-ARG DEBIAN_FRONTEND=noninteractive
-
-# Example: `cu102`, `cu113`, etc.
-ARG CUDA='cu113'
-
-RUN apt -y update
-RUN apt install -y libaio-dev
-RUN python3 -m pip install --no-cache-dir --upgrade pip
-
-ARG REF=main
-RUN git clone https://github.com/huggingface/transformers && cd transformers && git checkout $REF
-
-# Install **nightly** release PyTorch (flag `--pre`)
-# (PyTorch must be installed before pre-compiling any DeepSpeed c++/cuda ops.)
-# (https://www.deepspeed.ai/tutorials/advanced-install/#pre-install-deepspeed-ops)
-RUN python3 -m pip install --no-cache-dir -U --pre torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/nightly/$CUDA
-
-RUN python3 -m pip install --no-cache-dir ./transformers[deepspeed-testing]
-
-# Pre-build **nightly** release of DeepSpeed, so it would be ready for testing (otherwise, the 1st deepspeed test will timeout)
-RUN python3 -m pip uninstall -y deepspeed
-# This has to be run inside the GPU VMs running the tests. (So far, it fails here due to GPU checks during compilation.)
-# Issue: https://github.com/microsoft/DeepSpeed/issues/2010
-# RUN git clone https://github.com/microsoft/DeepSpeed && cd DeepSpeed && rm -rf build && \
-#    DS_BUILD_CPU_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install . --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check 2>&1
-
-# When installing in editable mode, `transformers` is not recognized as a package.
-# this line must be added in order for python to be aware of transformers.
-RUN cd transformers && python3 setup.py develop
-
-# Disable for now as deepspeed is not installed above. To be enabled once the issue is fixed.
-# RUN python3 -c "from deepspeed.launcher.runner import main"

docker/transformers-pytorch-gpu/Dockerfile

@@ -1,31 +1,30 @@
-FROM nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04
+FROM nvidia/cuda:10.2-cudnn7-devel-ubuntu18.04
 LABEL maintainer="Hugging Face"
+LABEL repository="transformers"
 
-ARG DEBIAN_FRONTEND=noninteractive
+RUN apt update && \
+    apt install -y bash \
+                   build-essential \
+                   git \
+                   curl \
+                   ca-certificates \
+                   python3 \
+                   python3-pip && \
+    rm -rf /var/lib/apt/lists
 
-RUN apt update
-RUN apt install -y git libsndfile1-dev tesseract-ocr espeak-ng python3 python3-pip ffmpeg
-RUN python3 -m pip install --no-cache-dir --upgrade pip
+RUN python3 -m pip install --no-cache-dir --upgrade pip && \
+    python3 -m pip install --no-cache-dir \
+    mkl \
+    torch
 
-ARG REF=main
-RUN git clone https://github.com/huggingface/transformers && cd transformers && git checkout $REF
-RUN python3 -m pip install --no-cache-dir -e ./transformers[dev-torch,testing]
+RUN git clone https://github.com/NVIDIA/apex
+RUN cd apex && \
+    python3 setup.py install && \
+    pip install -v --no-cache-dir --global-option="--cpp_ext" --global-option="--cuda_ext" ./
 
-# If set to nothing, will install the latest version
-ARG PYTORCH='1.12.0'
-ARG TORCH_VISION=''
-ARG TORCH_AUDIO=''
+WORKDIR /workspace
+COPY . transformers/
+RUN cd transformers/ && \
+    python3 -m pip install --no-cache-dir .
 
-RUN [ ${#PYTORCH} -gt 0 ] && VERSION='torch=='$PYTORCH'.*' ||  VERSION='torch'; python3 -m pip install --no-cache-dir -U $VERSION --extra-index-url https://download.pytorch.org/whl/cu113
-RUN [ ${#TORCH_VISION} -gt 0 ] && VERSION='torchvision=='TORCH_VISION'.*' ||  VERSION='torchvision'; python3 -m pip install --no-cache-dir -U $VERSION --extra-index-url https://download.pytorch.org/whl/cu113
-RUN [ ${#TORCH_AUDIO} -gt 0 ] && VERSION='torchaudio=='TORCH_AUDIO'.*' ||  VERSION='torchaudio'; python3 -m pip install --no-cache-dir -U $VERSION --extra-index-url https://download.pytorch.org/whl/cu113
-
-RUN python3 -m pip uninstall -y tensorflow flax
-
-RUN python3 -m pip install --no-cache-dir torch-scatter -f https://data.pyg.org/whl/torch-$(python3 -c "from torch import version; print(version.__version__.split('+')[0])")+cu113.html
-RUN python3 -m pip install --no-cache-dir git+https://github.com/facebookresearch/detectron2.git pytesseract https://github.com/kpu/kenlm/archive/master.zip
-RUN python3 -m pip install -U "itsdangerous<2.1.0"
-
-# When installing in editable mode, `transformers` is not recognized as a package.
-# this line must be added in order for python to be aware of transformers.
-RUN cd transformers && python3 setup.py develop
+CMD ["/bin/bash"]

docker/transformers-pytorch-tpu/Dockerfile

@@ -1,7 +1,7 @@
 FROM google/cloud-sdk:slim
 
 # Build args.
-ARG GITHUB_REF=refs/heads/main
+ARG GITHUB_REF=refs/heads/master
 
 # TODO: This Dockerfile installs pytorch/xla 3.6 wheels. There are also 3.7
 # wheels available; see below.
@@ -53,7 +53,7 @@ RUN git clone https://github.com/huggingface/transformers.git && \
     git checkout CI && \
     cd .. && \
     pip install ./transformers && \
-    pip install -r ./transformers/examples/pytorch/_test_requirements.txt && \
+    pip install -r ./transformers/examples/requirements.txt && \
     pip install pytest
 
 RUN python -c "import torch_xla; print(torch_xla.__version__)"

docker/transformers-pytorch-tpu/bert-base-cased.jsonnet

@@ -27,7 +27,7 @@ local bertBaseCased = base.BaseTest {
   },
   command: utils.scriptCommand(
     |||
-      python -m pytest -s transformers/examples/pytorch/test_xla_examples.py -v
+      python -m pytest -s transformers/examples/test_xla_examples.py -v
       test_exit_code=$?
       echo "\nFinished running commands.\n"
       test $test_exit_code -eq 0

docker/transformers-tensorflow-gpu/Dockerfile

@@ -1,23 +1,25 @@
-FROM nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04
+FROM nvidia/cuda:10.1-cudnn7-runtime-ubuntu18.04
 LABEL maintainer="Hugging Face"
+LABEL repository="transformers"
 
-ARG DEBIAN_FRONTEND=noninteractive
+RUN apt update && \
+    apt install -y bash \
+                   build-essential \
+                   git \
+                   curl \
+                   ca-certificates \
+                   python3 \
+                   python3-pip && \
+    rm -rf /var/lib/apt/lists
 
-RUN apt update
-RUN apt install -y git libsndfile1-dev tesseract-ocr espeak-ng python3 python3-pip ffmpeg
-RUN python3 -m pip install --no-cache-dir --upgrade pip
+RUN python3 -m pip install --no-cache-dir --upgrade pip && \
+    python3 -m pip install --no-cache-dir \
+    mkl \
+    tensorflow
 
-ARG REF=main
-RUN git clone https://github.com/huggingface/transformers && cd transformers && git checkout $REF
-RUN python3 -m pip install --no-cache-dir -e ./transformers[dev-tensorflow,testing]
+WORKDIR /workspace
+COPY . transformers/
+RUN cd transformers/ && \
+    python3 -m pip install --no-cache-dir .
 
-# If set to nothing, will install the latest version
-ARG TENSORFLOW=''
-
-RUN [ ${#TENSORFLOW} -gt 0 ] && VERSION='tensorflow=='$TENSORFLOW'.*' ||  VERSION='tensorflow'; python3 -m pip install --no-cache-dir -U $VERSION
-RUN python3 -m pip uninstall -y torch flax
-RUN python3 -m pip install -U "itsdangerous<2.1.0"
-
-# When installing in editable mode, `transformers` is not recognized as a package.
-# this line must be added in order for python to be aware of transformers.
-RUN cd transformers && python3 setup.py develop
+CMD ["/bin/bash"]

docs/Makefile

@@ -0,0 +1,19 @@
+# Minimal makefile for Sphinx documentation
+#
+
+# You can set these variables from the command line.
+SPHINXOPTS    =
+SPHINXBUILD   = sphinx-build
+SOURCEDIR     = source
+BUILDDIR      = _build
+
+# Put it first so that "make" without argument is like "make help".
+help:
+	@$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
+
+.PHONY: help Makefile
+
+# Catch-all target: route all unknown targets to Sphinx using the new
+# "make mode" option.  $(O) is meant as a shortcut for $(SPHINXOPTS).
+%: Makefile
+	@$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)

docs/README.md

@@ -1,19 +1,3 @@
-<!---
-Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
--->
-
 # Generating the documentation
 
 To generate the documentation, you first have to build it. Several packages are necessary to build the doc,
@@ -23,212 +7,178 @@ you can install them with the following command, at the root of the code reposit
 pip install -e ".[docs]"
 ```
 
-Then you need to install our special tool that builds the documentation:
-
-```bash
-pip install git+https://github.com/huggingface/doc-builder
-```
-
 ---
 **NOTE**
 
-You only need to generate the documentation to inspect it locally (if you're planning changes and want to
+You only need to generate the documentation to inspect it locally (if you're planning changes and want to 
 check how they look like before committing for instance). You don't have to commit the built documentation.
 
 ---
 
+## Packages installed
+
+Here's an overview of all the packages installed. If you ran the previous command installing all packages from
+`requirements.txt`, you do not need to run the following commands.
+
+Building it requires the package `sphinx` that you can
+install using:
+
+```bash
+pip install -U sphinx
+```
+
+You would also need the custom installed [theme](https://github.com/readthedocs/sphinx_rtd_theme) by
+[Read The Docs](https://readthedocs.org/). You can install it using the following command:
+
+```bash
+pip install sphinx_rtd_theme
+```
+
+The third necessary package is the `recommonmark` package to accept Markdown as well as Restructured text:
+
+```bash
+pip install recommonmark
+```
+
 ## Building the documentation
 
-Once you have setup the `doc-builder` and additional packages, you can generate the documentation by 
-typing the following command:
+Once you have setup `sphinx`, you can build the documentation by running the following command in the `/docs` folder:
 
 ```bash
-doc-builder build transformers docs/source/ --build_dir ~/tmp/test-build
+make html
 ```
 
-You can adapt the `--build_dir` to set any temporary folder that you prefer. This command will create it and generate
-the MDX files that will be rendered as the documentation on the main website. You can inspect them in your favorite
-Markdown editor.
-
-## Previewing the documentation
-
-To preview the docs, first install the `watchdog` module with:
-
-```bash
-pip install watchdog
-```
-
-Then run the following command:
-
-```bash
-doc-builder preview {package_name} {path_to_docs}
-```
-
-For example:
-
-```bash
-doc-builder preview transformers docs/source/en/
-```
-
-The docs will be viewable at [http://localhost:3000](http://localhost:3000). You can also preview the docs once you have opened a PR. You will see a bot add a comment to a link where the documentation with your changes lives.
+A folder called ``_build/html`` should have been created. You can now open the file ``_build/html/index.html`` in your
+browser. 
 
 ---
 **NOTE**
 
-The `preview` command only works with existing doc files. When you add a completely new file, you need to update `_toctree.yml` & restart `preview` command (`ctrl-c` to stop it & call `doc-builder preview ...` again).
+If you are adding/removing elements from the toc-tree or from any structural item, it is recommended to clean the build
+directory before rebuilding. Run the following command to clean and build:
+
+```bash
+make clean && make html
+```
 
 ---
 
-## Adding a new element to the navigation bar
+It should build the static app that will be available under `/docs/_build/html`
 
-Accepted files are Markdown (.md or .mdx).
+## Adding a new element to the tree (toc-tree)
 
-Create a file with its extension and put it in the source directory. You can then link it to the toc-tree by putting
-the filename without the extension in the [`_toctree.yml`](https://github.com/huggingface/transformers/blob/main/docs/source/_toctree.yml) file.
+Accepted files are reStructuredText (.rst) and Markdown (.md). Create a file with its extension and put it
+in the source directory. You can then link it to the toc-tree by putting the filename without the extension.
 
-## Renaming section headers and moving sections
+## Preview the documentation in a pull request
 
-It helps to keep the old links working when renaming section header and/or moving sections from one document to another. This is because the old links are likely to be used in Issues, Forums and Social media and it'd be make for a much more superior user experience if users reading those months later could still easily navigate to the originally intended information.
-
-Therefore we simply keep a little map of moved sections at the end of the document where the original section was. The key is to preserve the original anchor.
-
-So if you renamed a section from: "Section A" to "Section B", then you can add at the end of the file:
-
-```
-Sections that were moved:
-
-[ <a href="#section-b">Section A</a><a id="section-a"></a> ]
-```
-and of course if you moved it to another file, then:
-
-```
-Sections that were moved:
-
-[ <a href="../new-file#section-b">Section A</a><a id="section-a"></a> ]
-```
-
-Use the relative style to link to the new file so that the versioned docs continue to work.
-
-For an example of a rich moved sections set please see the very end of [the Trainer doc](https://github.com/huggingface/transformers/blob/main/docs/source/main_classes/trainer.mdx).
+Once you have made your pull request, you can check what the documentation will look like after it's merged by
+following these steps:
 
+- Look at the checks at the bottom of the conversation page of your PR (you may need to click on "show all checks" to
+  expand them).
+- Click on "details" next to the `ci/circleci: build_doc` check.
+- In the new window, click on the "Artifacts" tab.
+- Locate the file "docs/_build/html/index.html" (or any specific page you want to check) and click on it to get a 
+  preview.
 
 ## Writing Documentation - Specification
 
 The `huggingface/transformers` documentation follows the
-[Google documentation](https://sphinxcontrib-napoleon.readthedocs.io/en/latest/example_google.html) style for docstrings,
-although we can write them directly in Markdown.
+[Google documentation](https://sphinxcontrib-napoleon.readthedocs.io/en/latest/example_google.html) style. It is
+mostly written in ReStructuredText 
+([Sphinx simple documentation](https://www.sphinx-doc.org/en/master/usage/restructuredtext/index.html), 
+[Sourceforge complete documentation](https://docutils.sourceforge.io/docs/ref/rst/restructuredtext.html)).
+
 
 ### Adding a new tutorial
 
 Adding a new tutorial or section is done in two steps:
 
 - Add a new file under `./source`. This file can either be ReStructuredText (.rst) or Markdown (.md).
-- Link that file in `./source/_toctree.yml` on the correct toc-tree.
+- Link that file in `./source/index.rst` on the correct toc-tree.
 
 Make sure to put your new file under the proper section. It's unlikely to go in the first section (*Get Started*), so
 depending on the intended targets (beginners, more advanced users or researchers) it should go in section two, three or
 four.
 
-### Translating
-
-When translating, refer to the guide at [./TRANSLATING.md](https://github.com/huggingface/transformers/blob/main/docs/TRANSLATING.md).
-
-
 ### Adding a new model
 
 When adding a new model:
-
-- Create a file `xxx.mdx` or under `./source/model_doc` (don't hesitate to copy an existing file as template).
-- Link that file in `./source/_toctree.yml`.
+ 
+- Create a file `xxx.rst` under `./source/model_doc` (don't hesitate to copy an existing file as template). 
+- Link that file in `./source/index.rst` on the `model_doc` toc-tree.
 - Write a short overview of the model:
     - Overview with paper & authors
     - Paper abstract
     - Tips and tricks and how to use it best
 - Add the classes that should be linked in the model. This generally includes the configuration, the tokenizer, and
   every model of that class (the base model, alongside models with additional heads), both in PyTorch and TensorFlow.
-  The order is generally:
-    - Configuration,
+  The order is generally: 
+    - Configuration, 
     - Tokenizer
     - PyTorch base model
     - PyTorch head models
     - TensorFlow base model
     - TensorFlow head models
-    - Flax base model
-    - Flax head models
-
-These classes should be added using our Markdown syntax. Usually as follows:
 
+These classes should be added using the RST syntax. Usually as follows:
 ```
-## XXXConfig
+XXXConfig
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-[[autodoc]] XXXConfig
+.. autoclass:: transformers.XXXConfig
+    :members:
 ```
 
 This will include every public method of the configuration that is documented. If for some reason you wish for a method
 not to be displayed in the documentation, you can do so by specifying which methods should be in the docs:
 
 ```
-## XXXTokenizer
+XXXTokenizer
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-[[autodoc]] XXXTokenizer
-    - build_inputs_with_special_tokens
-    - get_special_tokens_mask
-    - create_token_type_ids_from_sequences
-    - save_vocabulary
-```
+.. autoclass:: transformers.XXXTokenizer
+    :members: build_inputs_with_special_tokens, get_special_tokens_mask,
+        create_token_type_ids_from_sequences, save_vocabulary
 
-If you just want to add a method that is not documented (for instance magic method like `__call__` are not documented
-byt default) you can put the list of methods to add in a list that contains `all`:
-
-```
-## XXXTokenizer
-
-[[autodoc]] XXXTokenizer
-    - all
-    - __call__
 ```
 
 ### Writing source documentation
 
-Values that should be put in `code` should either be surrounded by backticks: \`like so\`. Note that argument names
-and objects like True, None or any strings should usually be put in `code`.
+Values that should be put in `code` should either be surrounded by double backticks: \`\`like so\`\` or be written as
+an object using the :obj: syntax: :obj:\`like so\`. Note that argument names and objects like True, None or any strings
+should usually be put in `code`.
 
-When mentioning a class, function or method, it is recommended to use our syntax for internal links so that our tool
-adds a link to its documentation with this syntax: \[\`XXXClass\`\] or \[\`function\`\]. This requires the class or 
-function to be in the main package.
+When mentionning a class, it is recommended to use the :class: syntax as the mentioned class will be automatically
+linked by Sphinx: :class:\`~transformers.XXXClass\`
 
-If you want to create a link to some internal class or function, you need to
-provide its path. For instance: \[\`utils.ModelOutput\`\]. This will be converted into a link with
-`utils.ModelOutput` in the description. To get rid of the path and only keep the name of the object you are
-linking to in the description, add a ~: \[\`~utils.ModelOutput\`\] will generate a link with `ModelOutput` in the description.
+When mentioning a function, it is recommended to use the :func: syntax as the mentioned function will be automatically
+linked by Sphinx: :func:\`~transformers.function\`.
 
-The same works for methods so you can either use \[\`XXXClass.method\`\] or \[~\`XXXClass.method\`\].
+When mentioning a method, it is recommended to use the :meth: syntax as the mentioned method will be automatically
+linked by Sphinx: :meth:\`~transformers.XXXClass.method\`.
+
+Links should be done as so (note the double underscore at the end): \`text for the link <./local-link-or-global-link#loc>\`__
 
 #### Defining arguments in a method
 
-Arguments should be defined with the `Args:` (or `Arguments:` or `Parameters:`) prefix, followed by a line return and
-an indentation. The argument should be followed by its type, with its shape if it is a tensor, a colon and its
-description:
-
-```
-    Args:
-        n_layers (`int`): The number of layers of the model.
-```
-
-If the description is too long to fit in one line, another indentation is necessary before writing the description
-after th argument.
+Arguments should be defined with the `Args:` prefix, followed by a line return and an indentation. 
+The argument should be followed by its type, with its shape if it is a tensor, and a line return.
+Another indentation is necessary before writing the description of the argument.
 
 Here's an example showcasing everything so far:
 
 ```
     Args:
-        input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
+        input_ids (:obj:`torch.LongTensor` of shape :obj:`(batch_size, sequence_length)`):
             Indices of input sequence tokens in the vocabulary.
 
-            Indices can be obtained using [`AlbertTokenizer`]. See [`~PreTrainedTokenizer.encode`] and
-            [`~PreTrainedTokenizer.__call__`] for details.
+            Indices can be obtained using :class:`~transformers.AlbertTokenizer`.
+            See :meth:`~transformers.PreTrainedTokenizer.encode` and
+            :meth:`~transformers.PreTrainedTokenizer.__call__` for details.
 
-            [What are input IDs?](../glossary#input-ids)
+            `What are input IDs? <../glossary.html#input-ids>`__
 ```
 
 For optional arguments or arguments with defaults we follow the following syntax: imagine we have a function with the
@@ -242,190 +192,53 @@ then its documentation should look like this:
 
 ```
     Args:
-        x (`str`, *optional*):
+        x (:obj:`str`, `optional`):
             This argument controls ...
-        a (`float`, *optional*, defaults to 1):
+        a (:obj:`float`, `optional`, defaults to 1):
             This argument is used to ...
 ```
 
-Note that we always omit the "defaults to \`None\`" when None is the default for any argument. Also note that even
+Note that we always omit the "defaults to :obj:\`None\`" when None is the default for any argument. Also note that even
 if the first line describing your argument type and its default gets long, you can't break it on several lines. You can
-however write as many lines as you want in the indented description (see the example above with `input_ids`).
+however write as many lines as you want in the indented description (see the example above with `input_ids`). 
 
-#### Writing a multi-line code block
+#### Writing a multi-line code block 
 
-Multi-line code blocks can be useful for displaying examples. They are done between two lines of three backticks as usual in Markdown:
+Multi-line code blocks can be useful for displaying examples. They are done like so:
 
-
-````
 ```
-# first line of code
-# second line
-# etc
+Example::
+
+    # first line of code
+    # second line
+    # etc
 ```
-````
+
+The `Example` string at the beginning can be replaced by anything as long as there are two semicolons following it.
 
 We follow the [doctest](https://docs.python.org/3/library/doctest.html) syntax for the examples to automatically test
 the results stay consistent with the library.
 
 #### Writing a return block
 
-The return block should be introduced with the `Returns:` prefix, followed by a line return and an indentation.
+Arguments should be defined with the `Args:` prefix, followed by a line return and an indentation. 
 The first line should be the type of the return, followed by a line return. No need to indent further for the elements
 building the return.
 
-Here's an example for a single value return:
-
-```
-    Returns:
-        `List[int]`: A list of integers in the range [0, 1] --- 1 for a special token, 0 for a sequence token.
-```
-
 Here's an example for tuple return, comprising several objects:
 
 ```
     Returns:
-        `tuple(torch.FloatTensor)` comprising various elements depending on the configuration ([`BertConfig`]) and inputs:
-        - ** loss** (*optional*, returned when `masked_lm_labels` is provided) `torch.FloatTensor` of shape `(1,)` --
-          Total loss as the sum of the masked language modeling loss and the next sequence prediction (classification) loss.
-        - **prediction_scores** (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.vocab_size)`) --
-          Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
+        :obj:`tuple(torch.FloatTensor)` comprising various elements depending on the configuration (:class:`~transformers.BertConfig`) and inputs:
+        loss (`optional`, returned when ``masked_lm_labels`` is provided) ``torch.FloatTensor`` of shape ``(1,)``:
+            Total loss as the sum of the masked language modeling loss and the next sequence prediction (classification) loss.
+        prediction_scores (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, sequence_length, config.vocab_size)`)
+            Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
 ```
 
-#### Adding an image
-
-Due to the rapidly growing repository, it is important to make sure that no files that would significantly weigh down the repository are added. This includes images, videos and other non-text files. We prefer to leverage a hf.co hosted `dataset` like
-the ones hosted on [`hf-internal-testing`](https://huggingface.co/hf-internal-testing) in which to place these files and reference
-them by URL. We recommend putting them in the following dataset: [huggingface/documentation-images](https://huggingface.co/datasets/huggingface/documentation-images).
-If an external contribution, feel free to add the images to your PR and ask a Hugging Face member to migrate your images
-to this dataset.
-
-## Styling the docstring
-
-We have an automatic script running with the `make style` comment that will make sure that:
-- the docstrings fully take advantage of the line width
-- all code examples are formatted using black, like the code of the Transformers library
-
-This script may have some weird failures if you made a syntax mistake or if you uncover a bug. Therefore, it's
-recommended to commit your changes before running `make style`, so you can revert the changes done by that script
-easily.
-
-# Testing documentation examples
-
-Good documentation oftens comes with an example of how a specific function or class should be used. 
-Each model class should contain at least one example showcasing
-how to use this model class in inference. *E.g.* the class [Wav2Vec2ForCTC](https://huggingface.co/docs/transformers/model_doc/wav2vec2#transformers.Wav2Vec2ForCTC) 
-includes an example of how to transcribe speech to text in the 
-[docstring of its forward function](https://huggingface.co/docs/transformers/model_doc/wav2vec2#transformers.Wav2Vec2ForCTC.forward).
-
-## Writing documenation examples
-
-The syntax for Example docstrings can look as follows:
+Here's an example for a single value return:
 
 ```
-    Example:
-
-    ```python
-    >>> from transformers import Wav2Vec2Processor, Wav2Vec2ForCTC
-    >>> from datasets import load_dataset
-    >>> import torch
-
-    >>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation")
-    >>> dataset = dataset.sort("id")
-    >>> sampling_rate = dataset.features["audio"].sampling_rate
-
-    >>> processor = Wav2Vec2Processor.from_pretrained("facebook/wav2vec2-base-960h")
-    >>> model = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-base-960h")
-
-    >>> # audio file is decoded on the fly
-    >>> inputs = processor(dataset[0]["audio"]["array"], sampling_rate=sampling_rate, return_tensors="pt")
-    >>> with torch.no_grad():
-    ...     logits = model(**inputs).logits
-    >>> predicted_ids = torch.argmax(logits, dim=-1)
-
-    >>> # transcribe speech
-    >>> transcription = processor.batch_decode(predicted_ids)
-    >>> transcription[0]
-    'MISTER QUILTER IS THE APOSTLE OF THE MIDDLE CLASSES AND WE ARE GLAD TO WELCOME HIS GOSPEL'
-    ```
+    Returns:
+        :obj:`List[int]`: A list of integers in the range [0, 1] --- 1 for a special token, 0 for a sequence token.
 ```
-
-The docstring should give a minimal, clear example of how the respective model 
-is to be used in inference and also include the expected (ideally sensible)
-output.
-Often, readers will try out the example before even going through the function 
-or class definitions. Therefore it is of utmost importance that the example 
-works as expected.
-
-## Docstring testing
-
-To do so each example should be included in the doctests. 
-We use pytests' [doctest integration](https://docs.pytest.org/doctest.html) to verify that all of our examples run correctly. 
-For Transformers, the doctests are run on a daily basis via GitHub Actions as can be 
-seen [here](https://github.com/huggingface/transformers/actions/workflows/doctests.yml).
-
-To include your example in the daily doctests, you need add the filename that
-contains the example docstring to the [documentation_tests.txt](../utils/documentation_tests.txt).
-
-### For Python files
-
-You will first need to run the following command (from the root of the repository) to prepare the doc file (doc-testing needs to add additional lines that we don't include in the doc source files):
-
-```bash
-python utils/prepare_for_doc_test.py src docs
-```
-
-If you work on a specific python module, say `modeling_wav2vec2.py`, you can run the command as follows (to avoid the unnecessary temporary changes in irrelevant files):
-
-```bash
-python utils/prepare_for_doc_test.py src/transformers/utils/doc.py src/transformers/models/wav2vec2/modeling_wav2vec2.py
-```
-(`utils/doc.py` should always be included)
-
-Then you can run all the tests in the docstrings of a given file with the following command, here is how we test the modeling file of Wav2Vec2 for instance:
-
-```bash
-pytest --doctest-modules src/transformers/models/wav2vec2/modeling_wav2vec2.py -sv --doctest-continue-on-failure
-```
-
-If you want to isolate a specific docstring, just add `::` after the file name then type the whole path of the function/class/method whose docstring you want to test. For instance, here is how to just test the forward method of `Wav2Vec2ForCTC`:
-
-```bash
-pytest --doctest-modules src/transformers/models/wav2vec2/modeling_wav2vec2.py::transformers.models.wav2vec2.modeling_wav2vec2.Wav2Vec2ForCTC.forward -sv --doctest-continue-on-failure
-```
-
-Once you're done, you can run the following command (still from the root of the repository) to undo the changes made by the first command before committing:
-
-```bash
-python utils/prepare_for_doc_test.py src docs --remove_new_line
-```
-
-### For Markdown files
-
-You will first need to run the following command (from the root of the repository) to prepare the doc file (doc-testing needs to add additional lines that we don't include in the doc source files):
-
-```bash
-python utils/prepare_for_doc_test.py src docs
-```
-
-Then you can test locally a given file with this command (here testing the quicktour):
-
-```bash
-pytest --doctest-modules docs/source/quicktour.mdx -sv --doctest-continue-on-failure --doctest-glob="*.mdx"
-```
-
-Once you're done, you can run the following command (still from the root of the repository) to undo the changes made by the first command before committing:
-
-```bash
-python utils/prepare_for_doc_test.py src docs --remove_new_line
-```
-
-### Writing doctests
-
-Here are a few tips to help you debug the doctests and make them pass:
-
-- The outputs of the code need to match the expected output **exactly**, so make sure you have the same outputs. In particular doctest will see a difference between single quotes and double quotes, or a missing parenthesis. The only exceptions to that rule are:
-  * whitespace: one give whitespace (space, tabulation, new line) is equivalent to any number of whitespace, so you can add new lines where there are spaces to make your output more readable.
-  * numerical values: you should never put more than 4 or 5 digits to expected results as different setups or library versions might get you slightly different results. `doctest` is configure to ignore any difference lower than the precision to which you wrote (so 1e-4 if you write 4 digits).
-- Don't leave a block of code that is very long to execute. If you can't make it fast, you can either not use the doctest syntax on it (so that it's ignored), or if you want to use the doctest syntax to show the results, you can add a comment `# doctest: +SKIP` at the end of the lines of code too long to execute
-- Each line of code that produces a result needs to have that result written below. You can ignore an output if you don't want to show it in your code example by adding a comment ` # doctest: +IGNORE_RESULT` at the end of the line of code producing it.

docs/TRANSLATING.md

@@ -1,58 +0,0 @@
-### Translating the Transformers documentation into your language
-
-As part of our mission to democratize machine learning, we'd love to make the Transformers library available in many more languages! Follow the steps below if you want to help translate the documentation into your language 🙏.
-
-**🗞️ Open an issue**
-
-To get started, navigate to the [Issues](https://github.com/huggingface/transformers/issues) page of this repo and check if anyone else has opened an issue for your language. If not, open a new issue by selecting the "Translation template" from the "New issue" button.
-
-Once an issue exists, post a comment to indicate which chapters you'd like to work on, and we'll add your name to the list.
-
-
-**🍴 Fork the repository**
-
-First, you'll need to [fork the Transformers repo](https://docs.github.com/en/get-started/quickstart/fork-a-repo). You can do this by clicking on the **Fork** button on the top-right corner of this repo's page.
-
-Once you've forked the repo, you'll want to get the files on your local machine for editing. You can do that by cloning the fork with Git as follows:
-
-```bash
-git clone https://github.com/YOUR-USERNAME/transformers.git
-```
-
-**📋 Copy-paste the English version with a new language code**
-
-The documentation files are in one leading directory:
-
-- [`docs/source`](https://github.com/huggingface/transformers/tree/main/docs/source): All the documentation materials are organized here by language.
-
-You'll only need to copy the files in the [`docs/source/en`](https://github.com/huggingface/transformers/tree/main/docs/source/en) directory, so first navigate to your fork of the repo and run the following:
-
-```bash
-cd ~/path/to/transformers/docs
-cp -r source/en source/LANG-ID
-```
-
-Here, `LANG-ID` should be one of the ISO 639-1 or ISO 639-2 language codes -- see [here](https://www.loc.gov/standards/iso639-2/php/code_list.php) for a handy table.
-
-**✍️ Start translating**
-
-The fun part comes - translating the text!
-
-The first thing we recommend is translating the part of the `_toctree.yml` file that corresponds to your doc chapter. This file is used to render the table of contents on the website. 
-
-> 🙋 If the `_toctree.yml` file doesn't yet exist for your language, you can create one by copy-pasting from the English version and deleting the sections unrelated to your chapter. Just make sure it exists in the `docs/source/LANG-ID/` directory!
-
-The fields you should add are `local` (with the name of the file containing the translation; e.g. `autoclass_tutorial`), and `title` (with the title of the doc in your language; e.g. `Load pretrained instances with an AutoClass`) -- as a reference, here is the `_toctree.yml` for [English](https://github.com/huggingface/transformers/blob/main/docs/source/en/_toctree.yml):
-
-```yaml
-- sections:
-  - local: pipeline_tutorial # Do not change this! Use the same name for your .md file
-    title: Pipelines for inference # Translate this!
-    ...
-  title: Tutorials # Translate this!
-```
-
-Once you have translated the `_toctree.yml` file, you can start translating the [MDX](https://mdxjs.com/) files associated with your docs chapter.
-
-> 🙋 If you'd like others to help you with the translation, you can either [open an issue](https://github.com/huggingface/transformers/issues) or tag @[espejelomar](https://twitter.com/espejelomar)
- on Twitter to gain some visibility.

docs/source/_config.py

@@ -1,14 +0,0 @@
-# docstyle-ignore
-INSTALL_CONTENT = """
-# Transformers installation
-! pip install transformers datasets
-# To install from source instead of the last release, comment the command above and uncomment the following one.
-# ! pip install git+https://github.com/huggingface/transformers.git
-"""
-
-notebook_first_cells = [{"type": "code", "content": INSTALL_CONTENT}]
-black_avoid_patterns = {
-    "{processor_class}": "FakeProcessorClass",
-    "{model_class}": "FakeModelClass",
-    "{object_class}": "FakeObjectClass",    
-}

docs/source/_static/css/code-snippets.css

@@ -0,0 +1,16 @@
+
+.highlight .c1, .highlight .sd{
+    color: #999
+}
+
+.highlight .nn, .highlight .k, .highlight .s1, .highlight .nb, .highlight .bp, .highlight .kc {
+    color: #FB8D68;
+}
+
+.highlight .kn, .highlight .nv, .highlight .s2, .highlight .ow {
+    color: #6670FF;
+}
+
+.highlight .gp {
+    color: #FB8D68;
+}

docs/source/_static/css/huggingface.css

@@ -0,0 +1,350 @@
+/* Our DOM objects */
+
+/* Colab dropdown */
+
+table.center-aligned-table td {
+    text-align: center;
+}
+
+table.center-aligned-table th {
+    text-align: center;
+    vertical-align: middle;
+}
+
+.colab-dropdown {
+    position: relative;
+    display: inline-block;
+}
+  
+.colab-dropdown-content {
+    display: none;
+    position: absolute;
+    background-color: #f9f9f9;
+    min-width: 117px;
+    box-shadow: 0px 8px 16px 0px rgba(0,0,0,0.2);
+    z-index: 1;
+}
+  
+.colab-dropdown-content button {
+    color: #6670FF;
+    background-color: #f9f9f9;
+    font-size: 12px;
+    border: none;
+    min-width: 117px;
+    padding: 5px 5px;
+    text-decoration: none;
+    display: block;
+}
+  
+.colab-dropdown-content button:hover {background-color: #eee;}
+  
+.colab-dropdown:hover .colab-dropdown-content {display: block;}
+
+/* Version control */
+
+.version-button {
+    background-color: #6670FF;
+    color: white;
+    border: none;
+    padding: 5px;
+    font-size: 15px;
+    cursor: pointer;
+}
+
+.version-button:hover, .version-button:focus {
+    background-color: #A6B0FF;
+}
+ 
+.version-dropdown {
+    display: none;
+    background-color: #6670FF;
+    min-width: 160px;
+    overflow: auto;
+    font-size: 15px;
+}
+  
+.version-dropdown a {
+    color: white;
+    padding: 3px 4px;
+    text-decoration: none;
+    display: block;
+}
+  
+.version-dropdown a:hover {
+    background-color: #A6B0FF;
+}
+  
+.version-show {
+    display: block;
+}
+
+/* Framework selector */
+
+.framework-selector {
+    display: flex;
+    flex-direction: row;
+    justify-content: flex-end;
+    margin-right: 30px;
+}
+
+.framework-selector > button {
+    background-color: white;
+    color: #6670FF;
+    border: 1px solid #6670FF;
+    padding: 5px;
+}
+
+.framework-selector > button.selected{
+    background-color: #6670FF;
+    color: white;
+    border: 1px solid #6670FF;
+    padding: 5px;
+}
+
+/* Copy button */
+
+a.copybtn {
+    margin: 3px;
+}
+
+/* The literal code blocks */
+.rst-content tt.literal, .rst-content tt.literal, .rst-content code.literal {
+    color: #6670FF;
+}
+
+/* To keep the logo centered */
+.wy-side-scroll {
+    width: auto;
+    font-size: 20px;
+}
+
+/* The div that holds the Hugging Face logo */
+.HuggingFaceDiv {
+    width: 100%
+}
+
+/* The research field on top of the toc tree */
+.wy-side-nav-search{
+    padding-top: 0;
+    background-color: #6670FF;
+}
+
+/* The toc tree */
+.wy-nav-side{
+    background-color: #6670FF;
+}
+
+/* The section headers in the toc tree */
+.wy-menu-vertical p.caption{
+    background-color: #4d59ff;
+    line-height: 40px;
+}
+
+/* The selected items in the toc tree */
+.wy-menu-vertical li.current{
+    background-color: #A6B0FF;
+}
+
+/* When a list item that does belong to the selected block from the toc tree is hovered */
+.wy-menu-vertical li.current a:hover{
+    background-color: #B6C0FF;
+}
+
+/* When a list item that does NOT belong to the selected block from the toc tree is hovered. */
+.wy-menu-vertical li a:hover{
+    background-color: #A7AFFB;
+}
+
+/* The text items on the toc tree */
+.wy-menu-vertical a {
+    color: #FFFFDD;
+    font-family: Calibre-Light, sans-serif;
+}
+.wy-menu-vertical header, .wy-menu-vertical p.caption{
+    color: white;
+    font-family: Calibre-Light, sans-serif;
+}
+
+/* The color inside the selected toc tree block */
+.wy-menu-vertical li.toctree-l2 a, .wy-menu-vertical li.toctree-l3 a, .wy-menu-vertical li.toctree-l4 a {
+    color: black;
+}
+
+/* Inside the depth-2 selected toc tree block */
+.wy-menu-vertical li.toctree-l2.current>a {
+    background-color: #B6C0FF
+}
+.wy-menu-vertical li.toctree-l2.current li.toctree-l3>a {
+    background-color: #C6D0FF
+}
+
+/* Inside the depth-3 selected toc tree block */
+.wy-menu-vertical li.toctree-l3.current li.toctree-l4>a{
+    background-color: #D6E0FF
+}
+
+/* Inside code snippets */
+.rst-content dl:not(.docutils) dt{
+    font-size: 15px;
+}
+
+/* Links */
+a {
+    color: #6670FF;
+}
+
+/* Content bars */
+.rst-content dl:not(.docutils) dt {
+    background-color: rgba(251, 141, 104, 0.1);
+    border-right: solid 2px #FB8D68;
+    border-left: solid 2px #FB8D68;
+    color: #FB8D68;
+    font-family: Calibre-Light, sans-serif;
+    border-top: none;
+    font-style: normal !important;
+}
+
+/* Expand button */
+.wy-menu-vertical li.toctree-l2 span.toctree-expand,
+.wy-menu-vertical li.on a span.toctree-expand, .wy-menu-vertical li.current>a span.toctree-expand,
+.wy-menu-vertical li.toctree-l3 span.toctree-expand{
+    color: black;
+}
+
+/* Max window size */
+.wy-nav-content{
+    max-width: 1200px;
+}
+
+/* Mobile header */
+.wy-nav-top{
+    background-color: #6670FF;
+}
+
+
+/* Source spans */
+.rst-content .viewcode-link, .rst-content .viewcode-back{
+    color: #6670FF;
+    font-size: 110%;
+    letter-spacing: 2px;
+    text-transform: uppercase;
+}
+
+/* It would be better for table to be visible without horizontal scrolling */
+.wy-table-responsive table td, .wy-table-responsive table th{
+    white-space: normal;
+}
+
+.footer {
+    margin-top: 20px;
+}
+
+.footer__Social {
+    display: flex;
+    flex-direction: row;
+}
+
+.footer__CustomImage {
+    margin: 2px 5px 0 0;
+}
+
+/* class and method names in doc */
+.rst-content dl:not(.docutils) tt.descname, .rst-content dl:not(.docutils) tt.descclassname, .rst-content dl:not(.docutils) tt.descname, .rst-content dl:not(.docutils) code.descname, .rst-content dl:not(.docutils) tt.descclassname, .rst-content dl:not(.docutils) code.descclassname{
+    font-family: Calibre, sans-serif;
+    font-size: 20px !important;
+}
+
+/* class name in doc*/
+.rst-content dl:not(.docutils) tt.descname, .rst-content dl:not(.docutils) tt.descname, .rst-content dl:not(.docutils) code.descname{
+    margin-right: 10px;
+    font-family: Calibre-Medium, sans-serif;
+}
+
+/* Method and class parameters */
+.sig-param{
+    line-height: 23px;
+}
+
+/* Class introduction "class" string at beginning */
+.rst-content dl:not(.docutils) .property{
+    font-size: 18px;
+    color: black;
+}
+
+
+/* FONTS */
+body{
+    font-family: Calibre, sans-serif;
+    font-size: 16px;
+}
+
+h1 {
+    font-family: Calibre-Thin, sans-serif;
+    font-size: 70px;
+}
+
+h2, .rst-content .toctree-wrapper p.caption, h3, h4, h5, h6, legend{
+    font-family: Calibre-Medium, sans-serif;
+}
+
+@font-face {
+    font-family: Calibre-Medium;
+    src: url(./Calibre-Medium.otf);
+    font-weight:400;
+}
+
+@font-face {
+    font-family: Calibre;
+    src: url(./Calibre-Regular.otf);
+    font-weight:400;
+}
+
+@font-face {
+    font-family: Calibre-Light;
+    src: url(./Calibre-Light.ttf);
+    font-weight:400;
+}
+
+@font-face {
+    font-family: Calibre-Thin;
+    src: url(./Calibre-Thin.otf);
+    font-weight:400;
+}
+
+
+/**
+ * Nav Links to other parts of huggingface.co
+ */
+ div.menu {
+    position: absolute;
+    top: 0;
+    right: 0;
+    padding-top: 20px;
+    padding-right: 20px;
+    z-index: 1000;
+}
+div.menu a {
+    font-size: 14px;
+    letter-spacing: 0.3px;
+    text-transform: uppercase;
+    color: white;
+    -webkit-font-smoothing: antialiased;
+    background: linear-gradient(0deg, #6671ffb8, #9a66ffb8 50%);
+    padding: 10px 16px 6px 16px;
+    border-radius: 3px;
+    margin-left: 12px;
+    position: relative;
+}
+div.menu a:active {
+    top: 1px;
+}
+@media (min-width: 768px) and (max-width: 1750px) {
+    .wy-breadcrumbs {
+        margin-top: 32px;
+    }
+}
+@media (max-width: 768px) {
+    div.menu {
+        display: none;
+    }
+}

docs/source/benchmarks.rst

@@ -0,0 +1,345 @@
+Benchmarks
+=======================================================================================================================
+
+Let's take a look at how 🤗 Transformer models can be benchmarked, best practices, and already available benchmarks.
+
+A notebook explaining in more detail how to benchmark 🤗 Transformer models can be found `here
+<https://github.com/huggingface/transformers/blob/master/notebooks/05-benchmark.ipynb>`__.
+
+How to benchmark 🤗 Transformer models
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+The classes :class:`~transformers.PyTorchBenchmark` and :class:`~transformers.TensorFlowBenchmark` allow to flexibly
+benchmark 🤗 Transformer models. The benchmark classes allow us to measure the `peak memory usage` and `required time`
+for both `inference` and `training`.
+
+.. note::
+
+  Hereby, `inference` is defined by a single forward pass, and `training` is defined by a single forward pass and
+  backward pass.
+
+The benchmark classes :class:`~transformers.PyTorchBenchmark` and :class:`~transformers.TensorFlowBenchmark` expect an
+object of type :class:`~transformers.PyTorchBenchmarkArguments` and
+:class:`~transformers.TensorFlowBenchmarkArguments`, respectively, for instantiation.
+:class:`~transformers.PyTorchBenchmarkArguments` and :class:`~transformers.TensorFlowBenchmarkArguments` are data
+classes and contain all relevant configurations for their corresponding benchmark class. In the following example, it
+is shown how a BERT model of type `bert-base-cased` can be benchmarked.
+
+.. code-block::
+
+    >>> ## PYTORCH CODE
+    >>> from transformers import PyTorchBenchmark, PyTorchBenchmarkArguments
+
+    >>> args = PyTorchBenchmarkArguments(models=["bert-base-uncased"], batch_sizes=[8], sequence_lengths=[8, 32, 128, 512])
+    >>> benchmark = PyTorchBenchmark(args)
+
+    >>> ## TENSORFLOW CODE
+    >>> from transformers import TensorFlowBenchmark, TensorFlowBenchmarkArguments
+
+    >>> args = TensorFlowBenchmarkArguments(models=["bert-base-uncased"], batch_sizes=[8], sequence_lengths=[8, 32, 128, 512])
+    >>> benchmark = TensorFlowBenchmark(args)
+
+
+Here, three arguments are given to the benchmark argument data classes, namely ``models``, ``batch_sizes``, and
+``sequence_lengths``. The argument ``models`` is required and expects a :obj:`list` of model identifiers from the
+`model hub <https://huggingface.co/models>`__ The :obj:`list` arguments ``batch_sizes`` and ``sequence_lengths`` define
+the size of the ``input_ids`` on which the model is benchmarked. There are many more parameters that can be configured
+via the benchmark argument data classes. For more detail on these one can either directly consult the files
+``src/transformers/benchmark/benchmark_args_utils.py``, ``src/transformers/benchmark/benchmark_args.py`` (for PyTorch)
+and ``src/transformers/benchmark/benchmark_args_tf.py`` (for Tensorflow). Alternatively, running the following shell
+commands from root will print out a descriptive list of all configurable parameters for PyTorch and Tensorflow
+respectively.
+
+.. code-block:: bash
+
+    ## PYTORCH CODE
+    python examples/benchmarking/run_benchmark.py --help
+
+    ## TENSORFLOW CODE
+    python examples/benchmarking/run_benchmark_tf.py --help
+
+
+An instantiated benchmark object can then simply be run by calling ``benchmark.run()``.
+
+.. code-block::
+
+    >>> ## PYTORCH CODE
+    >>> results = benchmark.run()
+    >>> print(results)
+    ====================       INFERENCE - SPEED - RESULT       ====================
+    --------------------------------------------------------------------------------
+    Model Name             Batch Size     Seq Length     Time in s                  
+    --------------------------------------------------------------------------------
+    bert-base-uncased          8               8             0.006     
+    bert-base-uncased          8               32            0.006     
+    bert-base-uncased          8              128            0.018     
+    bert-base-uncased          8              512            0.088     
+    --------------------------------------------------------------------------------
+
+    ====================      INFERENCE - MEMORY - RESULT       ====================
+    --------------------------------------------------------------------------------
+    Model Name             Batch Size     Seq Length    Memory in MB 
+    --------------------------------------------------------------------------------
+    bert-base-uncased          8               8             1227
+    bert-base-uncased          8               32            1281
+    bert-base-uncased          8              128            1307
+    bert-base-uncased          8              512            1539
+    --------------------------------------------------------------------------------
+
+    ====================        ENVIRONMENT INFORMATION         ====================
+    - transformers_version: 2.11.0
+    - framework: PyTorch
+    - use_torchscript: False
+    - framework_version: 1.4.0
+    - python_version: 3.6.10
+    - system: Linux
+    - cpu: x86_64
+    - architecture: 64bit
+    - date: 2020-06-29
+    - time: 08:58:43.371351
+    - fp16: False
+    - use_multiprocessing: True
+    - only_pretrain_model: False
+    - cpu_ram_mb: 32088
+    - use_gpu: True
+    - num_gpus: 1
+    - gpu: TITAN RTX
+    - gpu_ram_mb: 24217
+    - gpu_power_watts: 280.0
+    - gpu_performance_state: 2
+    - use_tpu: False
+
+    >>> ## TENSORFLOW CODE
+    >>> results = benchmark.run()
+    >>> print(results)
+    ====================       INFERENCE - SPEED - RESULT       ====================
+    --------------------------------------------------------------------------------
+    Model Name             Batch Size     Seq Length     Time in s                  
+    --------------------------------------------------------------------------------
+    bert-base-uncased          8               8             0.005
+    bert-base-uncased          8               32            0.008
+    bert-base-uncased          8              128            0.022
+    bert-base-uncased          8              512            0.105
+    --------------------------------------------------------------------------------
+
+    ====================      INFERENCE - MEMORY - RESULT       ====================
+    --------------------------------------------------------------------------------
+    Model Name             Batch Size     Seq Length    Memory in MB 
+    --------------------------------------------------------------------------------
+    bert-base-uncased          8               8             1330
+    bert-base-uncased          8               32            1330
+    bert-base-uncased          8              128            1330
+    bert-base-uncased          8              512            1770
+    --------------------------------------------------------------------------------
+
+    ====================        ENVIRONMENT INFORMATION         ====================
+    - transformers_version: 2.11.0
+    - framework: Tensorflow
+    - use_xla: False
+    - framework_version: 2.2.0
+    - python_version: 3.6.10
+    - system: Linux
+    - cpu: x86_64
+    - architecture: 64bit
+    - date: 2020-06-29
+    - time: 09:26:35.617317
+    - fp16: False
+    - use_multiprocessing: True
+    - only_pretrain_model: False
+    - cpu_ram_mb: 32088
+    - use_gpu: True
+    - num_gpus: 1
+    - gpu: TITAN RTX
+    - gpu_ram_mb: 24217
+    - gpu_power_watts: 280.0
+    - gpu_performance_state: 2
+    - use_tpu: False
+
+By default, the `time` and the `required memory` for `inference` are benchmarked. In the example output above the first
+two sections show the result corresponding to `inference time` and `inference memory`. In addition, all relevant
+information about the computing environment, `e.g.` the GPU type, the system, the library versions, etc... are printed
+out in the third section under `ENVIRONMENT INFORMATION`. This information can optionally be saved in a `.csv` file
+when adding the argument :obj:`save_to_csv=True` to :class:`~transformers.PyTorchBenchmarkArguments` and
+:class:`~transformers.TensorFlowBenchmarkArguments` respectively. In this case, every section is saved in a separate
+`.csv` file. The path to each `.csv` file can optionally be defined via the argument data classes.
+
+Instead of benchmarking pre-trained models via their model identifier, `e.g.` `bert-base-uncased`, the user can
+alternatively benchmark an arbitrary configuration of any available model class. In this case, a :obj:`list` of
+configurations must be inserted with the benchmark args as follows.
+
+.. code-block::
+
+    >>> ## PYTORCH CODE
+    >>> from transformers import PyTorchBenchmark, PyTorchBenchmarkArguments, BertConfig
+
+    >>> args = PyTorchBenchmarkArguments(models=["bert-base", "bert-384-hid", "bert-6-lay"], batch_sizes=[8], sequence_lengths=[8, 32, 128, 512])
+    >>> config_base = BertConfig()
+    >>> config_384_hid = BertConfig(hidden_size=384)
+    >>> config_6_lay = BertConfig(num_hidden_layers=6)
+
+    >>> benchmark = PyTorchBenchmark(args, configs=[config_base, config_384_hid, config_6_lay])
+    >>> benchmark.run()
+    ====================       INFERENCE - SPEED - RESULT       ====================
+    --------------------------------------------------------------------------------
+    Model Name             Batch Size     Seq Length       Time in s                  
+    --------------------------------------------------------------------------------
+    bert-base                  8              128            0.006
+    bert-base                  8              512            0.006
+    bert-base                  8              128            0.018     
+    bert-base                  8              512            0.088     
+    bert-384-hid              8               8             0.006     
+    bert-384-hid              8               32            0.006     
+    bert-384-hid              8              128            0.011     
+    bert-384-hid              8              512            0.054     
+    bert-6-lay                 8               8             0.003     
+    bert-6-lay                 8               32            0.004     
+    bert-6-lay                 8              128            0.009     
+    bert-6-lay                 8              512            0.044
+    --------------------------------------------------------------------------------
+
+    ====================      INFERENCE - MEMORY - RESULT       ====================
+    --------------------------------------------------------------------------------
+    Model Name             Batch Size     Seq Length      Memory in MB 
+    --------------------------------------------------------------------------------
+    bert-base                  8               8             1277
+    bert-base                  8               32            1281
+    bert-base                  8              128            1307     
+    bert-base                  8              512            1539     
+    bert-384-hid              8               8             1005     
+    bert-384-hid              8               32            1027     
+    bert-384-hid              8              128            1035     
+    bert-384-hid              8              512            1255     
+    bert-6-lay                 8               8             1097     
+    bert-6-lay                 8               32            1101     
+    bert-6-lay                 8              128            1127     
+    bert-6-lay                 8              512            1359
+    --------------------------------------------------------------------------------
+
+    ====================        ENVIRONMENT INFORMATION         ====================
+    - transformers_version: 2.11.0
+    - framework: PyTorch
+    - use_torchscript: False
+    - framework_version: 1.4.0
+    - python_version: 3.6.10
+    - system: Linux
+    - cpu: x86_64
+    - architecture: 64bit
+    - date: 2020-06-29
+    - time: 09:35:25.143267
+    - fp16: False
+    - use_multiprocessing: True
+    - only_pretrain_model: False
+    - cpu_ram_mb: 32088
+    - use_gpu: True
+    - num_gpus: 1
+    - gpu: TITAN RTX
+    - gpu_ram_mb: 24217
+    - gpu_power_watts: 280.0
+    - gpu_performance_state: 2
+    - use_tpu: False
+
+    >>> ## TENSORFLOW CODE
+    >>> from transformers import TensorFlowBenchmark, TensorFlowBenchmarkArguments, BertConfig
+
+    >>> args = TensorFlowBenchmarkArguments(models=["bert-base", "bert-384-hid", "bert-6-lay"], batch_sizes=[8], sequence_lengths=[8, 32, 128, 512])
+    >>> config_base = BertConfig()
+    >>> config_384_hid = BertConfig(hidden_size=384)
+    >>> config_6_lay = BertConfig(num_hidden_layers=6)
+
+    >>> benchmark = TensorFlowBenchmark(args, configs=[config_base, config_384_hid, config_6_lay])
+    >>> benchmark.run()
+    ====================       INFERENCE - SPEED - RESULT       ====================
+    --------------------------------------------------------------------------------
+    Model Name             Batch Size     Seq Length       Time in s                  
+    --------------------------------------------------------------------------------
+    bert-base                  8               8             0.005
+    bert-base                  8               32            0.008
+    bert-base                  8              128            0.022
+    bert-base                  8              512            0.106
+    bert-384-hid              8               8             0.005
+    bert-384-hid              8               32            0.007
+    bert-384-hid              8              128            0.018
+    bert-384-hid              8              512            0.064
+    bert-6-lay                 8               8             0.002
+    bert-6-lay                 8               32            0.003
+    bert-6-lay                 8              128            0.0011
+    bert-6-lay                 8              512            0.074
+    --------------------------------------------------------------------------------
+
+    ====================      INFERENCE - MEMORY - RESULT       ====================
+    --------------------------------------------------------------------------------
+    Model Name             Batch Size     Seq Length      Memory in MB 
+    --------------------------------------------------------------------------------
+    bert-base                  8               8             1330
+    bert-base                  8               32            1330
+    bert-base                  8              128            1330
+    bert-base                  8              512            1770
+    bert-384-hid              8               8             1330
+    bert-384-hid              8               32            1330
+    bert-384-hid              8              128            1330
+    bert-384-hid              8              512            1540
+    bert-6-lay                 8               8             1330
+    bert-6-lay                 8               32            1330
+    bert-6-lay                 8              128            1330
+    bert-6-lay                 8              512            1540
+    --------------------------------------------------------------------------------
+
+    ====================        ENVIRONMENT INFORMATION         ====================
+    - transformers_version: 2.11.0
+    - framework: Tensorflow
+    - use_xla: False
+    - framework_version: 2.2.0
+    - python_version: 3.6.10
+    - system: Linux
+    - cpu: x86_64
+    - architecture: 64bit
+    - date: 2020-06-29
+    - time: 09:38:15.487125
+    - fp16: False
+    - use_multiprocessing: True
+    - only_pretrain_model: False
+    - cpu_ram_mb: 32088
+    - use_gpu: True
+    - num_gpus: 1
+    - gpu: TITAN RTX
+    - gpu_ram_mb: 24217
+    - gpu_power_watts: 280.0
+    - gpu_performance_state: 2
+    - use_tpu: False
+
+
+Again, `inference time` and `required memory` for `inference` are measured, but this time for customized configurations
+of the :obj:`BertModel` class. This feature can especially be helpful when deciding for which configuration the model
+should be trained.
+
+
+Benchmark best practices
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+This section lists a couple of best practices one should be aware of when benchmarking a model.
+
+- Currently, only single device benchmarking is supported. When benchmarking on GPU, it is recommended that the user
+  specifies on which device the code should be run by setting the ``CUDA_VISIBLE_DEVICES`` environment variable in the
+  shell, `e.g.` ``export CUDA_VISIBLE_DEVICES=0`` before running the code.
+- The option :obj:`no_multi_processing` should only be set to :obj:`True` for testing and debugging. To ensure accurate
+  memory measurement it is recommended to run each memory benchmark in a separate process by making sure
+  :obj:`no_multi_processing` is set to :obj:`True`.
+- One should always state the environment information when sharing the results of a model benchmark. Results can vary
+  heavily between different GPU devices, library versions, etc., so that benchmark results on their own are not very
+  useful for the community.
+
+
+Sharing your benchmark
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Previously all available core models (10 at the time) have been benchmarked for `inference time`, across many different
+settings: using PyTorch, with and without TorchScript, using TensorFlow, with and without XLA. All of those tests were
+done across CPUs (except for TensorFlow XLA) and GPUs.
+
+The approach is detailed in the `following blogpost
+<https://medium.com/huggingface/benchmarking-transformers-pytorch-and-tensorflow-e2917fb891c2>`__ and the results are
+available `here
+<https://docs.google.com/spreadsheets/d/1sryqufw2D0XlUH4sq3e9Wnxu5EAQkaohzrJbd5HdQ_w/edit?usp=sharing>`__.
+
+With the new `benchmark` tools, it is easier than ever to share your benchmark results with the community `here
+<https://github.com/huggingface/transformers/blob/master/examples/benchmarking/README.md>`__.

docs/source/bertology.rst

@@ -0,0 +1,26 @@
+BERTology
+-----------------------------------------------------------------------------------------------------------------------
+
+There is a growing field of study concerned with investigating the inner working of large-scale transformers like BERT
+(that some call "BERTology"). Some good examples of this field are:
+
+
+* BERT Rediscovers the Classical NLP Pipeline by Ian Tenney, Dipanjan Das, Ellie Pavlick:
+  https://arxiv.org/abs/1905.05950
+* Are Sixteen Heads Really Better than One? by Paul Michel, Omer Levy, Graham Neubig: https://arxiv.org/abs/1905.10650
+* What Does BERT Look At? An Analysis of BERT's Attention by Kevin Clark, Urvashi Khandelwal, Omer Levy, Christopher D.
+  Manning: https://arxiv.org/abs/1906.04341
+
+In order to help this new field develop, we have included a few additional features in the BERT/GPT/GPT-2 models to
+help people access the inner representations, mainly adapted from the great work of Paul Michel
+(https://arxiv.org/abs/1905.10650):
+
+
+* accessing all the hidden-states of BERT/GPT/GPT-2,
+* accessing all the attention weights for each head of BERT/GPT/GPT-2,
+* retrieving heads output values and gradients to be able to compute head importance score and prune head as explained
+  in https://arxiv.org/abs/1905.10650.
+
+To help you understand and use these features, we have added a specific example script: `bertology.py
+<https://github.com/huggingface/transformers/blob/master/examples/bertology/run_bertology.py>`_ while extract
+information and prune a model pre-trained on GLUE.

docs/source/conf.py

@@ -0,0 +1,198 @@
+# -*- coding: utf-8 -*-
+#
+# Configuration file for the Sphinx documentation builder.
+#
+# This file does only contain a selection of the most common options. For a
+# full list see the documentation:
+# http://www.sphinx-doc.org/en/master/config
+
+# -- Path setup --------------------------------------------------------------
+
+# If extensions (or modules to document with autodoc) are in another directory,
+# add these directories to sys.path here. If the directory is relative to the
+# documentation root, use os.path.abspath to make it absolute, like shown here.
+#
+import os
+import sys
+sys.path.insert(0, os.path.abspath('../../src'))
+
+
+# -- Project information -----------------------------------------------------
+
+project = u'transformers'
+copyright = u'2020, huggingface'
+author = u'huggingface'
+
+# The short X.Y version
+version = u''
+# The full version, including alpha/beta/rc tags
+release = u'4.0.1'
+
+
+# -- General configuration ---------------------------------------------------
+
+# If your documentation needs a minimal Sphinx version, state it here.
+#
+# needs_sphinx = '1.0'
+
+# Add any Sphinx extension module names here, as strings. They can be
+# extensions coming with Sphinx (named 'sphinx.ext.*') or your custom
+# ones.
+extensions = [
+    'sphinx.ext.autodoc',
+    'sphinx.ext.coverage',
+    'sphinx.ext.napoleon',
+    'recommonmark',
+    'sphinx.ext.viewcode',
+    'sphinx_markdown_tables',
+    'sphinx_copybutton'
+]
+
+# Add any paths that contain templates here, relative to this directory.
+templates_path = ['_templates']
+
+# The suffix(es) of source filenames.
+# You can specify multiple suffix as a list of string:
+#
+source_suffix = ['.rst', '.md']
+# source_suffix = '.rst'
+
+# The master toctree document.
+master_doc = 'index'
+
+# The language for content autogenerated by Sphinx. Refer to documentation
+# for a list of supported languages.
+#
+# This is also used if you do content translation via gettext catalogs.
+# Usually you set "language" from the command line for these cases.
+language = None
+
+# List of patterns, relative to source directory, that match files and
+# directories to ignore when looking for source files.
+# This pattern also affects html_static_path and html_extra_path.
+exclude_patterns = [u'_build', 'Thumbs.db', '.DS_Store']
+
+# The name of the Pygments (syntax highlighting) style to use.
+pygments_style = None
+
+# Remove the prompt when copying examples
+copybutton_prompt_text = r">>> |\.\.\. "
+copybutton_prompt_is_regexp = True
+
+# -- Options for HTML output -------------------------------------------------
+
+# The theme to use for HTML and HTML Help pages.  See the documentation for
+# a list of builtin themes.
+#
+html_theme = 'sphinx_rtd_theme'
+
+# Theme options are theme-specific and customize the look and feel of a theme
+# further.  For a list of options available for each theme, see the
+# documentation.
+#
+html_theme_options = {
+    'analytics_id': 'UA-83738774-2'
+}
+
+# Add any paths that contain custom static files (such as style sheets) here,
+# relative to this directory. They are copied after the builtin static files,
+# so a file named "default.css" will overwrite the builtin "default.css".
+html_static_path = ['_static']
+
+# Custom sidebar templates, must be a dictionary that maps document names
+# to template names.
+#
+# The default sidebars (for documents that don't match any pattern) are
+# defined by theme itself.  Builtin themes are using these templates by
+# default: ``['localtoc.html', 'relations.html', 'sourcelink.html',
+# 'searchbox.html']``.
+#
+# html_sidebars = {}
+
+# This must be the name of an image file (path relative to the configuration 
+# directory) that is the favicon of the docs. Modern browsers use this as 
+# the icon for tabs, windows and bookmarks. It should be a Windows-style 
+# icon file (.ico).
+html_favicon = 'favicon.ico'
+
+
+# -- Options for HTMLHelp output ---------------------------------------------
+
+# Output file base name for HTML help builder.
+htmlhelp_basename = 'transformersdoc'
+
+
+# -- Options for LaTeX output ------------------------------------------------
+
+latex_elements = {
+    # The paper size ('letterpaper' or 'a4paper').
+    #
+    # 'papersize': 'letterpaper',
+
+    # The font size ('10pt', '11pt' or '12pt').
+    #
+    # 'pointsize': '10pt',
+
+    # Additional stuff for the LaTeX preamble.
+    #
+    # 'preamble': '',
+
+    # Latex figure (float) alignment
+    #
+    # 'figure_align': 'htbp',
+}
+
+# Grouping the document tree into LaTeX files. List of tuples
+# (source start file, target name, title,
+#  author, documentclass [howto, manual, or own class]).
+latex_documents = [
+    (master_doc, 'transformers.tex', u'transformers Documentation',
+     u'huggingface', 'manual'),
+]
+
+
+# -- Options for manual page output ------------------------------------------
+
+# One entry per manual page. List of tuples
+# (source start file, name, description, authors, manual section).
+man_pages = [
+    (master_doc, 'transformers', u'transformers Documentation',
+     [author], 1)
+]
+
+
+# -- Options for Texinfo output ----------------------------------------------
+
+# Grouping the document tree into Texinfo files. List of tuples
+# (source start file, target name, title, author,
+#  dir menu entry, description, category)
+texinfo_documents = [
+    (master_doc, 'transformers', u'transformers Documentation',
+     author, 'transformers', 'One line description of project.',
+     'Miscellaneous'),
+]
+
+
+# -- Options for Epub output -------------------------------------------------
+
+# Bibliographic Dublin Core info.
+epub_title = project
+
+# The unique identifier of the text. This can be a ISBN number
+# or the project homepage.
+#
+# epub_identifier = ''
+
+# A unique identification for the text.
+#
+# epub_uid = ''
+
+# A list of files that should not be packed into the epub file.
+epub_exclude_files = ['search.html']
+
+def setup(app):
+    app.add_css_file('css/huggingface.css')
+    app.add_css_file('css/code-snippets.css')
+    app.add_js_file('js/custom.js')
+
+# -- Extension configuration -------------------------------------------------

docs/source/contributing.md

@@ -0,0 +1 @@
+../../CONTRIBUTING.md

docs/source/converting_tensorflow_models.rst

@@ -0,0 +1,160 @@
+Converting Tensorflow Checkpoints
+=======================================================================================================================
+
+A command-line interface is provided to convert original Bert/GPT/GPT-2/Transformer-XL/XLNet/XLM checkpoints in models
+than be loaded using the ``from_pretrained`` methods of the library.
+
+.. note::
+    Since 2.3.0 the conversion script is now part of the transformers CLI (**transformers-cli**) available in any
+    transformers >= 2.3.0 installation.
+
+    The documentation below reflects the **transformers-cli convert** command format.
+
+BERT
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+You can convert any TensorFlow checkpoint for BERT (in particular `the pre-trained models released by Google
+<https://github.com/google-research/bert#pre-trained-models>`_\ ) in a PyTorch save file by using the
+`convert_bert_original_tf_checkpoint_to_pytorch.py
+<https://github.com/huggingface/transformers/blob/master/src/transformers/convert_bert_original_tf_checkpoint_to_pytorch.py>`_
+script.
+
+This CLI takes as input a TensorFlow checkpoint (three files starting with ``bert_model.ckpt``\ ) and the associated
+configuration file (\ ``bert_config.json``\ ), and creates a PyTorch model for this configuration, loads the weights
+from the TensorFlow checkpoint in the PyTorch model and saves the resulting model in a standard PyTorch save file that
+can be imported using ``torch.load()`` (see examples in `run_bert_extract_features.py
+<https://github.com/huggingface/pytorch-pretrained-BERT/tree/master/examples/run_bert_extract_features.py>`_\ ,
+`run_bert_classifier.py
+<https://github.com/huggingface/pytorch-pretrained-BERT/tree/master/examples/run_bert_classifier.py>`_ and
+`run_bert_squad.py <https://github.com/huggingface/pytorch-pretrained-BERT/tree/master/examples/run_bert_squad.py>`_\
+).
+
+You only need to run this conversion script **once** to get a PyTorch model. You can then disregard the TensorFlow
+checkpoint (the three files starting with ``bert_model.ckpt``\ ) but be sure to keep the configuration file (\
+``bert_config.json``\ ) and the vocabulary file (\ ``vocab.txt``\ ) as these are needed for the PyTorch model too.
+
+To run this specific conversion script you will need to have TensorFlow and PyTorch installed (\ ``pip install
+tensorflow``\ ). The rest of the repository only requires PyTorch.
+
+Here is an example of the conversion process for a pre-trained ``BERT-Base Uncased`` model:
+
+.. code-block:: shell
+
+   export BERT_BASE_DIR=/path/to/bert/uncased_L-12_H-768_A-12
+
+   transformers-cli convert --model_type bert \
+     --tf_checkpoint $BERT_BASE_DIR/bert_model.ckpt \
+     --config $BERT_BASE_DIR/bert_config.json \
+     --pytorch_dump_output $BERT_BASE_DIR/pytorch_model.bin
+
+You can download Google's pre-trained models for the conversion `here
+<https://github.com/google-research/bert#pre-trained-models>`__.
+
+ALBERT
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Convert TensorFlow model checkpoints of ALBERT to PyTorch using the
+`convert_albert_original_tf_checkpoint_to_pytorch.py
+<https://github.com/huggingface/transformers/blob/master/src/transformers/convert_bert_original_tf_checkpoint_to_pytorch.py>`_
+script.
+
+The CLI takes as input a TensorFlow checkpoint (three files starting with ``model.ckpt-best``\ ) and the accompanying
+configuration file (\ ``albert_config.json``\ ), then creates and saves a PyTorch model. To run this conversion you
+will need to have TensorFlow and PyTorch installed.
+
+Here is an example of the conversion process for the pre-trained ``ALBERT Base`` model:
+
+.. code-block:: shell
+
+   export ALBERT_BASE_DIR=/path/to/albert/albert_base
+
+   transformers-cli convert --model_type albert \
+     --tf_checkpoint $ALBERT_BASE_DIR/model.ckpt-best \
+     --config $ALBERT_BASE_DIR/albert_config.json \
+     --pytorch_dump_output $ALBERT_BASE_DIR/pytorch_model.bin
+
+You can download Google's pre-trained models for the conversion `here
+<https://github.com/google-research/albert#pre-trained-models>`__.
+
+OpenAI GPT
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Here is an example of the conversion process for a pre-trained OpenAI GPT model, assuming that your NumPy checkpoint
+save as the same format than OpenAI pretrained model (see `here <https://github.com/openai/finetune-transformer-lm>`__\
+)
+
+.. code-block:: shell
+
+   export OPENAI_GPT_CHECKPOINT_FOLDER_PATH=/path/to/openai/pretrained/numpy/weights
+
+   transformers-cli convert --model_type gpt \
+     --tf_checkpoint $OPENAI_GPT_CHECKPOINT_FOLDER_PATH \
+     --pytorch_dump_output $PYTORCH_DUMP_OUTPUT \
+     [--config OPENAI_GPT_CONFIG] \
+     [--finetuning_task_name OPENAI_GPT_FINETUNED_TASK] \
+
+
+OpenAI GPT-2
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Here is an example of the conversion process for a pre-trained OpenAI GPT-2 model (see `here
+<https://github.com/openai/gpt-2>`__\ )
+
+.. code-block:: shell
+
+   export OPENAI_GPT2_CHECKPOINT_PATH=/path/to/gpt2/pretrained/weights
+
+   transformers-cli convert --model_type gpt2 \
+     --tf_checkpoint $OPENAI_GPT2_CHECKPOINT_PATH \
+     --pytorch_dump_output $PYTORCH_DUMP_OUTPUT \
+     [--config OPENAI_GPT2_CONFIG] \
+     [--finetuning_task_name OPENAI_GPT2_FINETUNED_TASK]
+
+Transformer-XL
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Here is an example of the conversion process for a pre-trained Transformer-XL model (see `here
+<https://github.com/kimiyoung/transformer-xl/tree/master/tf#obtain-and-evaluate-pretrained-sota-models>`__\ )
+
+.. code-block:: shell
+
+   export TRANSFO_XL_CHECKPOINT_FOLDER_PATH=/path/to/transfo/xl/checkpoint
+
+   transformers-cli convert --model_type transfo_xl \
+     --tf_checkpoint $TRANSFO_XL_CHECKPOINT_FOLDER_PATH \
+     --pytorch_dump_output $PYTORCH_DUMP_OUTPUT \
+     [--config TRANSFO_XL_CONFIG] \
+     [--finetuning_task_name TRANSFO_XL_FINETUNED_TASK]
+
+
+XLNet
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Here is an example of the conversion process for a pre-trained XLNet model:
+
+.. code-block:: shell
+
+   export TRANSFO_XL_CHECKPOINT_PATH=/path/to/xlnet/checkpoint
+   export TRANSFO_XL_CONFIG_PATH=/path/to/xlnet/config
+
+   transformers-cli convert --model_type xlnet \
+     --tf_checkpoint $TRANSFO_XL_CHECKPOINT_PATH \
+     --config $TRANSFO_XL_CONFIG_PATH \
+     --pytorch_dump_output $PYTORCH_DUMP_OUTPUT \
+     [--finetuning_task_name XLNET_FINETUNED_TASK] \
+
+
+XLM
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Here is an example of the conversion process for a pre-trained XLM model:
+
+.. code-block:: shell
+
+   export XLM_CHECKPOINT_PATH=/path/to/xlm/checkpoint
+
+   transformers-cli convert --model_type xlm \
+     --tf_checkpoint $XLM_CHECKPOINT_PATH \
+     --pytorch_dump_output $PYTORCH_DUMP_OUTPUT
+    [--config XML_CONFIG] \
+    [--finetuning_task_name XML_FINETUNED_TASK]

docs/source/custom_datasets.rst

@@ -0,0 +1,714 @@
+Fine-tuning with custom datasets
+=======================================================================================================================
+
+.. note::
+
+    The datasets used in this tutorial are available and can be more easily accessed using the `🤗 NLP library
+    <https://github.com/huggingface/nlp>`_. We do not use this library to access the datasets here since this tutorial
+    meant to illustrate how to work with your own data. A brief of introduction can be found at the end of the tutorial
+    in the section ":ref:`nlplib`".
+
+This tutorial will take you through several examples of using 🤗 Transformers models with your own datasets. The guide
+shows one of many valid workflows for using these models and is meant to be illustrative rather than definitive. We
+show examples of reading in several data formats, preprocessing the data for several types of tasks, and then preparing
+the data into PyTorch/TensorFlow ``Dataset`` objects which can easily be used either with
+:class:`~transformers.Trainer`/:class:`~transformers.TFTrainer` or with native PyTorch/TensorFlow.
+
+We include several examples, each of which demonstrates a different type of common downstream task:
+
+  - :ref:`seq_imdb`
+  - :ref:`tok_ner`
+  - :ref:`qa_squad`
+  - :ref:`resources`
+
+.. _seq_imdb:
+
+Sequence Classification with IMDb Reviews
+-----------------------------------------------------------------------------------------------------------------------
+
+.. note::
+
+    This dataset can be explored in the Hugging Face model hub (`IMDb <https://huggingface.co/datasets/imdb>`_), and
+    can be alternatively downloaded with the 🤗 NLP library with ``load_dataset("imdb")``.
+
+In this example, we'll show how to download, tokenize, and train a model on the IMDb reviews dataset. This task takes
+the text of a review and requires the model to predict whether the sentiment of the review is positive or negative.
+Let's start by downloading the dataset from the `Large Movie Review Dataset
+<http://ai.stanford.edu/~amaas/data/sentiment/>`_ webpage.
+
+.. code-block:: bash
+
+    wget http://ai.stanford.edu/~amaas/data/sentiment/aclImdb_v1.tar.gz
+    tar -xf aclImdb_v1.tar.gz
+
+This data is organized into ``pos`` and ``neg`` folders with one text file per example. Let's write a function that can
+read this in.
+
+.. code-block:: python
+
+    from pathlib import Path
+
+    def read_imdb_split(split_dir):
+        split_dir = Path(split_dir)
+        texts = []
+        labels = []
+        for label_dir in ["pos", "neg"]:
+            for text_file in (split_dir/label_dir).iterdir():
+                texts.append(text_file.read_text())
+                labels.append(0 if label_dir is "neg" else 1)
+
+        return texts, labels
+
+    train_texts, train_labels = read_imdb_split('aclImdb/train')
+    test_texts, test_labels = read_imdb_split('aclImdb/test')
+
+We now have a train and test dataset, but let's also also create a validation set which we can use for for evaluation
+and tuning without training our test set results. Sklearn has a convenient utility for creating such splits:
+
+.. code-block:: python
+
+    from sklearn.model_selection import train_test_split
+    train_texts, val_texts, train_labels, val_labels = train_test_split(train_texts, train_labels, test_size=.2)
+
+Alright, we've read in our dataset. Now let's tackle tokenization. We'll eventually train a classifier using
+pre-trained DistilBert, so let's use the DistilBert tokenizer.
+
+.. code-block:: python
+
+    from transformers import DistilBertTokenizerFast
+    tokenizer = DistilBertTokenizerFast.from_pretrained('distilbert-base-uncased')
+
+Now we can simply pass our texts to the tokenizer. We'll pass ``truncation=True`` and ``padding=True``, which will
+ensure that all of our sequences are padded to the same length and are truncated to be no longer model's maximum input
+length. This will allow us to feed batches of sequences into the model at the same time.
+
+.. code-block:: python
+
+    train_encodings = tokenizer(train_texts, truncation=True, padding=True)
+    val_encodings = tokenizer(val_texts, truncation=True, padding=True)
+    test_encodings = tokenizer(test_texts, truncation=True, padding=True)
+
+Now, let's turn our labels and encodings into a Dataset object. In PyTorch, this is done by subclassing a
+``torch.utils.data.Dataset`` object and implementing ``__len__`` and ``__getitem__``. In TensorFlow, we pass our input
+encodings and labels to the ``from_tensor_slices`` constructor method. We put the data in this format so that the data
+can be easily batched such that each key in the batch encoding corresponds to a named parameter of the
+:meth:`~transformers.DistilBertForSequenceClassification.forward` method of the model we will train.
+
+.. code-block:: python
+
+    ## PYTORCH CODE
+    import torch
+
+    class IMDbDataset(torch.utils.data.Dataset):
+        def __init__(self, encodings, labels):
+            self.encodings = encodings
+            self.labels = labels
+
+        def __getitem__(self, idx):
+            item = {key: torch.tensor(val[idx]) for key, val in self.encodings.items()}
+            item['labels'] = torch.tensor(self.labels[idx])
+            return item
+
+        def __len__(self):
+            return len(self.labels)
+
+    train_dataset = IMDbDataset(train_encodings, train_labels)
+    val_dataset = IMDbDataset(val_encodings, val_labels)
+    test_dataset = IMDbDataset(test_encodings, test_labels)
+    ## TENSORFLOW CODE
+    import tensorflow as tf
+
+    train_dataset = tf.data.Dataset.from_tensor_slices((
+        dict(train_encodings),
+        train_labels
+    ))
+    val_dataset = tf.data.Dataset.from_tensor_slices((
+        dict(val_encodings),
+        val_labels
+    ))
+    test_dataset = tf.data.Dataset.from_tensor_slices((
+        dict(test_encodings),
+        test_labels
+    ))
+
+Now that our datasets our ready, we can fine-tune a model either with the 🤗
+:class:`~transformers.Trainer`/:class:`~transformers.TFTrainer` or with native PyTorch/TensorFlow. See :doc:`training
+<training>`.
+
+.. _ft_trainer:
+
+Fine-tuning with Trainer
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+The steps above prepared the datasets in the way that the trainer is expected. Now all we need to do is create a model
+to fine-tune, define the :class:`~transformers.TrainingArguments`/:class:`~transformers.TFTrainingArguments` and
+instantiate a :class:`~transformers.Trainer`/:class:`~transformers.TFTrainer`.
+
+.. code-block:: python
+
+    ## PYTORCH CODE
+    from transformers import DistilBertForSequenceClassification, Trainer, TrainingArguments
+
+    training_args = TrainingArguments(
+        output_dir='./results',          # output directory
+        num_train_epochs=3,              # total number of training epochs
+        per_device_train_batch_size=16,  # batch size per device during training
+        per_device_eval_batch_size=64,   # batch size for evaluation
+        warmup_steps=500,                # number of warmup steps for learning rate scheduler
+        weight_decay=0.01,               # strength of weight decay
+        logging_dir='./logs',            # directory for storing logs
+        logging_steps=10,
+    )
+
+    model = DistilBertForSequenceClassification.from_pretrained("distilbert-base-uncased")
+
+    trainer = Trainer(
+        model=model,                         # the instantiated 🤗 Transformers model to be trained
+        args=training_args,                  # training arguments, defined above
+        train_dataset=train_dataset,         # training dataset
+        eval_dataset=val_dataset             # evaluation dataset
+    )
+
+    trainer.train()
+    ## TENSORFLOW CODE
+    from transformers import TFDistilBertForSequenceClassification, TFTrainer, TFTrainingArguments
+
+    training_args = TFTrainingArguments(
+        output_dir='./results',          # output directory
+        num_train_epochs=3,              # total number of training epochs
+        per_device_train_batch_size=16,  # batch size per device during training
+        per_device_eval_batch_size=64,   # batch size for evaluation
+        warmup_steps=500,                # number of warmup steps for learning rate scheduler
+        weight_decay=0.01,               # strength of weight decay
+        logging_dir='./logs',            # directory for storing logs
+        logging_steps=10,
+    )
+
+    with training_args.strategy.scope():
+        model = TFDistilBertForSequenceClassification.from_pretrained("distilbert-base-uncased")
+
+    trainer = TFTrainer(
+        model=model,                         # the instantiated 🤗 Transformers model to be trained
+        args=training_args,                  # training arguments, defined above
+        train_dataset=train_dataset,         # training dataset
+        eval_dataset=val_dataset             # evaluation dataset
+    )
+
+    trainer.train()
+
+.. _ft_native:
+
+Fine-tuning with native PyTorch/TensorFlow
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+We can also train use native PyTorch or TensorFlow:
+
+.. code-block:: python
+
+    ## PYTORCH CODE
+    from torch.utils.data import DataLoader
+    from transformers import DistilBertForSequenceClassification, AdamW
+
+    device = torch.device('cuda') if torch.cuda.is_available() else torch.device('cpu')
+
+    model = DistilBertForSequenceClassification.from_pretrained('distilbert-base-uncased')
+    model.to(device)
+    model.train()
+
+    train_loader = DataLoader(train_dataset, batch_size=16, shuffle=True)
+
+    optim = AdamW(model.parameters(), lr=5e-5)
+
+    for epoch in range(3):
+        for batch in train_loader:
+            optim.zero_grad()
+            input_ids = batch['input_ids'].to(device)
+            attention_mask = batch['attention_mask'].to(device)
+            labels = batch['labels'].to(device)
+            outputs = model(input_ids, attention_mask=attention_mask, labels=labels)
+            loss = outputs[0]
+            loss.backward()
+            optim.step()
+
+    model.eval()
+    ## TENSORFLOW CODE
+    from transformers import TFDistilBertForSequenceClassification
+
+    model = TFDistilBertForSequenceClassification.from_pretrained('distilbert-base-uncased')
+
+    optimizer = tf.keras.optimizers.Adam(learning_rate=5e-5)
+    model.compile(optimizer=optimizer, loss=model.compute_loss) # can also use any keras loss fn
+    model.fit(train_dataset.shuffle(1000).batch(16), epochs=3, batch_size=16)
+
+.. _tok_ner:
+
+Token Classification with W-NUT Emerging Entities
+-----------------------------------------------------------------------------------------------------------------------
+
+.. note::
+
+    This dataset can be explored in the Hugging Face model hub (`WNUT-17 <https://huggingface.co/datasets/wnut_17>`_),
+    and can be alternatively downloaded with the 🤗 NLP library with ``load_dataset("wnut_17")``.
+
+Next we will look at token classification. Rather than classifying an entire sequence, this task classifies token by
+token. We'll demonstrate how to do this with `Named Entity Recognition
+<http://nlpprogress.com/english/named_entity_recognition.html>`_, which involves identifying tokens which correspond to
+a predefined set of "entities". Specifically, we'll use the `W-NUT Emerging and Rare entities
+<http://noisy-text.github.io/2017/emerging-rare-entities.html>`_ corpus. The data is given as a collection of
+pre-tokenized documents where each token is assigned a tag.
+
+Let's start by downloading the data.
+
+.. code-block:: bash
+
+    wget http://noisy-text.github.io/2017/files/wnut17train.conll
+
+In this case, we'll just download the train set, which is a single text file. Each line of the file contains either (1)
+a word and tag separated by a tab, or (2) a blank line indicating the end of a document. Let's write a function to read
+this in. We'll take in the file path and return ``token_docs`` which is a list of lists of token strings, and
+``token_tags`` which is a list of lists of tag strings.
+
+.. code-block:: python
+
+    from pathlib import Path
+    import re
+
+    def read_wnut(file_path):
+        file_path = Path(file_path)
+
+        raw_text = file_path.read_text().strip()
+        raw_docs = re.split(r'\n\t?\n', raw_text)
+        token_docs = []
+        tag_docs = []
+        for doc in raw_docs:
+            tokens = []
+            tags = []
+            for line in doc.split('\n'):
+                token, tag = line.split('\t')
+                tokens.append(token)
+                tags.append(tag)
+            token_docs.append(tokens)
+            tag_docs.append(tags)
+
+        return token_docs, tag_docs
+
+    texts, tags = read_wnut('wnut17train.conll')
+
+Just to see what this data looks like, let's take a look at a segment of the first document.
+
+.. code-block:: python
+
+    >>> print(texts[0][10:17], tags[0][10:17], sep='\n')
+    ['for', 'two', 'weeks', '.', 'Empire', 'State', 'Building']
+    ['O', 'O', 'O', 'O', 'B-location', 'I-location', 'I-location']
+
+``location`` is an entity type, ``B-`` indicates the beginning of an entity, and ``I-`` indicates consecutive positions
+of the same entity ("Empire State Building" is considered one entity). ``O`` indicates the token does not correspond to
+any entity.
+
+Now that we've read the data in, let's create a train/validation split:
+
+.. code-block:: python
+
+    from sklearn.model_selection import train_test_split
+    train_texts, val_texts, train_tags, val_tags = train_test_split(texts, tags, test_size=.2)
+
+Next, let's create encodings for our tokens and tags. For the tags, we can start by just create a simple mapping which
+we'll use in a moment:
+
+.. code-block:: python
+
+    unique_tags = set(tag for doc in tags for tag in doc)
+    tag2id = {tag: id for id, tag in enumerate(unique_tags)}
+    id2tag = {id: tag for tag, id in tag2id.items()}
+
+To encode the tokens, we'll use a pre-trained DistilBert tokenizer. We can tell the tokenizer that we're dealing with
+ready-split tokens rather than full sentence strings by passing ``is_split_into_words=True``. We'll also pass
+``padding=True`` and ``truncation=True`` to pad the sequences to be the same length. Lastly, we can tell the model to
+return information about the tokens which are split by the wordpiece tokenization process, which we will need in a
+moment.
+
+.. code-block:: python
+
+    from transformers import DistilBertTokenizerFast
+    tokenizer = DistilBertTokenizerFast.from_pretrained('distilbert-base-cased')
+    train_encodings = tokenizer(train_texts, is_split_into_words=True, return_offsets_mapping=True, padding=True, truncation=True)
+    val_encodings = tokenizer(val_texts, is_split_into_words=True, return_offsets_mapping=True, padding=True, truncation=True)
+
+Great, so now our tokens are nicely encoded in the format that they need to be in to feed them into our DistilBert
+model below.
+
+Now we arrive at a common obstacle with using pre-trained models for token-level classification: many of the tokens in
+the W-NUT corpus are not in DistilBert's vocabulary. Bert and many models like it use a method called WordPiece
+Tokenization, meaning that single words are split into multiple tokens such that each token is likely to be in the
+vocabulary. For example, DistilBert's tokenizer would split the Twitter handle ``@huggingface`` into the tokens ``['@',
+'hugging', '##face']``. This is a problem for us because we have exactly one tag per token. If the tokenizer splits a
+token into multiple sub-tokens, then we will end up with a mismatch between our tokens and our labels.
+
+One way to handle this is to only train on the tag labels for the first subtoken of a split token. We can do this in 🤗
+Transformers by setting the labels we wish to ignore to ``-100``. In the example above, if the label for
+``@HuggingFace`` is ``3`` (indexing ``B-corporation``), we would set the labels of ``['@', 'hugging', '##face']`` to
+``[3, -100, -100]``.
+
+Let's write a function to do this. This is where we will use the ``offset_mapping`` from the tokenizer as mentioned
+above. For each sub-token returned by the tokenizer, the offset mapping gives us a tuple indicating the sub-token's
+start position and end position relative to the original token it was split from. That means that if the first position
+in the tuple is anything other than ``0``, we will set its corresponding label to ``-100``. While we're at it, we can
+also set labels to ``-100`` if the second position of the offset mapping is ``0``, since this means it must be a
+special token like ``[PAD]`` or ``[CLS]``.
+
+.. note::
+
+    Due to a recently fixed bug, -1 must be used instead of -100 when using TensorFlow in 🤗 Transformers <= 3.02.
+
+.. code-block:: python
+
+    import numpy as np
+
+    def encode_tags(tags, encodings):
+        labels = [[tag2id[tag] for tag in doc] for doc in tags]
+        encoded_labels = []
+        for doc_labels, doc_offset in zip(labels, encodings.offset_mapping):
+            # create an empty array of -100
+            doc_enc_labels = np.ones(len(doc_offset),dtype=int) * -100
+            arr_offset = np.array(doc_offset)
+
+            # set labels whose first offset position is 0 and the second is not 0
+            doc_enc_labels[(arr_offset[:,0] == 0) & (arr_offset[:,1] != 0)] = doc_labels
+            encoded_labels.append(doc_enc_labels.tolist())
+
+        return encoded_labels
+
+    train_labels = encode_tags(train_tags, train_encodings)
+    val_labels = encode_tags(val_tags, val_encodings)
+
+The hard part is now done. Just as in the sequence classification example above, we can create a dataset object:
+
+.. code-block:: python
+
+    ## PYTORCH CODE
+    import torch
+
+    class WNUTDataset(torch.utils.data.Dataset):
+        def __init__(self, encodings, labels):
+            self.encodings = encodings
+            self.labels = labels
+
+        def __getitem__(self, idx):
+            item = {key: torch.tensor(val[idx]) for key, val in self.encodings.items()}
+            item['labels'] = torch.tensor(self.labels[idx])
+            return item
+
+        def __len__(self):
+            return len(self.labels)
+
+    train_encodings.pop("offset_mapping") # we don't want to pass this to the model
+    val_encodings.pop("offset_mapping")
+    train_dataset = WNUTDataset(train_encodings, train_labels)
+    val_dataset = WNUTDataset(val_encodings, val_labels)
+    ## TENSORFLOW CODE
+    import tensorflow as tf
+
+    train_encodings.pop("offset_mapping") # we don't want to pass this to the model
+    val_encodings.pop("offset_mapping")
+
+    train_dataset = tf.data.Dataset.from_tensor_slices((
+        dict(train_encodings),
+        train_labels
+    ))
+    val_dataset = tf.data.Dataset.from_tensor_slices((
+        dict(val_encodings),
+        val_labels
+    ))
+
+Now load in a token classification model and specify the number of labels:
+
+.. code-block:: python
+
+    ## PYTORCH CODE
+    from transformers import DistilBertForTokenClassification
+    model = DistilBertForTokenClassification.from_pretrained('distilbert-base-cased', num_labels=len(unique_tags))
+    ## TENSORFLOW CODE
+    from transformers import TFDistilBertForTokenClassification
+    model = TFDistilBertForTokenClassification.from_pretrained('distilbert-base-cased', num_labels=len(unique_tags))
+
+The data and model are both ready to go. You can train the model either with
+:class:`~transformers.Trainer`/:class:`~transformers.TFTrainer` or with native PyTorch/TensorFlow, exactly as in the
+sequence classification example above.
+
+  - :ref:`ft_trainer`
+  - :ref:`ft_native`
+
+.. _qa_squad:
+
+Question Answering with SQuAD 2.0
+-----------------------------------------------------------------------------------------------------------------------
+
+.. note::
+
+    This dataset can be explored in the Hugging Face model hub (`SQuAD V2
+    <https://huggingface.co/datasets/squad_v2>`_), and can be alternatively downloaded with the 🤗 NLP library with
+    ``load_dataset("squad_v2")``.
+
+Question answering comes in many forms. In this example, we'll look at the particular type of extractive QA that
+involves answering a question about a passage by highlighting the segment of the passage that answers the question.
+This involves fine-tuning a model which predicts a start position and an end position in the passage. We will use the
+`Stanford Question Answering Dataset (SQuAD) 2.0 <https://rajpurkar.github.io/SQuAD-explorer/>`_.
+
+We will start by downloading the data:
+
+.. code-block:: bash
+
+    mkdir squad
+    wget https://rajpurkar.github.io/SQuAD-explorer/dataset/train-v2.0.json -O squad/train-v2.0.json
+    wget https://rajpurkar.github.io/SQuAD-explorer/dataset/dev-v2.0.json -O squad/dev-v2.0.json
+
+Each split is in a structured json file with a number of questions and answers for each passage (or context). We'll
+take this apart into parallel lists of contexts, questions, and answers (note that the contexts here are repeated since
+there are multiple questions per context):
+
+.. code-block:: python
+
+    import json
+    from pathlib import Path
+
+    def read_squad(path):
+        path = Path(path)
+        with open(path, 'rb') as f:
+            squad_dict = json.load(f)
+
+        contexts = []
+        questions = []
+        answers = []
+        for group in squad_dict['data']:
+            for passage in group['paragraphs']:
+                context = passage['context']
+                for qa in passage['qas']:
+                    question = qa['question']
+                    for answer in qa['answers']:
+                        contexts.append(context)
+                        questions.append(question)
+                        answers.append(answer)
+
+        return contexts, questions, answers
+
+    train_contexts, train_questions, train_answers = read_squad('squad/train-v2.0.json')
+    val_contexts, val_questions, val_answers = read_squad('squad/dev-v2.0.json')
+
+The contexts and questions are just strings. The answers are dicts containing the subsequence of the passage with the
+correct answer as well as an integer indicating the character at which the answer begins. In order to train a model on
+this data we need (1) the tokenized context/question pairs, and (2) integers indicating at which *token* positions the
+answer begins and ends.
+
+First, let's get the *character* position at which the answer ends in the passage (we are given the starting position).
+Sometimes SQuAD answers are off by one or two characters, so we will also adjust for that.
+
+.. code-block:: python
+
+    def add_end_idx(answers, contexts):
+        for answer, context in zip(answers, contexts):
+            gold_text = answer['text']
+            start_idx = answer['answer_start']
+            end_idx = start_idx + len(gold_text)
+
+            # sometimes squad answers are off by a character or two – fix this
+            if context[start_idx:end_idx] == gold_text:
+                answer['answer_end'] = end_idx
+            elif context[start_idx-1:end_idx-1] == gold_text:
+                answer['answer_start'] = start_idx - 1
+                answer['answer_end'] = end_idx - 1     # When the gold label is off by one character
+            elif context[start_idx-2:end_idx-2] == gold_text:
+                answer['answer_start'] = start_idx - 2
+                answer['answer_end'] = end_idx - 2     # When the gold label is off by two characters
+
+    add_end_idx(train_answers, train_contexts)
+    add_end_idx(val_answers, val_contexts)
+
+Now ``train_answers`` and ``val_answers`` include the character end positions and the corrected start positions. Next,
+let's tokenize our context/question pairs. 🤗 Tokenizers can accept parallel lists of sequences and encode them together
+as sequence pairs.
+
+.. code-block:: python
+
+    from transformers import DistilBertTokenizerFast
+    tokenizer = DistilBertTokenizerFast.from_pretrained('distilbert-base-uncased')
+
+    train_encodings = tokenizer(train_contexts, train_questions, truncation=True, padding=True)
+    val_encodings = tokenizer(val_contexts, val_questions, truncation=True, padding=True)
+
+Next we need to convert our character start/end positions to token start/end positions. When using 🤗 Fast Tokenizers,
+we can use the built in :func:`~transformers.BatchEncoding.char_to_token` method.
+
+.. code-block:: python
+
+    def add_token_positions(encodings, answers):
+        start_positions = []
+        end_positions = []
+        for i in range(len(answers)):
+            start_positions.append(encodings.char_to_token(i, answers[i]['answer_start']))
+            end_positions.append(encodings.char_to_token(i, answers[i]['answer_end'] - 1))
+            # if None, the answer passage has been truncated
+            if start_positions[-1] is None:
+                start_positions[-1] = tokenizer.model_max_length
+            if end_positions[-1] is None:
+                end_positions[-1] = tokenizer.model_max_length
+        encodings.update({'start_positions': start_positions, 'end_positions': end_positions})
+
+    add_token_positions(train_encodings, train_answers)
+    add_token_positions(val_encodings, val_answers)
+
+Our data is ready. Let's just put it in a PyTorch/TensorFlow dataset so that we can easily use it for training. In
+PyTorch, we define a custom ``Dataset`` class. In TensorFlow, we pass a tuple of ``(inputs_dict, labels_dict)`` to the
+``from_tensor_slices`` method.
+
+.. code-block:: python
+
+    ## PYTORCH CODE
+    import torch
+
+    class SquadDataset(torch.utils.data.Dataset):
+        def __init__(self, encodings):
+            self.encodings = encodings
+
+        def __getitem__(self, idx):
+            return {key: torch.tensor(val[idx]) for key, val in self.encodings.items()}
+
+        def __len__(self):
+            return len(self.encodings.input_ids)
+
+    train_dataset = SquadDataset(train_encodings)
+    val_dataset = SquadDataset(val_encodings)
+    ## TENSORFLOW CODE
+    import tensorflow as tf
+
+    train_dataset = tf.data.Dataset.from_tensor_slices((
+        {key: train_encodings[key] for key in ['input_ids', 'attention_mask']},
+        {key: train_encodings[key] for key in ['start_positions', 'end_positions']}
+    ))
+    val_dataset = tf.data.Dataset.from_tensor_slices((
+        {key: val_encodings[key] for key in ['input_ids', 'attention_mask']},
+        {key: val_encodings[key] for key in ['start_positions', 'end_positions']}
+    ))
+
+Now we can use a DistilBert model with a QA head for training:
+
+.. code-block:: python
+
+    ## PYTORCH CODE
+    from transformers import DistilBertForQuestionAnswering
+    model = DistilBertForQuestionAnswering.from_pretrained("distilbert-base-uncased")
+    ## TENSORFLOW CODE
+    from transformers import TFDistilBertForQuestionAnswering
+    model = TFDistilBertForQuestionAnswering.from_pretrained("distilbert-base-uncased")
+
+
+The data and model are both ready to go. You can train the model with
+:class:`~transformers.Trainer`/:class:`~transformers.TFTrainer` exactly as in the sequence classification example
+above. If using native PyTorch, replace ``labels`` with ``start_positions`` and ``end_positions`` in the training
+example. If using Keras's ``fit``, we need to make a minor modification to handle this example since it involves
+multiple model outputs.
+
+  - :ref:`ft_trainer`
+
+.. code-block:: python
+
+    ## PYTORCH CODE
+    from torch.utils.data import DataLoader
+    from transformers import AdamW
+
+    device = torch.device('cuda') if torch.cuda.is_available() else torch.device('cpu')
+
+    model.to(device)
+    model.train()
+
+    train_loader = DataLoader(train_dataset, batch_size=16, shuffle=True)
+
+    optim = AdamW(model.parameters(), lr=5e-5)
+
+    for epoch in range(3):
+        for batch in train_loader:
+            optim.zero_grad()
+            input_ids = batch['input_ids'].to(device)
+            attention_mask = batch['attention_mask'].to(device)
+            start_positions = batch['start_positions'].to(device)
+            end_positions = batch['end_positions'].to(device)
+            outputs = model(input_ids, attention_mask=attention_mask, start_positions=start_positions, end_positions=end_positions)
+            loss = outputs[0]
+            loss.backward()
+            optim.step()
+
+    model.eval()
+    ## TENSORFLOW CODE
+    # Keras will expect a tuple when dealing with labels
+    train_dataset = train_dataset.map(lambda x, y: (x, (y['start_positions'], y['end_positions'])))
+
+    # Keras will assign a separate loss for each output and add them together. So we'll just use the standard CE loss
+    # instead of using the built-in model.compute_loss, which expects a dict of outputs and averages the two terms.
+    # Note that this means the loss will be 2x of when using TFTrainer since we're adding instead of averaging them.
+    loss = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)
+    model.distilbert.return_dict = False # if using 🤗 Transformers >3.02, make sure outputs are tuples
+
+    optimizer = tf.keras.optimizers.Adam(learning_rate=5e-5)
+    model.compile(optimizer=optimizer, loss=loss) # can also use any keras loss fn
+    model.fit(train_dataset.shuffle(1000).batch(16), epochs=3, batch_size=16)
+
+.. _resources:
+
+Additional Resources
+-----------------------------------------------------------------------------------------------------------------------
+
+  - `How to train a new language model from scratch using Transformers and Tokenizers
+    <https://huggingface.co/blog/how-to-train>`_. Blog post showing the steps to load in Esperanto data and train a
+    masked language model from scratch.
+  - :doc:`Preprocessing <preprocessing>`. Docs page on data preprocessing.
+  - :doc:`Training <training>`. Docs page on training and fine-tuning.
+
+.. _nlplib:
+
+Using the 🤗 NLP Datasets & Metrics library
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+This tutorial demonstrates how to read in datasets from various raw text formats and prepare them for training with 🤗
+Transformers so that you can do the same thing with your own custom datasets. However, we recommend users use the `🤗
+NLP library <https://github.com/huggingface/nlp>`_ for working with the 150+ datasets included in the `hub
+<https://huggingface.co/datasets>`_, including the three datasets used in this tutorial. As a very brief overview, we
+will show how to use the NLP library to download and prepare the IMDb dataset from the first example, :ref:`seq_imdb`.
+
+Start by downloading the dataset:
+
+.. code-block:: python
+
+    from nlp import load_dataset
+    train = load_dataset("imdb", split="train")
+
+Each dataset has multiple columns corresponding to different features. Let's see what our columns are.
+
+.. code-block:: python
+
+    >>> print(train.column_names)
+    ['label', 'text']
+
+Great. Now let's tokenize the text. We can do this using the ``map`` method. We'll also rename the ``label`` column to
+``labels`` to match the model's input arguments.
+
+.. code-block:: python
+
+    train = train.map(lambda batch: tokenizer(batch["text"], truncation=True, padding=True), batched=True)
+    train.rename_column_("label", "labels")
+
+Lastly, we can use the ``set_format`` method to determine which columns and in what data format we want to access
+dataset elements.
+
+.. code-block:: python
+
+    ## PYTORCH CODE
+    >>> train.set_format("torch", columns=["input_ids", "attention_mask", "labels"])
+    >>> {key: val.shape for key, val in train[0].items()})
+    {'labels': torch.Size([]), 'input_ids': torch.Size([512]), 'attention_mask': torch.Size([512])}
+    ## TENSORFLOW CODE
+    >>> train.set_format("tensorflow", columns=["input_ids", "attention_mask", "labels"])
+    >>> {key: val.shape for key, val in train[0].items()})
+    {'labels': TensorShape([]), 'input_ids': TensorShape([512]), 'attention_mask': TensorShape([512])}
+
+We now have a fully-prepared dataset. Check out `the 🤗 NLP docs <https://huggingface.co/nlp/processing.html>`_ for a
+more thorough introduction.

docs/source/en/_config.py

@@ -1,14 +0,0 @@
-# docstyle-ignore
-INSTALL_CONTENT = """
-# Transformers installation
-! pip install transformers datasets
-# To install from source instead of the last release, comment the command above and uncomment the following one.
-# ! pip install git+https://github.com/huggingface/transformers.git
-"""
-
-notebook_first_cells = [{"type": "code", "content": INSTALL_CONTENT}]
-black_avoid_patterns = {
-    "{processor_class}": "FakeProcessorClass",
-    "{model_class}": "FakeModelClass",
-    "{object_class}": "FakeObjectClass",    
-}

docs/source/en/_toctree.yml

@@ -1,479 +0,0 @@
-- sections:
-  - local: index
-    title: 🤗 Transformers
-  - local: quicktour
-    title: Quick tour
-  - local: installation
-    title: Installation
-  title: Get started
-- sections:
-  - local: pipeline_tutorial
-    title: Pipelines for inference
-  - local: autoclass_tutorial
-    title: Load pretrained instances with an AutoClass
-  - local: preprocessing
-    title: Preprocess
-  - local: training
-    title: Fine-tune a pretrained model
-  - local: accelerate
-    title: Distributed training with 🤗 Accelerate
-  - local: model_sharing
-    title: Share a model
-  title: Tutorials
-- sections:
-  - local: fast_tokenizers
-    title: Use tokenizers from 🤗 Tokenizers
-  - local: create_a_model
-    title: Create a custom architecture
-  - local: custom_models
-    title: Sharing custom models
-  - sections:
-    - local: tasks/sequence_classification
-      title: Text classification
-    - local: tasks/token_classification
-      title: Token classification
-    - local: tasks/question_answering
-      title: Question answering
-    - local: tasks/language_modeling
-      title: Language modeling
-    - local: tasks/translation
-      title: Translation
-    - local: tasks/summarization
-      title: Summarization
-    - local: tasks/multiple_choice
-      title: Multiple choice
-    - local: tasks/audio_classification
-      title: Audio classification
-    - local: tasks/asr
-      title: Automatic speech recognition
-    - local: tasks/image_classification
-      title: Image classification
-    title: Fine-tune for downstream tasks
-  - local: run_scripts
-    title: Train with a script
-  - local: sagemaker
-    title: Run training on Amazon SageMaker
-  - local: multilingual
-    title: Inference for multilingual models
-  - local: converting_tensorflow_models
-    title: Converting TensorFlow Checkpoints
-  - local: serialization
-    title: Export 🤗 Transformers models
-  - sections:
-    - local: performance
-      title: Overview
-    - local: perf_train_gpu_one
-      title: Training on one GPU
-    - local: perf_train_gpu_many
-      title: Training on many GPUs
-    - local: perf_train_cpu
-      title: Training on CPU
-    - local: perf_train_cpu_many
-      title: Training on many CPUs
-    - local: perf_train_tpu
-      title: Training on TPUs
-    - local: perf_train_special
-      title: Training on Specialized Hardware
-    - local: perf_infer_cpu
-      title: Inference on CPU
-    - local: perf_infer_gpu_one
-      title: Inference on one GPU
-    - local: perf_infer_gpu_many
-      title: Inference on many GPUs
-    - local: perf_infer_special
-      title: Inference on Specialized Hardware
-    - local: perf_hardware
-      title: Custom hardware for training
-    title: Performance and scalability
-  - local: big_models
-    title: Instantiating a big model
-  - local: benchmarks
-    title: Benchmarks
-  - local: migration
-    title: Migrating from previous packages
-  - local: troubleshooting
-    title: Troubleshoot
-  - local: debugging
-    title: Debugging
-  - local: notebooks
-    title: 🤗 Transformers Notebooks
-  - local: community
-    title: Community
-  - local: contributing
-    title: How to contribute to transformers?
-  - local: add_new_model
-    title: How to add a model to 🤗 Transformers?
-  - local: add_new_pipeline
-    title: How to create a custom pipeline?
-  - local: testing
-    title: Testing
-  - local: pr_checks
-    title: Checks on a Pull Request
-  title: How-to guides
-- sections:
-  - local: philosophy
-    title: Philosophy
-  - local: glossary
-    title: Glossary
-  - local: task_summary
-    title: Summary of the tasks
-  - local: model_summary
-    title: Summary of the models
-  - local: tokenizer_summary
-    title: Summary of the tokenizers
-  - local: pad_truncation
-    title: Padding and truncation
-  - local: bertology
-    title: BERTology
-  - local: perplexity
-    title: Perplexity of fixed-length models
-  title: Conceptual guides
-- sections:
-  - sections:
-    - local: main_classes/callback
-      title: Callbacks
-    - local: main_classes/configuration
-      title: Configuration
-    - local: main_classes/data_collator
-      title: Data Collator
-    - local: main_classes/keras_callbacks
-      title: Keras callbacks
-    - local: main_classes/logging
-      title: Logging
-    - local: main_classes/model
-      title: Models
-    - local: main_classes/text_generation
-      title: Text Generation
-    - local: main_classes/onnx
-      title: ONNX
-    - local: main_classes/optimizer_schedules
-      title: Optimization
-    - local: main_classes/output
-      title: Model outputs
-    - local: main_classes/pipelines
-      title: Pipelines
-    - local: main_classes/processors
-      title: Processors
-    - local: main_classes/tokenizer
-      title: Tokenizer
-    - local: main_classes/trainer
-      title: Trainer
-    - local: main_classes/deepspeed
-      title: DeepSpeed Integration
-    - local: main_classes/feature_extractor
-      title: Feature Extractor
-    title: Main Classes
-  - sections:
-    - local: model_doc/auto
-      title: Auto Classes
-    - isExpanded: false
-      sections:
-      - local: model_doc/albert
-        title: ALBERT
-      - local: model_doc/bart
-        title: BART
-      - local: model_doc/barthez
-        title: BARThez
-      - local: model_doc/bartpho
-        title: BARTpho
-      - local: model_doc/bert
-        title: BERT
-      - local: model_doc/bert-generation
-        title: BertGeneration
-      - local: model_doc/bert-japanese
-        title: BertJapanese
-      - local: model_doc/bertweet
-        title: Bertweet
-      - local: model_doc/big_bird
-        title: BigBird
-      - local: model_doc/bigbird_pegasus
-        title: BigBirdPegasus
-      - local: model_doc/blenderbot
-        title: Blenderbot
-      - local: model_doc/blenderbot-small
-        title: Blenderbot Small
-      - local: model_doc/bloom
-        title: BLOOM
-      - local: model_doc/bort
-        title: BORT
-      - local: model_doc/byt5
-        title: ByT5
-      - local: model_doc/camembert
-        title: CamemBERT
-      - local: model_doc/canine
-        title: CANINE
-      - local: model_doc/codegen
-        title: CodeGen
-      - local: model_doc/convbert
-        title: ConvBERT
-      - local: model_doc/cpm
-        title: CPM
-      - local: model_doc/ctrl
-        title: CTRL
-      - local: model_doc/deberta
-        title: DeBERTa
-      - local: model_doc/deberta-v2
-        title: DeBERTa-v2
-      - local: model_doc/dialogpt
-        title: DialoGPT
-      - local: model_doc/distilbert
-        title: DistilBERT
-      - local: model_doc/dpr
-        title: DPR
-      - local: model_doc/electra
-        title: ELECTRA
-      - local: model_doc/encoder-decoder
-        title: Encoder Decoder Models
-      - local: model_doc/flaubert
-        title: FlauBERT
-      - local: model_doc/fnet
-        title: FNet
-      - local: model_doc/fsmt
-        title: FSMT
-      - local: model_doc/funnel
-        title: Funnel Transformer
-      - local: model_doc/openai-gpt
-        title: GPT
-      - local: model_doc/gpt_neo
-        title: GPT Neo
-      - local: model_doc/gpt_neox
-        title: GPT NeoX
-      - local: model_doc/gptj
-        title: GPT-J
-      - local: model_doc/gpt2
-        title: GPT2
-      - local: model_doc/herbert
-        title: HerBERT
-      - local: model_doc/ibert
-        title: I-BERT
-      - local: model_doc/layoutlm
-        title: LayoutLM
-      - local: model_doc/led
-        title: LED
-      - local: model_doc/longformer
-        title: Longformer
-      - local: model_doc/longt5
-        title: LongT5
-      - local: model_doc/luke
-        title: LUKE
-      - local: model_doc/m2m_100
-        title: M2M100
-      - local: model_doc/marian
-        title: MarianMT
-      - local: model_doc/mbart
-        title: MBart and MBart-50
-      - local: model_doc/megatron-bert
-        title: MegatronBERT
-      - local: model_doc/megatron_gpt2
-        title: MegatronGPT2
-      - local: model_doc/mluke
-        title: mLUKE
-      - local: model_doc/mobilebert
-        title: MobileBERT
-      - local: model_doc/mpnet
-        title: MPNet
-      - local: model_doc/mt5
-        title: MT5
-      - local: model_doc/mvp
-        title: MVP
-      - local: model_doc/nezha
-        title: NEZHA
-      - local: model_doc/nllb
-        title: NLLB
-      - local: model_doc/nystromformer
-        title: Nyströmformer
-      - local: model_doc/opt
-        title: OPT
-      - local: model_doc/pegasus
-        title: Pegasus
-      - local: model_doc/phobert
-        title: PhoBERT
-      - local: model_doc/plbart
-        title: PLBart
-      - local: model_doc/prophetnet
-        title: ProphetNet
-      - local: model_doc/qdqbert
-        title: QDQBert
-      - local: model_doc/rag
-        title: RAG
-      - local: model_doc/realm
-        title: REALM
-      - local: model_doc/reformer
-        title: Reformer
-      - local: model_doc/rembert
-        title: RemBERT
-      - local: model_doc/retribert
-        title: RetriBERT
-      - local: model_doc/roberta
-        title: RoBERTa
-      - local: model_doc/roformer
-        title: RoFormer
-      - local: model_doc/splinter
-        title: Splinter
-      - local: model_doc/squeezebert
-        title: SqueezeBERT
-      - local: model_doc/t5
-        title: T5
-      - local: model_doc/t5v1.1
-        title: T5v1.1
-      - local: model_doc/tapas
-        title: TAPAS
-      - local: model_doc/tapex
-        title: TAPEX
-      - local: model_doc/transfo-xl
-        title: Transformer XL
-      - local: model_doc/ul2
-        title: UL2
-      - local: model_doc/xglm
-        title: XGLM
-      - local: model_doc/xlm
-        title: XLM
-      - local: model_doc/xlm-prophetnet
-        title: XLM-ProphetNet
-      - local: model_doc/xlm-roberta
-        title: XLM-RoBERTa
-      - local: model_doc/xlm-roberta-xl
-        title: XLM-RoBERTa-XL
-      - local: model_doc/xlnet
-        title: XLNet
-      - local: model_doc/yoso
-        title: YOSO
-      title: Text models
-    - isExpanded: false
-      sections:
-      - local: model_doc/beit
-        title: BEiT
-      - local: model_doc/convnext
-        title: ConvNeXT
-      - local: model_doc/cvt
-        title: CvT
-      - local: model_doc/deit
-        title: DeiT
-      - local: model_doc/detr
-        title: DETR
-      - local: model_doc/dit
-        title: DiT
-      - local: model_doc/dpt
-        title: DPT
-      - local: model_doc/glpn
-        title: GLPN
-      - local: model_doc/imagegpt
-        title: ImageGPT
-      - local: model_doc/levit
-        title: LeViT
-      - local: model_doc/maskformer
-        title: MaskFormer
-      - local: model_doc/mobilevit
-        title: MobileViT
-      - local: model_doc/owlvit
-        title: OWL-ViT
-      - local: model_doc/poolformer
-        title: PoolFormer
-      - local: model_doc/regnet
-        title: RegNet
-      - local: model_doc/resnet
-        title: ResNet
-      - local: model_doc/segformer
-        title: SegFormer
-      - local: model_doc/swin
-        title: Swin Transformer
-      - local: model_doc/swinv2
-        title: Swin Transformer V2
-      - local: model_doc/van
-        title: VAN
-      - local: model_doc/videomae
-        title: VideoMAE
-      - local: model_doc/vit
-        title: Vision Transformer (ViT)
-      - local: model_doc/vit_mae
-        title: ViTMAE
-      - local: model_doc/yolos
-        title: YOLOS
-      title: Vision models
-    - isExpanded: false
-      sections:
-      - local: model_doc/hubert
-        title: Hubert
-      - local: model_doc/mctct
-        title: MCTCT
-      - local: model_doc/sew
-        title: SEW
-      - local: model_doc/sew-d
-        title: SEW-D
-      - local: model_doc/speech_to_text
-        title: Speech2Text
-      - local: model_doc/speech_to_text_2
-        title: Speech2Text2
-      - local: model_doc/unispeech
-        title: UniSpeech
-      - local: model_doc/unispeech-sat
-        title: UniSpeech-SAT
-      - local: model_doc/wav2vec2
-        title: Wav2Vec2
-      - local: model_doc/wav2vec2-conformer
-        title: Wav2Vec2-Conformer
-      - local: model_doc/wav2vec2_phoneme
-        title: Wav2Vec2Phoneme
-      - local: model_doc/wavlm
-        title: WavLM
-      - local: model_doc/xls_r
-        title: XLS-R
-      - local: model_doc/xlsr_wav2vec2
-        title: XLSR-Wav2Vec2
-      title: Audio models
-    - isExpanded: false
-      sections:
-      - local: model_doc/clip
-        title: CLIP
-      - local: model_doc/data2vec
-        title: Data2Vec
-      - local: model_doc/flava
-        title: FLAVA
-      - local: model_doc/groupvit
-        title: GroupViT
-      - local: model_doc/layoutlmv2
-        title: LayoutLMV2
-      - local: model_doc/layoutlmv3
-        title: LayoutLMV3
-      - local: model_doc/layoutxlm
-        title: LayoutXLM
-      - local: model_doc/lxmert
-        title: LXMERT
-      - local: model_doc/perceiver
-        title: Perceiver
-      - local: model_doc/speech-encoder-decoder
-        title: Speech Encoder Decoder Models
-      - local: model_doc/trocr
-        title: TrOCR
-      - local: model_doc/vilt
-        title: ViLT
-      - local: model_doc/vision-encoder-decoder
-        title: Vision Encoder Decoder Models
-      - local: model_doc/vision-text-dual-encoder
-        title: Vision Text Dual Encoder
-      - local: model_doc/visual_bert
-        title: VisualBERT
-      title: Multimodal models
-    - isExpanded: false
-      sections:
-      - local: model_doc/decision_transformer
-        title: Decision Transformer
-      - local: model_doc/trajectory_transformer
-        title: Trajectory Transformer
-      title: Reinforcement learning models
-    title: Models
-  - sections:
-    - local: internal/modeling_utils
-      title: Custom Layers and Utilities
-    - local: internal/pipelines_utils
-      title: Utilities for pipelines
-    - local: internal/tokenization_utils
-      title: Utilities for Tokenizers
-    - local: internal/trainer_utils
-      title: Utilities for Trainer
-    - local: internal/generation_utils
-      title: Utilities for Generation
-    - local: internal/file_utils
-      title: General Utilities
-    title: Internal Helpers
-  title: API

docs/source/en/accelerate.mdx

@@ -1,132 +0,0 @@
-<!--Copyright 2022 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Distributed training with 🤗 Accelerate
-
-As models get bigger, parallelism has emerged as a strategy for training larger models on limited hardware and accelerating training speed by several orders of magnitude. At Hugging Face, we created the [🤗 Accelerate](https://huggingface.co/docs/accelerate/index.html) library to help users easily train a 🤗 Transformers model on any type of distributed setup, whether it is multiple GPU's on one machine or multiple GPU's across several machines. In this tutorial, learn how to customize your native PyTorch training loop to enable training in a distributed environment.
-
-## Setup
-
-Get started by installing 🤗 Accelerate:
-
-```bash
-pip install accelerate
-```
-
-Then import and create an [`Accelerator`](https://huggingface.co/docs/accelerate/accelerator.html#accelerate.Accelerator) object. `Accelerator` will automatically detect your type of distributed setup and initialize all the necessary components for training. You don't need to explicitly place your model on a device.
-
-```py
->>> from accelerate import Accelerator
-
->>> accelerator = Accelerator()
-```
-
-## Prepare to accelerate
-
-The next step is to pass all the relevant training objects to the [`prepare`](https://huggingface.co/docs/accelerate/accelerator.html#accelerate.Accelerator.prepare) method. This includes your training and evaluation DataLoaders, a model and an optimizer:
-
-```py
->>> train_dataloader, eval_dataloader, model, optimizer = accelerator.prepare(
-...     train_dataloader, eval_dataloader, model, optimizer
-... )
-```
-
-## Backward
-
-The last addition is to replace the typical `loss.backward()` in your training loop with 🤗 Accelerate's [`backward`](https://huggingface.co/docs/accelerate/accelerator.html#accelerate.Accelerator.backward) method:
-
-```py
->>> for epoch in range(num_epochs):
-...     for batch in train_dataloader:
-...         outputs = model(**batch)
-...         loss = outputs.loss
-...         accelerator.backward(loss)
-
-...         optimizer.step()
-...         lr_scheduler.step()
-...         optimizer.zero_grad()
-...         progress_bar.update(1)
-```
-
-As you can see in the following code, you only need to add four additional lines of code to your training loop to enable distributed training!
-
-```diff
-+ from accelerate import Accelerator
-  from transformers import AdamW, AutoModelForSequenceClassification, get_scheduler
-
-+ accelerator = Accelerator()
-
-  model = AutoModelForSequenceClassification.from_pretrained(checkpoint, num_labels=2)
-  optimizer = AdamW(model.parameters(), lr=3e-5)
-
-- device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
-- model.to(device)
-
-+ train_dataloader, eval_dataloader, model, optimizer = accelerator.prepare(
-+     train_dataloader, eval_dataloader, model, optimizer
-+ )
-
-  num_epochs = 3
-  num_training_steps = num_epochs * len(train_dataloader)
-  lr_scheduler = get_scheduler(
-      "linear",
-      optimizer=optimizer,
-      num_warmup_steps=0,
-      num_training_steps=num_training_steps
-  )
-
-  progress_bar = tqdm(range(num_training_steps))
-
-  model.train()
-  for epoch in range(num_epochs):
-      for batch in train_dataloader:
--         batch = {k: v.to(device) for k, v in batch.items()}
-          outputs = model(**batch)
-          loss = outputs.loss
--         loss.backward()
-+         accelerator.backward(loss)
-
-          optimizer.step()
-          lr_scheduler.step()
-          optimizer.zero_grad()
-          progress_bar.update(1)
-```
-
-## Train
-
-Once you've added the relevant lines of code, launch your training in a script or a notebook like Colaboratory.
-
-### Train with a script
-
-If you are running your training from a script, run the following command to create and save a configuration file:
-
-```bash
-accelerate config
-```
-
-Then launch your training with:
-
-```bash
-accelerate launch train.py
-```
-
-### Train with a notebook
-
-🤗 Accelerate can also run in a notebook if you're planning on using Colaboratory's TPUs. Wrap all the code responsible for training in a function, and pass it to `notebook_launcher`:
-
-```py
->>> from accelerate import notebook_launcher
-
->>> notebook_launcher(training_function)
-```
-
-For more information about 🤗 Accelerate and it's rich features, refer to the [documentation](https://huggingface.co/docs/accelerate/index.html).

docs/source/en/add_new_model.mdx

@@ -1,845 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
--->
-
-# How to add a model to 🤗 Transformers?
-
-Adding a new model is often difficult and requires an in-depth knowledge of the 🤗 Transformers library and ideally also
-of the model's original repository. At Hugging Face, we are trying to empower the community more and more to add models
-independently. Thus, for some new models that the community wants to be added to 🤗 Transformers, we create a customized
-*call-for-model-addition* that explains step-by-step how to add the requested model. With this
-*call-for-model-addition*, we want to teach a motivated and experienced contributor of the community how to port a
-model to 🤗 Transformers.
-
-If this sounds like something you would be interested in, feel free to check out the currently open
-“calls-for-model-addition” [here](https://github.com/huggingface/transformers/tree/main/templates/adding_a_new_model/open_model_proposals/README.md)
-and to contact us.
-
-If selected, you will then work closely with one member of the Hugging Face team to integrate the model into 🤗
-Transformers. By doing so, you will both gain a theoretical and deep practical understanding of the proposed model. But
-more importantly, you will have made a major open-source contribution to 🤗 Transformers. Along the way, you will:
-
--  get insights into open-source best practices
--  understand the design principles of one of the most popular NLP libraries
--  learn how to do efficiently test large NLP models
--  learn how to integrate Python utilities like `black`, `isort`, `make fix-copies` into a library to always
-  ensure clean and readable code
-
-We are also more than happy if you want to add a model that cannot be found in the “calls-for-model-addition” folder.
-The following sections explain in detail how to add a new model. It might also be very helpful to check out already
-added models to see if those resemble the model you would like to add [here](https://github.com/huggingface/transformers/pulls?q=is%3Apr+label%3A%22PR+for+Model+Addition%22+is%3Aclosed).
-
-To start, let's try to get a general overview of the Transformers library.
-
-## General overview of 🤗 Transformers
-
-First, you should get a general overview of 🤗 Transformers. 🤗 Transformers is a very opinionated library, so there is a
-chance that you don't agree with some of the library's philosophies or design choices. From our experience, however, we
-found that the fundamental design choices and philosophies of the library are crucial to efficiently scale 🤗
-Transformers while keeping maintenance costs at a reasonable level.
-
-A good first starting point to better understand the library is to read the [documentation of our philosophy](philosophy). As a result of our way of working, there are some choices that we try to apply to all models:
-
-- Composition is generally favored over-abstraction
-- Duplicating code is not always bad if it strongly improves the readability or accessibility of a model
-- Model files are as self-contained as possible so that when you read the code of a specific model, you ideally only
-  have to look into the respective `modeling_....py` file.
-
-In our opinion, the library's code is not just a means to provide a product, *e.g.* the ability to use BERT for
-inference, but also as the very product that we want to improve. Hence, when adding a model, the user is not only the
-person that will use your model, but also everybody that will read, try to understand, and possibly tweak your code.
-
-With this in mind, let's go a bit deeper into the general library design.
-
-### Overview of models
-
-To successfully add a model, it is important to understand the interaction between your model and its config,
-[`PreTrainedModel`], and [`PretrainedConfig`]. For exemplary purposes, we will
-call the model to be added to 🤗 Transformers `BrandNewBert`.
-
-Let's take a look:
-
-<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers_overview.png"/>
-
-As you can see, we do make use of inheritance in 🤗 Transformers, but we keep the level of abstraction to an absolute
-minimum. There are never more than two levels of abstraction for any model in the library. `BrandNewBertModel`
-inherits from `BrandNewBertPreTrainedModel` which in turn inherits from [`PreTrainedModel`] and
-that's it. As a general rule, we want to make sure that a new model only depends on
-[`PreTrainedModel`]. The important functionalities that are automatically provided to every new
-model are [`~PreTrainedModel.from_pretrained`] and
-[`~PreTrainedModel.save_pretrained`], which are used for serialization and deserialization. All of the
-other important functionalities, such as `BrandNewBertModel.forward` should be completely defined in the new
-`modeling_brand_new_bert.py` script. Next, we want to make sure that a model with a specific head layer, such as
-`BrandNewBertForMaskedLM` does not inherit from `BrandNewBertModel`, but rather uses `BrandNewBertModel`
-as a component that can be called in its forward pass to keep the level of abstraction low. Every new model requires a
-configuration class, called `BrandNewBertConfig`. This configuration is always stored as an attribute in
-[`PreTrainedModel`], and thus can be accessed via the `config` attribute for all classes
-inheriting from `BrandNewBertPreTrainedModel`:
-
-```python
-model = BrandNewBertModel.from_pretrained("brandy/brand_new_bert")
-model.config  # model has access to its config
-```
-
-Similar to the model, the configuration inherits basic serialization and deserialization functionalities from
-[`PretrainedConfig`]. Note that the configuration and the model are always serialized into two
-different formats - the model to a *pytorch_model.bin* file and the configuration to a *config.json* file. Calling
-[`~PreTrainedModel.save_pretrained`] will automatically call
-[`~PretrainedConfig.save_pretrained`], so that both model and configuration are saved.
-
-
-### Code style
-
-When coding your new model, keep in mind that Transformers is an opinionated library and we have a few quirks of our
-own regarding how code should be written :-)
-
-1. The forward pass of your model should be fully written in the modeling file while being fully independent of other
-   models in the library. If you want to reuse a block from another model, copy the code and paste it with a
-   `# Copied from` comment on top (see [here](https://github.com/huggingface/transformers/blob/v4.17.0/src/transformers/models/roberta/modeling_roberta.py#L160)
-   for a good example).
-2. The code should be fully understandable, even by a non-native English speaker. This means you should pick
-   descriptive variable names and avoid abbreviations. As an example, `activation` is preferred to `act`.
-   One-letter variable names are strongly discouraged unless it's an index in a for loop. 
-3. More generally we prefer longer explicit code to short magical one.
-4. Avoid subclassing `nn.Sequential` in PyTorch but subclass `nn.Module` and write the forward pass, so that anyone
-   using your code can quickly debug it by adding print statements or breaking points.
-5. Your function signature should be type-annotated. For the rest, good variable names are way more readable and
-   understandable than type annotations.
-
-### Overview of tokenizers
-
-Not quite ready yet :-( This section will be added soon!
-
-## Step-by-step recipe to add a model to 🤗 Transformers
-
-Everyone has different preferences of how to port a model so it can be very helpful for you to take a look at summaries
-of how other contributors ported models to Hugging Face. Here is a list of community blog posts on how to port a model:
-
-1. [Porting GPT2 Model](https://medium.com/huggingface/from-tensorflow-to-pytorch-265f40ef2a28) by [Thomas](https://huggingface.co/thomwolf)
-2. [Porting WMT19 MT Model](https://huggingface.co/blog/porting-fsmt) by [Stas](https://huggingface.co/stas)
-
-From experience, we can tell you that the most important things to keep in mind when adding a model are:
-
--  Don't reinvent the wheel! Most parts of the code you will add for the new 🤗 Transformers model already exist
-  somewhere in 🤗 Transformers. Take some time to find similar, already existing models and tokenizers you can copy
-  from. [grep](https://www.gnu.org/software/grep/) and [rg](https://github.com/BurntSushi/ripgrep) are your
-  friends. Note that it might very well happen that your model's tokenizer is based on one model implementation, and
-  your model's modeling code on another one. *E.g.* FSMT's modeling code is based on BART, while FSMT's tokenizer code
-  is based on XLM.
--  It's more of an engineering challenge than a scientific challenge. You should spend more time on creating an
-  efficient debugging environment than trying to understand all theoretical aspects of the model in the paper.
--  Ask for help, when you're stuck! Models are the core component of 🤗 Transformers so that we at Hugging Face are more
-  than happy to help you at every step to add your model. Don't hesitate to ask if you notice you are not making
-  progress.
-
-In the following, we try to give you a general recipe that we found most useful when porting a model to 🤗 Transformers.
-
-The following list is a summary of everything that has to be done to add a model and can be used by you as a To-Do
-List:
-
--  1. ☐ (Optional) Understood theoretical aspects
--  2. ☐ Prepared transformers dev environment
--  3. ☐ Set up debugging environment of the original repository
--  4. ☐ Created script that successfully runs forward pass using original repository and checkpoint
--  5. ☐ Successfully added the model skeleton to Transformers
--  6. ☐ Successfully converted original checkpoint to Transformers checkpoint
--  7. ☐ Successfully ran forward pass in Transformers that gives identical output to original checkpoint
--  8. ☐ Finished model tests in Transformers
--  9. ☐ Successfully added Tokenizer in Transformers
--  10. ☐ Run end-to-end integration tests
--  11. ☐ Finished docs
--  12. ☐ Uploaded model weights to the hub
--  13. ☐ Submitted the pull request
--  14. ☐ (Optional) Added a demo notebook
-
-To begin with, we usually recommend to start by getting a good theoretical understanding of `BrandNewBert`. However,
-if you prefer to understand the theoretical aspects of the model *on-the-job*, then it is totally fine to directly dive
-into the `BrandNewBert`'s code-base. This option might suit you better, if your engineering skills are better than
-your theoretical skill, if you have trouble understanding `BrandNewBert`'s paper, or if you just enjoy programming
-much more than reading scientific papers.
-
-### 1. (Optional) Theoretical aspects of BrandNewBert
-
-You should take some time to read *BrandNewBert's* paper, if such descriptive work exists. There might be large
-sections of the paper that are difficult to understand. If this is the case, this is fine - don't worry! The goal is
-not to get a deep theoretical understanding of the paper, but to extract the necessary information required to
-effectively re-implement the model in 🤗 Transformers. That being said, you don't have to spend too much time on the
-theoretical aspects, but rather focus on the practical ones, namely:
-
--  What type of model is *brand_new_bert*? BERT-like encoder-only model? GPT2-like decoder-only model? BART-like
-  encoder-decoder model? Look at the [model_summary](model_summary) if you're not familiar with the differences between those.
--  What are the applications of *brand_new_bert*? Text classification? Text generation? Seq2Seq tasks, *e.g.,*
-  summarization?
--  What is the novel feature of the model making it different from BERT/GPT-2/BART?
--  Which of the already existing [🤗 Transformers models](https://huggingface.co/transformers/#contents) is most
-  similar to *brand_new_bert*?
--  What type of tokenizer is used? A sentencepiece tokenizer? Word piece tokenizer? Is it the same tokenizer as used
-  for BERT or BART?
-
-After you feel like you have gotten a good overview of the architecture of the model, you might want to write to the
-Hugging Face team with any questions you might have. This might include questions regarding the model's architecture,
-its attention layer, etc. We will be more than happy to help you.
-
-### 2. Next prepare your environment
-
-1. Fork the [repository](https://github.com/huggingface/transformers) by clicking on the ‘Fork' button on the
-   repository's page. This creates a copy of the code under your GitHub user account.
-
-2. Clone your `transformers` fork to your local disk, and add the base repository as a remote:
-
-```bash
-git clone https://github.com/[your Github handle]/transformers.git
-cd transformers
-git remote add upstream https://github.com/huggingface/transformers.git
-```
-
-3. Set up a development environment, for instance by running the following command:
-
-```bash
-python -m venv .env
-source .env/bin/activate
-pip install -e ".[dev]"
-```
-
-and return to the parent directory
-
-```bash
-cd ..
-```
-
-4. We recommend adding the PyTorch version of *brand_new_bert* to Transformers. To install PyTorch, please follow the
-   instructions on https://pytorch.org/get-started/locally/.
-
-**Note:** You don't need to have CUDA installed. Making the new model work on CPU is sufficient.
-
-5. To port *brand_new_bert*, you will also need access to its original repository:
-
-```bash
-git clone https://github.com/org_that_created_brand_new_bert_org/brand_new_bert.git 
-cd brand_new_bert
-pip install -e .
-```
-
-Now you have set up a development environment to port *brand_new_bert* to 🤗 Transformers.
-
-### 3.-4. Run a pretrained checkpoint using the original repository
-
-At first, you will work on the original *brand_new_bert* repository. Often, the original implementation is very
-“researchy”. Meaning that documentation might be lacking and the code can be difficult to understand. But this should
-be exactly your motivation to reimplement *brand_new_bert*. At Hugging Face, one of our main goals is to *make people
-stand on the shoulders of giants* which translates here very well into taking a working model and rewriting it to make
-it as **accessible, user-friendly, and beautiful** as possible. This is the number-one motivation to re-implement
-models into 🤗 Transformers - trying to make complex new NLP technology accessible to **everybody**.
-
-You should start thereby by diving into the original repository.
-
-Successfully running the official pretrained model in the original repository is often **the most difficult** step.
-From our experience, it is very important to spend some time getting familiar with the original code-base. You need to
-figure out the following:
-
-- Where to find the pretrained weights?
-- How to load the pretrained weights into the corresponding model?
-- How to run the tokenizer independently from the model?
-- Trace one forward pass so that you know which classes and functions are required for a simple forward pass. Usually,
-  you only have to reimplement those functions.
-- Be able to locate the important components of the model: Where is the model's class? Are there model sub-classes,
-  *e.g.* EncoderModel, DecoderModel? Where is the self-attention layer? Are there multiple different attention layers,
-  *e.g.* *self-attention*, *cross-attention*...?
-- How can you debug the model in the original environment of the repo? Do you have to add *print* statements, can you
-  work with an interactive debugger like *ipdb*, or should you use an efficient IDE to debug the model, like PyCharm?
-
-It is very important that before you start the porting process, that you can **efficiently** debug code in the original
-repository! Also, remember that you are working with an open-source library, so do not hesitate to open an issue, or
-even a pull request in the original repository. The maintainers of this repository are most likely very happy about
-someone looking into their code!
-
-At this point, it is really up to you which debugging environment and strategy you prefer to use to debug the original
-model. We strongly advise against setting up a costly GPU environment, but simply work on a CPU both when starting to
-dive into the original repository and also when starting to write the 🤗 Transformers implementation of the model. Only
-at the very end, when the model has already been successfully ported to 🤗 Transformers, one should verify that the
-model also works as expected on GPU.
-
-In general, there are two possible debugging environments for running the original model
-
--  [Jupyter notebooks](https://jupyter.org/) / [google colab](https://colab.research.google.com/notebooks/intro.ipynb)
--  Local python scripts.
-
-Jupyter notebooks have the advantage that they allow for cell-by-cell execution which can be helpful to better split
-logical components from one another and to have faster debugging cycles as intermediate results can be stored. Also,
-notebooks are often easier to share with other contributors, which might be very helpful if you want to ask the Hugging
-Face team for help. If you are familiar with Jupiter notebooks, we strongly recommend you to work with them.
-
-The obvious disadvantage of Jupyter notebooks is that if you are not used to working with them you will have to spend
-some time adjusting to the new programming environment and that you might not be able to use your known debugging tools
-anymore, like `ipdb`.
-
-For each code-base, a good first step is always to load a **small** pretrained checkpoint and to be able to reproduce a
-single forward pass using a dummy integer vector of input IDs as an input. Such a script could look like this (in
-pseudocode):
-
-```python
-model = BrandNewBertModel.load_pretrained_checkpoint("/path/to/checkpoint/")
-input_ids = [0, 4, 5, 2, 3, 7, 9]  # vector of input ids
-original_output = model.predict(input_ids)
-```
-
-Next, regarding the debugging strategy, there are generally a few from which to choose from:
-
-- Decompose the original model into many small testable components and run a forward pass on each of those for
-  verification
-- Decompose the original model only into the original *tokenizer* and the original *model*, run a forward pass on
-  those, and use intermediate print statements or breakpoints for verification
-
-Again, it is up to you which strategy to choose. Often, one or the other is advantageous depending on the original code
-base.
-
-If the original code-base allows you to decompose the model into smaller sub-components, *e.g.* if the original
-code-base can easily be run in eager mode, it is usually worth the effort to do so. There are some important advantages
-to taking the more difficult road in the beginning:
-
-- at a later stage when comparing the original model to the Hugging Face implementation, you can verify automatically
-  for each component individually that the corresponding component of the 🤗 Transformers implementation matches instead
-  of relying on visual comparison via print statements
-- it can give you some rope to decompose the big problem of porting a model into smaller problems of just porting
-  individual components and thus structure your work better
-- separating the model into logical meaningful components will help you to get a better overview of the model's design
-  and thus to better understand the model
-- at a later stage those component-by-component tests help you to ensure that no regression occurs as you continue
-  changing your code
-
-[Lysandre's](https://gist.github.com/LysandreJik/db4c948f6b4483960de5cbac598ad4ed) integration checks for ELECTRA
-gives a nice example of how this can be done.
-
-However, if the original code-base is very complex or only allows intermediate components to be run in a compiled mode,
-it might be too time-consuming or even impossible to separate the model into smaller testable sub-components. A good
-example is [T5's MeshTensorFlow](https://github.com/tensorflow/mesh/tree/master/mesh_tensorflow) library which is
-very complex and does not offer a simple way to decompose the model into its sub-components. For such libraries, one
-often relies on verifying print statements.
-
-No matter which strategy you choose, the recommended procedure is often the same in that you should start to debug the
-starting layers first and the ending layers last.
-
-It is recommended that you retrieve the output, either by print statements or sub-component functions, of the following
-layers in the following order:
-
-1. Retrieve the input IDs passed to the model
-2. Retrieve the word embeddings
-3. Retrieve the input of the first Transformer layer
-4. Retrieve the output of the first Transformer layer
-5. Retrieve the output of the following n - 1 Transformer layers
-6. Retrieve the output of the whole BrandNewBert Model
-
-Input IDs should thereby consists of an array of integers, *e.g.* `input_ids = [0, 4, 4, 3, 2, 4, 1, 7, 19]`
-
-The outputs of the following layers often consist of multi-dimensional float arrays and can look like this:
-
-```
-[[
- [-0.1465, -0.6501,  0.1993,  ...,  0.1451,  0.3430,  0.6024],
- [-0.4417, -0.5920,  0.3450,  ..., -0.3062,  0.6182,  0.7132],
- [-0.5009, -0.7122,  0.4548,  ..., -0.3662,  0.6091,  0.7648],
- ...,
- [-0.5613, -0.6332,  0.4324,  ..., -0.3792,  0.7372,  0.9288],
- [-0.5416, -0.6345,  0.4180,  ..., -0.3564,  0.6992,  0.9191],
- [-0.5334, -0.6403,  0.4271,  ..., -0.3339,  0.6533,  0.8694]]],
-```
-
-We expect that every model added to 🤗 Transformers passes a couple of integration tests, meaning that the original
-model and the reimplemented version in 🤗 Transformers have to give the exact same output up to a precision of 0.001!
-Since it is normal that the exact same model written in different libraries can give a slightly different output
-depending on the library framework, we accept an error tolerance of 1e-3 (0.001). It is not enough if the model gives
-nearly the same output, they have to be the almost identical. Therefore, you will certainly compare the intermediate
-outputs of the 🤗 Transformers version multiple times against the intermediate outputs of the original implementation of
-*brand_new_bert* in which case an **efficient** debugging environment of the original repository is absolutely
-important. Here is some advice is to make your debugging environment as efficient as possible.
-
-- Find the best way of debugging intermediate results. Is the original repository written in PyTorch? Then you should
-  probably take the time to write a longer script that decomposes the original model into smaller sub-components to
-  retrieve intermediate values. Is the original repository written in Tensorflow 1? Then you might have to rely on
-  TensorFlow print operations like [tf.print](https://www.tensorflow.org/api_docs/python/tf/print) to output
-  intermediate values. Is the original repository written in Jax? Then make sure that the model is **not jitted** when
-  running the forward pass, *e.g.* check-out [this link](https://github.com/google/jax/issues/196).
-- Use the smallest pretrained checkpoint you can find. The smaller the checkpoint, the faster your debug cycle
-  becomes. It is not efficient if your pretrained model is so big that your forward pass takes more than 10 seconds.
-  In case only very large checkpoints are available, it might make more sense to create a dummy model in the new
-  environment with randomly initialized weights and save those weights for comparison with the 🤗 Transformers version
-  of your model
-- Make sure you are using the easiest way of calling a forward pass in the original repository. Ideally, you want to
-  find the function in the original repository that **only** calls a single forward pass, *i.e.* that is often called
-  `predict`, `evaluate`, `forward` or `__call__`. You don't want to debug a function that calls `forward`
-  multiple times, *e.g.* to generate text, like `autoregressive_sample`, `generate`.
-- Try to separate the tokenization from the model's *forward* pass. If the original repository shows examples where
-  you have to input a string, then try to find out where in the forward call the string input is changed to input ids
-  and start from this point. This might mean that you have to possibly write a small script yourself or change the
-  original code so that you can directly input the ids instead of an input string.
-- Make sure that the model in your debugging setup is **not** in training mode, which often causes the model to yield
-  random outputs due to multiple dropout layers in the model. Make sure that the forward pass in your debugging
-  environment is **deterministic** so that the dropout layers are not used. Or use *transformers.utils.set_seed*
-  if the old and new implementations are in the same framework.
-
-The following section gives you more specific details/tips on how you can do this for *brand_new_bert*.
-
-### 5.-14. Port BrandNewBert to 🤗 Transformers
-
-Next, you can finally start adding new code to 🤗 Transformers. Go into the clone of your 🤗 Transformers' fork:
-
-```bash
-cd transformers
-```
-
-In the special case that you are adding a model whose architecture exactly matches the model architecture of an
-existing model you only have to add a conversion script as described in [this section](#write-a-conversion-script).
-In this case, you can just re-use the whole model architecture of the already existing model.
-
-Otherwise, let's start generating a new model. You have two choices here:
-
-- `transformers-cli add-new-model-like` to add a new model like an existing one
-- `transformers-cli add-new-model` to add a new model from our template (will look like BERT or Bart depending on the type of model you select)
-
-In both cases, you will be prompted with a questionnaire to fill the basic information of your model. The second command requires to install `cookiecutter`, you can find more information on it [here](https://github.com/huggingface/transformers/tree/main/templates/adding_a_new_model).
-
-**Open a Pull Request on the main huggingface/transformers repo**
-
-Before starting to adapt the automatically generated code, now is the time to open a “Work in progress (WIP)” pull
-request, *e.g.* “[WIP] Add *brand_new_bert*”, in 🤗 Transformers so that you and the Hugging Face team can work
-side-by-side on integrating the model into 🤗 Transformers.
-
-You should do the following:
-
-1. Create a branch with a descriptive name from your main branch
-
-```bash
-git checkout -b add_brand_new_bert
-```
-
-2. Commit the automatically generated code:
-
-```bash
-git add .
-git commit
-```
-
-3. Fetch and rebase to current main
-
-```bash
-git fetch upstream
-git rebase upstream/main
-```
-
-4. Push the changes to your account using:
-
-```bash
-git push -u origin a-descriptive-name-for-my-changes
-```
-
-5. Once you are satisfied, go to the webpage of your fork on GitHub. Click on “Pull request”. Make sure to add the
-   GitHub handle of some members of the Hugging Face team as reviewers, so that the Hugging Face team gets notified for
-   future changes.
-
-6. Change the PR into a draft by clicking on “Convert to draft” on the right of the GitHub pull request web page.
-
-In the following, whenever you have done some progress, don't forget to commit your work and push it to your account so
-that it shows in the pull request. Additionally, you should make sure to update your work with the current main from
-time to time by doing:
-
-```bash
-git fetch upstream
-git merge upstream/main
-```
-
-In general, all questions you might have regarding the model or your implementation should be asked in your PR and
-discussed/solved in the PR. This way, the Hugging Face team will always be notified when you are committing new code or
-if you have a question. It is often very helpful to point the Hugging Face team to your added code so that the Hugging
-Face team can efficiently understand your problem or question.
-
-To do so, you can go to the “Files changed” tab where you see all of your changes, go to a line regarding which you
-want to ask a question, and click on the “+” symbol to add a comment. Whenever a question or problem has been solved,
-you can click on the “Resolve” button of the created comment.
-
-In the same way, the Hugging Face team will open comments when reviewing your code. We recommend asking most questions
-on GitHub on your PR. For some very general questions that are not very useful for the public, feel free to ping the
-Hugging Face team by Slack or email.
-
-**5. Adapt the generated models code for brand_new_bert**
-
-At first, we will focus only on the model itself and not care about the tokenizer. All the relevant code should be
-found in the generated files `src/transformers/models/brand_new_bert/modeling_brand_new_bert.py` and
-`src/transformers/models/brand_new_bert/configuration_brand_new_bert.py`.
-
-Now you can finally start coding :). The generated code in
-`src/transformers/models/brand_new_bert/modeling_brand_new_bert.py` will either have the same architecture as BERT if
-it's an encoder-only model or BART if it's an encoder-decoder model. At this point, you should remind yourself what
-you've learned in the beginning about the theoretical aspects of the model: *How is the model different from BERT or
-BART?*". Implement those changes which often means to change the *self-attention* layer, the order of the normalization
-layer, etc… Again, it is often useful to look at the similar architecture of already existing models in Transformers to
-get a better feeling of how your model should be implemented.
-
-**Note** that at this point, you don't have to be very sure that your code is fully correct or clean. Rather, it is
-advised to add a first *unclean*, copy-pasted version of the original code to
-`src/transformers/models/brand_new_bert/modeling_brand_new_bert.py` until you feel like all the necessary code is
-added. From our experience, it is much more efficient to quickly add a first version of the required code and
-improve/correct the code iteratively with the conversion script as described in the next section. The only thing that
-has to work at this point is that you can instantiate the 🤗 Transformers implementation of *brand_new_bert*, *i.e.* the
-following command should work:
-
-```python
-from transformers import BrandNewBertModel, BrandNewBertConfig
-
-model = BrandNewBertModel(BrandNewBertConfig())
-```
-
-The above command will create a model according to the default parameters as defined in `BrandNewBertConfig()` with
-random weights, thus making sure that the `init()` methods of all components works.
-
-**6. Write a conversion script**
-
-Next, you should write a conversion script that lets you convert the checkpoint you used to debug *brand_new_bert* in
-the original repository to a checkpoint compatible with your just created 🤗 Transformers implementation of
-*brand_new_bert*. It is not advised to write the conversion script from scratch, but rather to look through already
-existing conversion scripts in 🤗 Transformers for one that has been used to convert a similar model that was written in
-the same framework as *brand_new_bert*. Usually, it is enough to copy an already existing conversion script and
-slightly adapt it for your use case. Don't hesitate to ask the Hugging Face team to point you to a similar already
-existing conversion script for your model.
-
-- If you are porting a model from TensorFlow to PyTorch, a good starting point might be BERT's conversion script [here](https://github.com/huggingface/transformers/blob/7acfa95afb8194f8f9c1f4d2c6028224dbed35a2/src/transformers/models/bert/modeling_bert.py#L91)
-- If you are porting a model from PyTorch to PyTorch, a good starting point might be BART's conversion script [here](https://github.com/huggingface/transformers/blob/main/src/transformers/models/bart/convert_bart_original_pytorch_checkpoint_to_pytorch.py)
-
-In the following, we'll quickly explain how PyTorch models store layer weights and define layer names. In PyTorch, the
-name of a layer is defined by the name of the class attribute you give the layer. Let's define a dummy model in
-PyTorch, called `SimpleModel` as follows:
-
-```python
-from torch import nn
-
-
-class SimpleModel(nn.Module):
-    def __init__(self):
-        super().__init__()
-        self.dense = nn.Linear(10, 10)
-        self.intermediate = nn.Linear(10, 10)
-        self.layer_norm = nn.LayerNorm(10)
-```
-
-Now we can create an instance of this model definition which will fill all weights: `dense`, `intermediate`,
-`layer_norm` with random weights. We can print the model to see its architecture
-
-```python
-model = SimpleModel()
-
-print(model)
-```
-
-This will print out the following:
-
-```
-SimpleModel(
-  (dense): Linear(in_features=10, out_features=10, bias=True)
-  (intermediate): Linear(in_features=10, out_features=10, bias=True)
-  (layer_norm): LayerNorm((10,), eps=1e-05, elementwise_affine=True)
-)
-```
-
-We can see that the layer names are defined by the name of the class attribute in PyTorch. You can print out the weight
-values of a specific layer:
-
-```python
-print(model.dense.weight.data)
-```
-
-to see that the weights were randomly initialized
-
-```
-tensor([[-0.0818,  0.2207, -0.0749, -0.0030,  0.0045, -0.1569, -0.1598,  0.0212,
-         -0.2077,  0.2157],
-        [ 0.1044,  0.0201,  0.0990,  0.2482,  0.3116,  0.2509,  0.2866, -0.2190,
-          0.2166, -0.0212],
-        [-0.2000,  0.1107, -0.1999, -0.3119,  0.1559,  0.0993,  0.1776, -0.1950,
-         -0.1023, -0.0447],
-        [-0.0888, -0.1092,  0.2281,  0.0336,  0.1817, -0.0115,  0.2096,  0.1415,
-         -0.1876, -0.2467],
-        [ 0.2208, -0.2352, -0.1426, -0.2636, -0.2889, -0.2061, -0.2849, -0.0465,
-          0.2577,  0.0402],
-        [ 0.1502,  0.2465,  0.2566,  0.0693,  0.2352, -0.0530,  0.1859, -0.0604,
-          0.2132,  0.1680],
-        [ 0.1733, -0.2407, -0.1721,  0.1484,  0.0358, -0.0633, -0.0721, -0.0090,
-          0.2707, -0.2509],
-        [-0.1173,  0.1561,  0.2945,  0.0595, -0.1996,  0.2988, -0.0802,  0.0407,
-          0.1829, -0.1568],
-        [-0.1164, -0.2228, -0.0403,  0.0428,  0.1339,  0.0047,  0.1967,  0.2923,
-          0.0333, -0.0536],
-        [-0.1492, -0.1616,  0.1057,  0.1950, -0.2807, -0.2710, -0.1586,  0.0739,
-          0.2220,  0.2358]]).
-```
-
-In the conversion script, you should fill those randomly initialized weights with the exact weights of the
-corresponding layer in the checkpoint. *E.g.*
-
-```python
-# retrieve matching layer weights, e.g. by
-# recursive algorithm
-layer_name = "dense"
-pretrained_weight = array_of_dense_layer
-
-model_pointer = getattr(model, "dense")
-
-model_pointer.weight.data = torch.from_numpy(pretrained_weight)
-```
-
-While doing so, you must verify that each randomly initialized weight of your PyTorch model and its corresponding
-pretrained checkpoint weight exactly match in both **shape and name**. To do so, it is **necessary** to add assert
-statements for the shape and print out the names of the checkpoints weights. E.g. you should add statements like:
-
-```python
-assert (
-    model_pointer.weight.shape == pretrained_weight.shape
-), f"Pointer shape of random weight {model_pointer.shape} and array shape of checkpoint weight {pretrained_weight.shape} mismatched"
-```
-
-Besides, you should also print out the names of both weights to make sure they match, *e.g.*
-
-```python
-logger.info(f"Initialize PyTorch weight {layer_name} from {pretrained_weight.name}")
-```
-
-If either the shape or the name doesn't match, you probably assigned the wrong checkpoint weight to a randomly
-initialized layer of the 🤗 Transformers implementation.
-
-An incorrect shape is most likely due to an incorrect setting of the config parameters in `BrandNewBertConfig()` that
-do not exactly match those that were used for the checkpoint you want to convert. However, it could also be that
-PyTorch's implementation of a layer requires the weight to be transposed beforehand.
-
-Finally, you should also check that **all** required weights are initialized and print out all checkpoint weights that
-were not used for initialization to make sure the model is correctly converted. It is completely normal, that the
-conversion trials fail with either a wrong shape statement or wrong name assignment. This is most likely because either
-you used incorrect parameters in `BrandNewBertConfig()`, have a wrong architecture in the 🤗 Transformers
-implementation, you have a bug in the `init()` functions of one of the components of the 🤗 Transformers
-implementation or you need to transpose one of the checkpoint weights.
-
-This step should be iterated with the previous step until all weights of the checkpoint are correctly loaded in the
-Transformers model. Having correctly loaded the checkpoint into the 🤗 Transformers implementation, you can then save
-the model under a folder of your choice `/path/to/converted/checkpoint/folder` that should then contain both a
-`pytorch_model.bin` file and a `config.json` file:
-
-```python
-model.save_pretrained("/path/to/converted/checkpoint/folder")
-```
-
-**7. Implement the forward pass**
-
-Having managed to correctly load the pretrained weights into the 🤗 Transformers implementation, you should now make
-sure that the forward pass is correctly implemented. In [Get familiar with the original repository](#run-a-pretrained-checkpoint-using-the-original-repository), you have already created a script that runs a forward
-pass of the model using the original repository. Now you should write an analogous script using the 🤗 Transformers
-implementation instead of the original one. It should look as follows:
-
-```python
-model = BrandNewBertModel.from_pretrained("/path/to/converted/checkpoint/folder")
-input_ids = [0, 4, 4, 3, 2, 4, 1, 7, 19]
-output = model(input_ids).last_hidden_states
-```
-
-It is very likely that the 🤗 Transformers implementation and the original model implementation don't give the exact
-same output the very first time or that the forward pass throws an error. Don't be disappointed - it's expected! First,
-you should make sure that the forward pass doesn't throw any errors. It often happens that the wrong dimensions are
-used leading to a *Dimensionality mismatch* error or that the wrong data type object is used, *e.g.* `torch.long`
-instead of `torch.float32`. Don't hesitate to ask the Hugging Face team for help, if you don't manage to solve
-certain errors.
-
-The final part to make sure the 🤗 Transformers implementation works correctly is to ensure that the outputs are
-equivalent to a precision of `1e-3`. First, you should ensure that the output shapes are identical, *i.e.*
-`outputs.shape` should yield the same value for the script of the 🤗 Transformers implementation and the original
-implementation. Next, you should make sure that the output values are identical as well. This one of the most difficult
-parts of adding a new model. Common mistakes why the outputs are not identical are:
-
-- Some layers were not added, *i.e.* an *activation* layer was not added, or the residual connection was forgotten
-- The word embedding matrix was not tied
-- The wrong positional embeddings are used because the original implementation uses on offset
-- Dropout is applied during the forward pass. To fix this make sure *model.training is False* and that no dropout
-  layer is falsely activated during the forward pass, *i.e.* pass *self.training* to [PyTorch's functional dropout](https://pytorch.org/docs/stable/nn.functional.html?highlight=dropout#torch.nn.functional.dropout)
-
-The best way to fix the problem is usually to look at the forward pass of the original implementation and the 🤗
-Transformers implementation side-by-side and check if there are any differences. Ideally, you should debug/print out
-intermediate outputs of both implementations of the forward pass to find the exact position in the network where the 🤗
-Transformers implementation shows a different output than the original implementation. First, make sure that the
-hard-coded `input_ids` in both scripts are identical. Next, verify that the outputs of the first transformation of
-the `input_ids` (usually the word embeddings) are identical. And then work your way up to the very last layer of the
-network. At some point, you will notice a difference between the two implementations, which should point you to the bug
-in the 🤗 Transformers implementation. From our experience, a simple and efficient way is to add many print statements
-in both the original implementation and 🤗 Transformers implementation, at the same positions in the network
-respectively, and to successively remove print statements showing the same values for intermediate presentations.
-
-When you're confident that both implementations yield the same output, verifying the outputs with
-`torch.allclose(original_output, output, atol=1e-3)`, you're done with the most difficult part! Congratulations - the
-work left to be done should be a cakewalk 😊.
-
-**8. Adding all necessary model tests**
-
-At this point, you have successfully added a new model. However, it is very much possible that the model does not yet
-fully comply with the required design. To make sure, the implementation is fully compatible with 🤗 Transformers, all
-common tests should pass. The Cookiecutter should have automatically added a test file for your model, probably under
-the same `tests/test_modeling_brand_new_bert.py`. Run this test file to verify that all common tests pass:
-
-```bash
-pytest tests/test_modeling_brand_new_bert.py
-```
-
-Having fixed all common tests, it is now crucial to ensure that all the nice work you have done is well tested, so that
-
-- a) The community can easily understand your work by looking at specific tests of *brand_new_bert*
-- b) Future changes to your model will not break any important feature of the model.
-
-At first, integration tests should be added. Those integration tests essentially do the same as the debugging scripts
-you used earlier to implement the model to 🤗 Transformers. A template of those model tests is already added by the
-Cookiecutter, called `BrandNewBertModelIntegrationTests` and only has to be filled out by you. To ensure that those
-tests are passing, run
-
-```bash
-RUN_SLOW=1 pytest -sv tests/test_modeling_brand_new_bert.py::BrandNewBertModelIntegrationTests
-```
-
-<Tip>
-
-In case you are using Windows, you should replace `RUN_SLOW=1` with `SET RUN_SLOW=1`
-
-</Tip>
-
-Second, all features that are special to *brand_new_bert* should be tested additionally in a separate test under
-`BrandNewBertModelTester`/``BrandNewBertModelTest`. This part is often forgotten but is extremely useful in two
-ways:
-
-- It helps to transfer the knowledge you have acquired during the model addition to the community by showing how the
-  special features of *brand_new_bert* should work.
-- Future contributors can quickly test changes to the model by running those special tests.
-
-
-**9. Implement the tokenizer**
-
-Next, we should add the tokenizer of *brand_new_bert*. Usually, the tokenizer is equivalent or very similar to an
-already existing tokenizer of 🤗 Transformers.
-
-It is very important to find/extract the original tokenizer file and to manage to load this file into the 🤗
-Transformers' implementation of the tokenizer.
-
-To ensure that the tokenizer works correctly, it is recommended to first create a script in the original repository
-that inputs a string and returns the `input_ids``. It could look similar to this (in pseudo-code):
-
-```python
-input_str = "This is a long example input string containing special characters .$?-, numbers 2872 234 12 and words."
-model = BrandNewBertModel.load_pretrained_checkpoint("/path/to/checkpoint/")
-input_ids = model.tokenize(input_str)
-```
-
-You might have to take a deeper look again into the original repository to find the correct tokenizer function or you
-might even have to do changes to your clone of the original repository to only output the `input_ids`. Having written
-a functional tokenization script that uses the original repository, an analogous script for 🤗 Transformers should be
-created. It should look similar to this:
-
-```python
-from transformers import BrandNewBertTokenizer
-
-input_str = "This is a long example input string containing special characters .$?-, numbers 2872 234 12 and words."
-
-tokenizer = BrandNewBertTokenizer.from_pretrained("/path/to/tokenizer/folder/")
-
-input_ids = tokenizer(input_str).input_ids
-```
-
-When both `input_ids` yield the same values, as a final step a tokenizer test file should also be added.
-
-Analogous to the modeling test files of *brand_new_bert*, the tokenization test files of *brand_new_bert* should
-contain a couple of hard-coded integration tests.
-
-**10. Run End-to-end integration tests**
-
-Having added the tokenizer, you should also add a couple of end-to-end integration tests using both the model and the
-tokenizer to `tests/test_modeling_brand_new_bert.py` in 🤗 Transformers. Such a test should show on a meaningful
-text-to-text sample that the 🤗 Transformers implementation works as expected. A meaningful text-to-text sample can
-include *e.g.* a source-to-target-translation pair, an article-to-summary pair, a question-to-answer pair, etc… If none
-of the ported checkpoints has been fine-tuned on a downstream task it is enough to simply rely on the model tests. In a
-final step to ensure that the model is fully functional, it is advised that you also run all tests on GPU. It can
-happen that you forgot to add some `.to(self.device)` statements to internal tensors of the model, which in such a
-test would show in an error. In case you have no access to a GPU, the Hugging Face team can take care of running those
-tests for you.
-
-**11. Add Docstring**
-
-Now, all the necessary functionality for *brand_new_bert* is added - you're almost done! The only thing left to add is
-a nice docstring and a doc page. The Cookiecutter should have added a template file called
-`docs/source/model_doc/brand_new_bert.rst` that you should fill out. Users of your model will usually first look at
-this page before using your model. Hence, the documentation must be understandable and concise. It is very useful for
-the community to add some *Tips* to show how the model should be used. Don't hesitate to ping the Hugging Face team
-regarding the docstrings.
-
-Next, make sure that the docstring added to `src/transformers/models/brand_new_bert/modeling_brand_new_bert.py` is
-correct and included all necessary inputs and outputs. We have a detailed guide about writing documentation and our docstring format [here](writing-documentation). It is always to good to remind oneself that documentation should
-be treated at least as carefully as the code in 🤗 Transformers since the documentation is usually the first contact
-point of the community with the model.
-
-**Code refactor**
-
-Great, now you have added all the necessary code for *brand_new_bert*. At this point, you should correct some potential
-incorrect code style by running:
-
-```bash
-make style
-```
-
-and verify that your coding style passes the quality check:
-
-```bash
-make quality
-```
-
-There are a couple of other very strict design tests in 🤗 Transformers that might still be failing, which shows up in
-the tests of your pull request. This is often because of some missing information in the docstring or some incorrect
-naming. The Hugging Face team will surely help you if you're stuck here.
-
-Lastly, it is always a good idea to refactor one's code after having ensured that the code works correctly. With all
-tests passing, now it's a good time to go over the added code again and do some refactoring.
-
-You have now finished the coding part, congratulation! 🎉 You are Awesome! 😎
-
-**12. Upload the models to the model hub**
-
-In this final part, you should convert and upload all checkpoints to the model hub and add a model card for each
-uploaded model checkpoint. You can get familiar with the hub functionalities by reading our [Model sharing and uploading Page](model_sharing). You should work alongside the Hugging Face team here to decide on a fitting name for each
-checkpoint and to get the required access rights to be able to upload the model under the author's organization of
-*brand_new_bert*. The `push_to_hub` method, present in all models in `transformers`, is a quick and efficient way to push your checkpoint to the hub. A little snippet is pasted below:
-
-```python
-brand_new_bert.push_to_hub("brand_new_bert")
-# Uncomment the following line to push to an organization.
-# brand_new_bert.push_to_hub("<organization>/brand_new_bert")
-```
-
-It is worth spending some time to create fitting model cards for each checkpoint. The model cards should highlight the
-specific characteristics of this particular checkpoint, *e.g.* On which dataset was the checkpoint
-pretrained/fine-tuned on? On what down-stream task should the model be used? And also include some code on how to
-correctly use the model.
-
-**13. (Optional) Add notebook**
-
-It is very helpful to add a notebook that showcases in-detail how *brand_new_bert* can be used for inference and/or
-fine-tuned on a downstream task. This is not mandatory to merge your PR, but very useful for the community.
-
-**14. Submit your finished PR**
-
-You're done programming now and can move to the last step, which is getting your PR merged into main. Usually, the
-Hugging Face team should have helped you already at this point, but it is worth taking some time to give your finished
-PR a nice description and eventually add comments to your code, if you want to point out certain design choices to your
-reviewer.
-
-### Share your work!!
-
-Now, it's time to get some credit from the community for your work! Having completed a model addition is a major
-contribution to Transformers and the whole NLP community. Your code and the ported pre-trained models will certainly be
-used by hundreds and possibly even thousands of developers and researchers. You should be proud of your work and share
-your achievement with the community.
-
-**You have made another model that is super easy to access for everyone in the community! 🤯**

docs/source/en/add_new_pipeline.mdx

@@ -1,254 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
--->
-
-# How to create a custom pipeline?
-
-In this guide, we will see how to create a custom pipeline and share it on the [Hub](hf.co/models) or add it to the
-Transformers library.
-
-First and foremost, you need to decide the raw entries the pipeline will be able to take. It can be strings, raw bytes,
-dictionaries or whatever seems to be the most likely desired input. Try to keep these inputs as pure Python as possible
-as it makes compatibility easier (even through other languages via JSON). Those will be the `inputs` of the
-pipeline (`preprocess`).
-
-Then define the `outputs`. Same policy as the `inputs`. The simpler, the better. Those will be the outputs of
-`postprocess` method.
-
-Start by inheriting the base class `Pipeline`. with the 4 methods needed to implement `preprocess`,
-`_forward`, `postprocess` and `_sanitize_parameters`.
-
-
-```python
-from transformers import Pipeline
-
-
-class MyPipeline(Pipeline):
-    def _sanitize_parameters(self, **kwargs):
-        preprocess_kwargs = {}
-        if "maybe_arg" in kwargs:
-            preprocess_kwargs["maybe_arg"] = kwargs["maybe_arg"]
-        return preprocess_kwargs, {}, {}
-
-    def preprocess(self, inputs, maybe_arg=2):
-        model_input = Tensor(inputs["input_ids"])
-        return {"model_input": model_input}
-
-    def _forward(self, model_inputs):
-        # model_inputs == {"model_input": model_input}
-        outputs = self.model(**model_inputs)
-        # Maybe {"logits": Tensor(...)}
-        return outputs
-
-    def postprocess(self, model_outputs):
-        best_class = model_outputs["logits"].softmax(-1)
-        return best_class
-```
-
-The structure of this breakdown is to support relatively seamless support for CPU/GPU, while supporting doing
-pre/postprocessing on the CPU on different threads
-
-`preprocess` will take the originally defined inputs, and turn them into something feedable to the model. It might
-contain more information and is usually a `Dict`.
-
-`_forward` is the implementation detail and is not meant to be called directly. `forward` is the preferred
-called method as it contains safeguards to make sure everything is working on the expected device. If anything is
-linked to a real model it belongs in the `_forward` method, anything else is in the preprocess/postprocess.
-
-`postprocess` methods will take the output of `_forward` and turn it into the final output that were decided
-earlier.
-
-`_sanitize_parameters` exists to allow users to pass any parameters whenever they wish, be it at initialization
-time `pipeline(...., maybe_arg=4)` or at call time `pipe = pipeline(...); output = pipe(...., maybe_arg=4)`.
-
-The returns of `_sanitize_parameters` are the 3 dicts of kwargs that will be passed directly to `preprocess`,
-`_forward` and `postprocess`. Don't fill anything if the caller didn't call with any extra parameter. That
-allows to keep the default arguments in the function definition which is always more "natural".
-
-A classic example would be a `top_k` argument in the post processing in classification tasks.
-
-```python
->>> pipe = pipeline("my-new-task")
->>> pipe("This is a test")
-[{"label": "1-star", "score": 0.8}, {"label": "2-star", "score": 0.1}, {"label": "3-star", "score": 0.05}
-{"label": "4-star", "score": 0.025}, {"label": "5-star", "score": 0.025}]
-
->>> pipe("This is a test", top_k=2)
-[{"label": "1-star", "score": 0.8}, {"label": "2-star", "score": 0.1}]
-```
-
-In order to achieve that, we'll update our `postprocess` method with a default parameter to `5`. and edit
-`_sanitize_parameters` to allow this new parameter.
-
-
-```python
-def postprocess(self, model_outputs, top_k=5):
-    best_class = model_outputs["logits"].softmax(-1)
-    # Add logic to handle top_k
-    return best_class
-
-
-def _sanitize_parameters(self, **kwargs):
-    preprocess_kwargs = {}
-    if "maybe_arg" in kwargs:
-        preprocess_kwargs["maybe_arg"] = kwargs["maybe_arg"]
-
-    postprocess_kwargs = {}
-    if "top_k" in kwargs:
-        postprocess_kwargs["top_k"] = kwargs["top_k"]
-    return preprocess_kwargs, {}, postprocess_kwargs
-```
-
-Try to keep the inputs/outputs very simple and ideally JSON-serializable as it makes the pipeline usage very easy
-without requiring users to understand new kind of objects. It's also relatively common to support many different types
-of arguments for ease of use (audio files, can be filenames, URLs or pure bytes)
-
-
-
-## Adding it to the list of supported tasks
-
-To register your `new-task` to the list of supported tasks, you have to add it to the `PIPELINE_REGISTRY`:
-
-```python
-from transformers.pipelines import PIPELINE_REGISTRY
-
-PIPELINE_REGISTRY.register_pipeline(
-    "new-task",
-    pipeline_class=MyPipeline,
-    pt_model=AutoModelForSequenceClassification,
-)
-```
-
-You can specify a default model if you want, in which case it should come with a specific revision (which can be the name of a branch or a commit hash, here we took `"abcdef"`) as well was the type:
-
-```python
-PIPELINE_REGISTRY.register_pipeline(
-    "new-task",
-    pipeline_class=MyPipeline,
-    pt_model=AutoModelForSequenceClassification,
-    default={"pt": ("user/awesome_model", "abcdef")},
-    type="text",  # current support type: text, audio, image, multimodal
-)
-```
-
-## Share your pipeline on the Hub
-
-To share your custom pipeline on the Hub, you just have to save the custom code of your `Pipeline` subclass in a
-python file. For instance, let's say we want to use a custom pipeline for sentence pair classification like this:
-
-```py
-import numpy as np
-
-from transformers import Pipeline
-
-
-def softmax(outputs):
-    maxes = np.max(outputs, axis=-1, keepdims=True)
-    shifted_exp = np.exp(outputs - maxes)
-    return shifted_exp / shifted_exp.sum(axis=-1, keepdims=True)
-
-
-class PairClassificationPipeline(Pipeline):
-    def _sanitize_parameters(self, **kwargs):
-        preprocess_kwargs = {}
-        if "second_text" in kwargs:
-            preprocess_kwargs["second_text"] = kwargs["second_text"]
-        return preprocess_kwargs, {}, {}
-
-    def preprocess(self, text, second_text=None):
-        return self.tokenizer(text, text_pair=second_text, return_tensors=self.framework)
-
-    def _forward(self, model_inputs):
-        return self.model(**model_inputs)
-
-    def postprocess(self, model_outputs):
-        logits = model_outputs.logits[0].numpy()
-        probabilities = softmax(logits)
-
-        best_class = np.argmax(probabilities)
-        label = self.model.config.id2label[best_class]
-        score = probabilities[best_class].item()
-        logits = logits.tolist()
-        return {"label": label, "score": score, "logits": logits}
-```
-
-The implementation is framework agnostic, and will work for PyTorch and TensorFlow models. If we have saved this in
-a file named `pair_classification.py`, we can then import it and register it like this:
-
-```py
-from pair_classification import PairClassificationPipeline
-from transformers.pipelines import PIPELINE_REGISTRY
-from transformers import AutoModelForSequenceClassification, TFAutoModelForSequenceClassification
-
-PIPELINE_REGISTRY.register_pipeline(
-    "pair-classification",
-    pipeline_class=PairClassificationPipeline,
-    pt_model=AutoModelForSequenceClassification,
-    tf_model=TFAutoModelForSequenceClassification,
-)
-```
-
-Once this is done, we can use it with a pretrained model. For instance `sgugger/finetuned-bert-mrpc` has been
-fine-tuned on the MRPC dataset, which classifies pairs of sentences as paraphrases or not.
-
-```py
-from transformers import pipeline
-
-classifier = pipeline("pair-classification", model="sgugger/finetuned-bert-mrpc")
-```
-
-Then we can share it on the Hub by using the `save_pretrained` method in a `Repository`:
-
-```py
-from huggingface_hub import Repository
-
-repo = Repository("test-dynamic-pipeline", clone_from="{your_username}/test-dynamic-pipeline")
-classifier.save_pretrained("test-dynamic-pipeline")
-repo.push_to_hub()
-```
-
-This will copy the file where you defined `PairClassificationPipeline` inside the folder `"test-dynamic-pipeline"`,
-along with saving the model and tokenizer of the pipeline, before pushing everything in the repository
-`{your_username}/test-dynamic-pipeline`. After that anyone can use it as long as they provide the option
-`trust_remote_code=True`:
-
-```py
-from transformers import pipeline
-
-classifier = pipeline(model="{your_username}/test-dynamic-pipeline", trust_remote_code=True)
-```
-
-## Add the pipeline to Transformers
-
-If you want to contribute your pipeline to Transformers, you will need to add a new module in the `pipelines` submodule
-with the code of your pipeline, then add it in the list of tasks defined in `pipelines/__init__.py`.
-
-Then you will need to add tests. Create a new file `tests/test_pipelines_MY_PIPELINE.py` with example with the other tests.
-
-The `run_pipeline_test` function will be very generic and run on small random models on every possible
-architecture as defined by `model_mapping` and `tf_model_mapping`.
-
-This is very important to test future compatibility, meaning if someone adds a new model for
-`XXXForQuestionAnswering` then the pipeline test will attempt to run on it. Because the models are random it's
-impossible to check for actual values, that's why There is a helper `ANY` that will simply attempt to match the
-output of the pipeline TYPE.
-
-You also *need* to implement 2 (ideally 4) tests.
-
-- `test_small_model_pt` : Define 1 small model for this pipeline (doesn't matter if the results don't make sense)
-  and test the pipeline outputs. The results should be the same as `test_small_model_tf`.
-- `test_small_model_tf` : Define 1 small model for this pipeline (doesn't matter if the results don't make sense)
-  and test the pipeline outputs. The results should be the same as `test_small_model_pt`.
-- `test_large_model_pt` (`optional`): Tests the pipeline on a real pipeline where the results are supposed to
-  make sense. These tests are slow and should be marked as such. Here the goal is to showcase the pipeline and to make
-  sure there is no drift in future releases
-- `test_large_model_tf` (`optional`): Tests the pipeline on a real pipeline where the results are supposed to
-  make sense. These tests are slow and should be marked as such. Here the goal is to showcase the pipeline and to make
-  sure there is no drift in future releases

docs/source/en/autoclass_tutorial.mdx

@@ -1,119 +0,0 @@
-<!--Copyright 2022 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Load pretrained instances with an AutoClass
-
-With so many different Transformer architectures, it can be challenging to create one for your checkpoint. As a part of 🤗 Transformers core philosophy to make the library easy, simple and flexible to use, an `AutoClass` automatically infer and load the correct architecture from a given checkpoint. The `from_pretrained` method lets you quickly load a pretrained model for any architecture so you don't have to devote time and resources to train a model from scratch. Producing this type of checkpoint-agnostic code means if your code works for one checkpoint, it will work with another checkpoint - as long as it was trained for a similar task - even if the architecture is different.
-
-<Tip>
-
-Remember, architecture refers to the skeleton of the model and checkpoints are the weights for a given architecture. For example, [BERT](https://huggingface.co/bert-base-uncased) is an architecture, while `bert-base-uncased` is a checkpoint. Model is a general term that can mean either architecture or checkpoint.
-
-</Tip>
-
-In this tutorial, learn to:
-
-* Load a pretrained tokenizer.
-* Load a pretrained feature extractor.
-* Load a pretrained processor.
-* Load a pretrained model.
-
-## AutoTokenizer
-
-Nearly every NLP task begins with a tokenizer. A tokenizer converts your input into a format that can be processed by the model.
-
-Load a tokenizer with [`AutoTokenizer.from_pretrained`]:
-
-```py
->>> from transformers import AutoTokenizer
-
->>> tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
-```
-
-Then tokenize your input as shown below:
-
-```py
->>> sequence = "In a hole in the ground there lived a hobbit."
->>> print(tokenizer(sequence))
-{'input_ids': [101, 1999, 1037, 4920, 1999, 1996, 2598, 2045, 2973, 1037, 7570, 10322, 4183, 1012, 102], 
- 'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], 
- 'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]}
-```
-
-## AutoFeatureExtractor
-
-For audio and vision tasks, a feature extractor processes the audio signal or image into the correct input format.
-
-Load a feature extractor with [`AutoFeatureExtractor.from_pretrained`]:
-
-```py
->>> from transformers import AutoFeatureExtractor
-
->>> feature_extractor = AutoFeatureExtractor.from_pretrained(
-...     "ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition"
-... )
-```
-
-## AutoProcessor
-
-Multimodal tasks require a processor that combines two types of preprocessing tools. For example, the [LayoutLMV2](model_doc/layoutlmv2) model requires a feature extractor to handle images and a tokenizer to handle text; a processor combines both of them.
-
-Load a processor with [`AutoProcessor.from_pretrained`]:
-
-```py
->>> from transformers import AutoProcessor
-
->>> processor = AutoProcessor.from_pretrained("microsoft/layoutlmv2-base-uncased")
-```
-
-## AutoModel
-
-<frameworkcontent>
-<pt>
-Finally, the `AutoModelFor` classes let you load a pretrained model for a given task (see [here](model_doc/auto) for a complete list of available tasks). For example, load a model for sequence classification with [`AutoModelForSequenceClassification.from_pretrained`]:
-
-```py
->>> from transformers import AutoModelForSequenceClassification
-
->>> model = AutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased")
-```
-
-Easily reuse the same checkpoint to load an architecture for a different task:
-
-```py
->>> from transformers import AutoModelForTokenClassification
-
->>> model = AutoModelForTokenClassification.from_pretrained("distilbert-base-uncased")
-```
-
-Generally, we recommend using the `AutoTokenizer` class and the `AutoModelFor` class to load pretrained instances of models. This will ensure you load the correct architecture every time. In the next [tutorial](preprocessing), learn how to use your newly loaded tokenizer, feature extractor and processor to preprocess a dataset for fine-tuning.
-</pt>
-<tf>
-Finally, the `TFAutoModelFor` classes let you load a pretrained model for a given task (see [here](model_doc/auto) for a complete list of available tasks). For example, load a model for sequence classification with [`TFAutoModelForSequenceClassification.from_pretrained`]:
-
-```py
->>> from transformers import TFAutoModelForSequenceClassification
-
->>> model = TFAutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased")
-```
-
-Easily reuse the same checkpoint to load an architecture for a different task:
-
-```py
->>> from transformers import TFAutoModelForTokenClassification
-
->>> model = TFAutoModelForTokenClassification.from_pretrained("distilbert-base-uncased")
-```
-
-Generally, we recommend using the `AutoTokenizer` class and the `TFAutoModelFor` class to load pretrained instances of models. This will ensure you load the correct architecture every time. In the next [tutorial](preprocessing), learn how to use your newly loaded tokenizer, feature extractor and processor to preprocess a dataset for fine-tuning.
-</tf>
-</frameworkcontent>

docs/source/en/benchmarks.mdx

@@ -1,383 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Benchmarks
-
-<Tip warning={true}>
-
-Hugging Face's Benchmarking tools are deprecated and it is advised to use external Benchmarking libraries to measure the speed 
-and memory complexity of Transformer models.
-
-</Tip>
-
-[[open-in-colab]]
-
-Let's take a look at how 🤗 Transformers models can be benchmarked, best practices, and already available benchmarks.
-
-A notebook explaining in more detail how to benchmark 🤗 Transformers models can be found [here](https://github.com/huggingface/notebooks/tree/main/examples/benchmark.ipynb).
-
-## How to benchmark 🤗 Transformers models
-
-The classes [`PyTorchBenchmark`] and [`TensorFlowBenchmark`] allow to flexibly benchmark 🤗 Transformers models. The benchmark classes allow us to measure the _peak memory usage_ and _required time_ for both _inference_ and _training_.
-
-<Tip>
-
-Hereby, _inference_ is defined by a single forward pass, and _training_ is defined by a single forward pass and
-backward pass.
-
-</Tip>
-
-The benchmark classes [`PyTorchBenchmark`] and [`TensorFlowBenchmark`] expect an object of type [`PyTorchBenchmarkArguments`] and
-[`TensorFlowBenchmarkArguments`], respectively, for instantiation. [`PyTorchBenchmarkArguments`] and [`TensorFlowBenchmarkArguments`] are data classes and contain all relevant configurations for their corresponding benchmark class. In the following example, it is shown how a BERT model of type _bert-base-cased_ can be benchmarked.
-
-<frameworkcontent>
-<pt>
-```py
->>> from transformers import PyTorchBenchmark, PyTorchBenchmarkArguments
-
->>> args = PyTorchBenchmarkArguments(models=["bert-base-uncased"], batch_sizes=[8], sequence_lengths=[8, 32, 128, 512])
->>> benchmark = PyTorchBenchmark(args)
-```
-</pt>
-<tf>
-```py
->>> from transformers import TensorFlowBenchmark, TensorFlowBenchmarkArguments
-
->>> args = TensorFlowBenchmarkArguments(
-...     models=["bert-base-uncased"], batch_sizes=[8], sequence_lengths=[8, 32, 128, 512]
-... )
->>> benchmark = TensorFlowBenchmark(args)
-```
-</tf>
-</frameworkcontent>
-
-Here, three arguments are given to the benchmark argument data classes, namely `models`, `batch_sizes`, and
-`sequence_lengths`. The argument `models` is required and expects a `list` of model identifiers from the
-[model hub](https://huggingface.co/models) The `list` arguments `batch_sizes` and `sequence_lengths` define
-the size of the `input_ids` on which the model is benchmarked. There are many more parameters that can be configured
-via the benchmark argument data classes. For more detail on these one can either directly consult the files
-`src/transformers/benchmark/benchmark_args_utils.py`, `src/transformers/benchmark/benchmark_args.py` (for PyTorch)
-and `src/transformers/benchmark/benchmark_args_tf.py` (for Tensorflow). Alternatively, running the following shell
-commands from root will print out a descriptive list of all configurable parameters for PyTorch and Tensorflow
-respectively.
-
-<frameworkcontent>
-<pt>
-```bash
-python examples/pytorch/benchmarking/run_benchmark.py --help
-```
-
-An instantiated benchmark object can then simply be run by calling `benchmark.run()`.
-
-```py
->>> results = benchmark.run()
->>> print(results)
-====================       INFERENCE - SPEED - RESULT       ====================
---------------------------------------------------------------------------------
-Model Name             Batch Size     Seq Length     Time in s                  
---------------------------------------------------------------------------------
-bert-base-uncased          8               8             0.006     
-bert-base-uncased          8               32            0.006     
-bert-base-uncased          8              128            0.018     
-bert-base-uncased          8              512            0.088     
---------------------------------------------------------------------------------
-
-====================      INFERENCE - MEMORY - RESULT       ====================
---------------------------------------------------------------------------------
-Model Name             Batch Size     Seq Length    Memory in MB 
---------------------------------------------------------------------------------
-bert-base-uncased          8               8             1227
-bert-base-uncased          8               32            1281
-bert-base-uncased          8              128            1307
-bert-base-uncased          8              512            1539
---------------------------------------------------------------------------------
-
-====================        ENVIRONMENT INFORMATION         ====================
-
-- transformers_version: 2.11.0
-- framework: PyTorch
-- use_torchscript: False
-- framework_version: 1.4.0
-- python_version: 3.6.10
-- system: Linux
-- cpu: x86_64
-- architecture: 64bit
-- date: 2020-06-29
-- time: 08:58:43.371351
-- fp16: False
-- use_multiprocessing: True
-- only_pretrain_model: False
-- cpu_ram_mb: 32088
-- use_gpu: True
-- num_gpus: 1
-- gpu: TITAN RTX
-- gpu_ram_mb: 24217
-- gpu_power_watts: 280.0
-- gpu_performance_state: 2
-- use_tpu: False
-```
-</pt>
-<tf>
-```bash
-python examples/tensorflow/benchmarking/run_benchmark_tf.py --help
-```
-
-An instantiated benchmark object can then simply be run by calling `benchmark.run()`.
-
-```py
->>> results = benchmark.run()
->>> print(results)
->>> results = benchmark.run()
->>> print(results)
-====================       INFERENCE - SPEED - RESULT       ====================
---------------------------------------------------------------------------------
-Model Name             Batch Size     Seq Length     Time in s                  
---------------------------------------------------------------------------------
-bert-base-uncased          8               8             0.005
-bert-base-uncased          8               32            0.008
-bert-base-uncased          8              128            0.022
-bert-base-uncased          8              512            0.105
---------------------------------------------------------------------------------
-
-====================      INFERENCE - MEMORY - RESULT       ====================
---------------------------------------------------------------------------------
-Model Name             Batch Size     Seq Length    Memory in MB 
---------------------------------------------------------------------------------
-bert-base-uncased          8               8             1330
-bert-base-uncased          8               32            1330
-bert-base-uncased          8              128            1330
-bert-base-uncased          8              512            1770
---------------------------------------------------------------------------------
-
-====================        ENVIRONMENT INFORMATION         ====================
-
-- transformers_version: 2.11.0
-- framework: Tensorflow
-- use_xla: False
-- framework_version: 2.2.0
-- python_version: 3.6.10
-- system: Linux
-- cpu: x86_64
-- architecture: 64bit
-- date: 2020-06-29
-- time: 09:26:35.617317
-- fp16: False
-- use_multiprocessing: True
-- only_pretrain_model: False
-- cpu_ram_mb: 32088
-- use_gpu: True
-- num_gpus: 1
-- gpu: TITAN RTX
-- gpu_ram_mb: 24217
-- gpu_power_watts: 280.0
-- gpu_performance_state: 2
-- use_tpu: False
-```
-</tf>
-</frameworkcontent>
-
-By default, the _time_ and the _required memory_ for _inference_ are benchmarked. In the example output above the first
-two sections show the result corresponding to _inference time_ and _inference memory_. In addition, all relevant
-information about the computing environment, _e.g._ the GPU type, the system, the library versions, etc... are printed
-out in the third section under _ENVIRONMENT INFORMATION_. This information can optionally be saved in a _.csv_ file
-when adding the argument `save_to_csv=True` to [`PyTorchBenchmarkArguments`] and
-[`TensorFlowBenchmarkArguments`] respectively. In this case, every section is saved in a separate
-_.csv_ file. The path to each _.csv_ file can optionally be defined via the argument data classes.
-
-Instead of benchmarking pre-trained models via their model identifier, _e.g._ `bert-base-uncased`, the user can
-alternatively benchmark an arbitrary configuration of any available model class. In this case, a `list` of
-configurations must be inserted with the benchmark args as follows.
-
-<frameworkcontent>
-<pt>
-```py
->>> from transformers import PyTorchBenchmark, PyTorchBenchmarkArguments, BertConfig
-
->>> args = PyTorchBenchmarkArguments(
-...     models=["bert-base", "bert-384-hid", "bert-6-lay"], batch_sizes=[8], sequence_lengths=[8, 32, 128, 512]
-... )
->>> config_base = BertConfig()
->>> config_384_hid = BertConfig(hidden_size=384)
->>> config_6_lay = BertConfig(num_hidden_layers=6)
-
->>> benchmark = PyTorchBenchmark(args, configs=[config_base, config_384_hid, config_6_lay])
->>> benchmark.run()
-====================       INFERENCE - SPEED - RESULT       ====================
---------------------------------------------------------------------------------
-Model Name             Batch Size     Seq Length       Time in s                  
---------------------------------------------------------------------------------
-bert-base                  8              128            0.006
-bert-base                  8              512            0.006
-bert-base                  8              128            0.018     
-bert-base                  8              512            0.088     
-bert-384-hid              8               8             0.006     
-bert-384-hid              8               32            0.006     
-bert-384-hid              8              128            0.011     
-bert-384-hid              8              512            0.054     
-bert-6-lay                 8               8             0.003     
-bert-6-lay                 8               32            0.004     
-bert-6-lay                 8              128            0.009     
-bert-6-lay                 8              512            0.044
---------------------------------------------------------------------------------
-
-====================      INFERENCE - MEMORY - RESULT       ====================
---------------------------------------------------------------------------------
-Model Name             Batch Size     Seq Length      Memory in MB 
---------------------------------------------------------------------------------
-bert-base                  8               8             1277
-bert-base                  8               32            1281
-bert-base                  8              128            1307     
-bert-base                  8              512            1539     
-bert-384-hid              8               8             1005     
-bert-384-hid              8               32            1027     
-bert-384-hid              8              128            1035     
-bert-384-hid              8              512            1255     
-bert-6-lay                 8               8             1097     
-bert-6-lay                 8               32            1101     
-bert-6-lay                 8              128            1127     
-bert-6-lay                 8              512            1359
---------------------------------------------------------------------------------
-
-====================        ENVIRONMENT INFORMATION         ====================
-
-- transformers_version: 2.11.0
-- framework: PyTorch
-- use_torchscript: False
-- framework_version: 1.4.0
-- python_version: 3.6.10
-- system: Linux
-- cpu: x86_64
-- architecture: 64bit
-- date: 2020-06-29
-- time: 09:35:25.143267
-- fp16: False
-- use_multiprocessing: True
-- only_pretrain_model: False
-- cpu_ram_mb: 32088
-- use_gpu: True
-- num_gpus: 1
-- gpu: TITAN RTX
-- gpu_ram_mb: 24217
-- gpu_power_watts: 280.0
-- gpu_performance_state: 2
-- use_tpu: False
-```
-</pt>
-<tf>
-```py
->>> from transformers import TensorFlowBenchmark, TensorFlowBenchmarkArguments, BertConfig
-
->>> args = TensorFlowBenchmarkArguments(
-...     models=["bert-base", "bert-384-hid", "bert-6-lay"], batch_sizes=[8], sequence_lengths=[8, 32, 128, 512]
-... )
->>> config_base = BertConfig()
->>> config_384_hid = BertConfig(hidden_size=384)
->>> config_6_lay = BertConfig(num_hidden_layers=6)
-
->>> benchmark = TensorFlowBenchmark(args, configs=[config_base, config_384_hid, config_6_lay])
->>> benchmark.run()
-====================       INFERENCE - SPEED - RESULT       ====================
---------------------------------------------------------------------------------
-Model Name             Batch Size     Seq Length       Time in s                  
---------------------------------------------------------------------------------
-bert-base                  8               8             0.005
-bert-base                  8               32            0.008
-bert-base                  8              128            0.022
-bert-base                  8              512            0.106
-bert-384-hid              8               8             0.005
-bert-384-hid              8               32            0.007
-bert-384-hid              8              128            0.018
-bert-384-hid              8              512            0.064
-bert-6-lay                 8               8             0.002
-bert-6-lay                 8               32            0.003
-bert-6-lay                 8              128            0.0011
-bert-6-lay                 8              512            0.074
---------------------------------------------------------------------------------
-
-====================      INFERENCE - MEMORY - RESULT       ====================
---------------------------------------------------------------------------------
-Model Name             Batch Size     Seq Length      Memory in MB 
---------------------------------------------------------------------------------
-bert-base                  8               8             1330
-bert-base                  8               32            1330
-bert-base                  8              128            1330
-bert-base                  8              512            1770
-bert-384-hid              8               8             1330
-bert-384-hid              8               32            1330
-bert-384-hid              8              128            1330
-bert-384-hid              8              512            1540
-bert-6-lay                 8               8             1330
-bert-6-lay                 8               32            1330
-bert-6-lay                 8              128            1330
-bert-6-lay                 8              512            1540
---------------------------------------------------------------------------------
-
-====================        ENVIRONMENT INFORMATION         ====================
-
-- transformers_version: 2.11.0
-- framework: Tensorflow
-- use_xla: False
-- framework_version: 2.2.0
-- python_version: 3.6.10
-- system: Linux
-- cpu: x86_64
-- architecture: 64bit
-- date: 2020-06-29
-- time: 09:38:15.487125
-- fp16: False
-- use_multiprocessing: True
-- only_pretrain_model: False
-- cpu_ram_mb: 32088
-- use_gpu: True
-- num_gpus: 1
-- gpu: TITAN RTX
-- gpu_ram_mb: 24217
-- gpu_power_watts: 280.0
-- gpu_performance_state: 2
-- use_tpu: False
-```
-</tf>
-</frameworkcontent>
-
-Again, _inference time_ and _required memory_ for _inference_ are measured, but this time for customized configurations
-of the `BertModel` class. This feature can especially be helpful when deciding for which configuration the model
-should be trained.
-
-
-## Benchmark best practices
-
-This section lists a couple of best practices one should be aware of when benchmarking a model.
-
-- Currently, only single device benchmarking is supported. When benchmarking on GPU, it is recommended that the user
-  specifies on which device the code should be run by setting the `CUDA_VISIBLE_DEVICES` environment variable in the
-  shell, _e.g._ `export CUDA_VISIBLE_DEVICES=0` before running the code.
-- The option `no_multi_processing` should only be set to `True` for testing and debugging. To ensure accurate
-  memory measurement it is recommended to run each memory benchmark in a separate process by making sure
-  `no_multi_processing` is set to `True`.
-- One should always state the environment information when sharing the results of a model benchmark. Results can vary
-  heavily between different GPU devices, library versions, etc., so that benchmark results on their own are not very
-  useful for the community.
-
-
-## Sharing your benchmark
-
-Previously all available core models (10 at the time) have been benchmarked for _inference time_, across many different
-settings: using PyTorch, with and without TorchScript, using TensorFlow, with and without XLA. All of those tests were
-done across CPUs (except for TensorFlow XLA) and GPUs.
-
-The approach is detailed in the [following blogpost](https://medium.com/huggingface/benchmarking-transformers-pytorch-and-tensorflow-e2917fb891c2) and the results are
-available [here](https://docs.google.com/spreadsheets/d/1sryqufw2D0XlUH4sq3e9Wnxu5EAQkaohzrJbd5HdQ_w/edit?usp=sharing).
-
-With the new _benchmark_ tools, it is easier than ever to share your benchmark results with the community
-
-- [PyTorch Benchmarking Results](https://github.com/huggingface/transformers/tree/main/examples/pytorch/benchmarking/README.md).
-- [TensorFlow Benchmarking Results](https://github.com/huggingface/transformers/tree/main/examples/tensorflow/benchmarking/README.md).

docs/source/en/bertology.mdx

@@ -1,36 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# BERTology
-
-There is a growing field of study concerned with investigating the inner working of large-scale transformers like BERT
-(that some call "BERTology"). Some good examples of this field are:
-
-
-- BERT Rediscovers the Classical NLP Pipeline by Ian Tenney, Dipanjan Das, Ellie Pavlick:
-  https://arxiv.org/abs/1905.05950
-- Are Sixteen Heads Really Better than One? by Paul Michel, Omer Levy, Graham Neubig: https://arxiv.org/abs/1905.10650
-- What Does BERT Look At? An Analysis of BERT's Attention by Kevin Clark, Urvashi Khandelwal, Omer Levy, Christopher D.
-  Manning: https://arxiv.org/abs/1906.04341
-
-In order to help this new field develop, we have included a few additional features in the BERT/GPT/GPT-2 models to
-help people access the inner representations, mainly adapted from the great work of Paul Michel
-(https://arxiv.org/abs/1905.10650):
-
-
-- accessing all the hidden-states of BERT/GPT/GPT-2,
-- accessing all the attention weights for each head of BERT/GPT/GPT-2,
-- retrieving heads output values and gradients to be able to compute head importance score and prune head as explained
-  in https://arxiv.org/abs/1905.10650.
-
-To help you understand and use these features, we have added a specific example script: [bertology.py](https://github.com/huggingface/transformers/tree/main/examples/research_projects/bertology/run_bertology.py) while extract information and prune a model pre-trained on
-GLUE.

docs/source/en/big_models.mdx

@@ -1,119 +0,0 @@
-<!--Copyright 2022 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Instantiating a big model
-
-When you want to use a very big pretrained model, one challenge is to minimize the use of the RAM. The usual workflow
-from PyTorch is:
-
-1. Create your model with random weights.
-2. Load your pretrained weights.
-3. Put those pretrained weights in your random model.
-
-Step 1 and 2 both require a full version of the model in memory, which is not a problem in most cases, but if your model starts weighing several GigaBytes, those two copies can make you got our of RAM. Even worse, if you are using `torch.distributed` to launch a distributed training, each process will load the pretrained model and store these two copies in RAM.
-
-<Tip>
-
-Note that the randomly created model is initialized with "empty" tensors, which take the space in memory without filling it (thus the random values are whatever was in this chunk of memory at a given time). The random initialization following the appropriate distribution for the kind of model/parameters instatiated (like a normal distribution for instance) is only performed after step 3 on the non-initialized weights, to be as fast as possible! 
-
-</Tip>
-
-In this guide, we explore the solutions Transformers offer to deal with this issue. Note that this is an area of active development, so the APIs explained here may change slightly in the future.
-
-## Sharded checkpoints
-
-Since version 4.18.0, model checkpoints that end up taking more than 10GB of space are automatically sharded in smaller pieces. In terms of having one single checkpoint when you do `model.save_pretrained(save_dir)`, you will end up with several partial checkpoints (each of which being of size < 10GB) and an index that maps parameter names to the files they are stored in.
-
-You can control the maximum size before sharding with the `max_shard_size` parameter, so for the sake of an example, we'll use a normal-size models with a small shard size: let's take a traditional BERT model.
-
-```py
-from transformers import AutoModel
-
-model = AutoModel.from_pretrained("bert-base-cased")
-```
-
-If you save it using [`~PreTrainedModel.save_pretrained`], you will get a new folder with two files: the config of the model and its weights:
-
-```py
->>> import os
->>> import tempfile
-
->>> with tempfile.TemporaryDirectory() as tmp_dir:
-...     model.save_pretrained(tmp_dir)
-...     print(sorted(os.listdir(tmp_dir)))
-['config.json', 'pytorch_model.bin']
-```
-
-Now let's use a maximum shard size of 200MB:
-
-```py
->>> with tempfile.TemporaryDirectory() as tmp_dir:
-...     model.save_pretrained(tmp_dir, max_shard_size="200MB")
-...     print(sorted(os.listdir(tmp_dir)))
-['config.json', 'pytorch_model-00001-of-00003.bin', 'pytorch_model-00002-of-00003.bin', 'pytorch_model-00003-of-00003.bin', 'pytorch_model.bin.index.json']
-```
-
-On top of the configuration of the model, we see three different weights files, and an `index.json` file which is our index. A checkpoint like this can be fully reloaded using the [`~PreTrainedModel.from_pretrained`] method:
-
-```py
->>> with tempfile.TemporaryDirectory() as tmp_dir:
-...     model.save_pretrained(tmp_dir, max_shard_size="200MB")
-...     new_model = AutoModel.from_pretrained(tmp_dir)
-```
-
-The main advantage of doing this for big models is that during step 2 of the workflow shown above, each shard of the checkpoint is loaded after the previous one, capping the memory usage in RAM to the model size plus the size of the biggest shard.
-
-Beind the scenes, the index file is used to determine which keys are in the checkpoint, and where the corresponding weights are stored. We can load that index like any json and get a dictionary:
-
-```py
->>> import json
-
->>> with tempfile.TemporaryDirectory() as tmp_dir:
-...     model.save_pretrained(tmp_dir, max_shard_size="200MB")
-...     with open(os.path.join(tmp_dir, "pytorch_model.bin.index.json"), "r") as f:
-...         index = json.load(f)
-
->>> print(index.keys())
-dict_keys(['metadata', 'weight_map'])
-```
-
-The metadata just consists of the total size of the model for now. We plan to add several other informations in the future:
-
-```py
->>> index["metadata"]
-{'total_size': 433245184}
-```
-
-The weights map is the main part of this index, which maps each parameter name (as usually found in a PyTorch model `state_dict`) to the file it's stored in:
-
-```py
->>> index["weight_map"]
-{'embeddings.LayerNorm.bias': 'pytorch_model-00001-of-00003.bin',
- 'embeddings.LayerNorm.weight': 'pytorch_model-00001-of-00003.bin',
- ...
-```
-
-If you want to directly load such a sharded checkpoint inside a model without using [`~PreTrainedModel.from_pretrained`] (like you would do `model.load_state_dict()` for a full checkpoint) you should use [`~modeling_utils.load_sharded_checkpoint`]:
-
-```py
->>> from transformers.modeling_utils import load_sharded_checkpoint
-
->>> with tempfile.TemporaryDirectory() as tmp_dir:
-...     model.save_pretrained(tmp_dir, max_shard_size="200MB")
-...     load_sharded_checkpoint(model, tmp_dir)
-```
-
-## Low memory loading
-
-Sharded checkpoints reduce the memory usage during step 2 of the workflow mentioned above, but in order to use that model in a low memory setting, we recommend leveraging our tools based on the Accelerate library.
-
-Please read the following guide for more information: [Large model loading using Accelerate](./main_classes/model#large-model-loading)

docs/source/en/community.mdx

@@ -1,65 +0,0 @@
-# Community
-
-This page regroups resources around 🤗 Transformers developed by the community.
-
-## Community resources:
-
-| Resource     |      Description      |      Author      |
-|:----------|:-------------|------:|
-| [Hugging Face Transformers Glossary Flashcards](https://www.darigovresearch.com/huggingface-transformers-glossary-flashcards) | A set of flashcards based on the [Transformers Docs Glossary](glossary) that has been put into a form which can be easily learnt/revised using [Anki ](https://apps.ankiweb.net/) an open source, cross platform app specifically designed for long term knowledge retention. See this [Introductory video on how to use the flashcards](https://www.youtube.com/watch?v=Dji_h7PILrw). | [Darigov Research](https://www.darigovresearch.com/) |
-
-## Community notebooks:
-
-| Notebook     |      Description      |      Author      |      |
-|:----------|:-------------|:-------------|------:|
-| [Fine-tune a pre-trained Transformer to generate lyrics](https://github.com/AlekseyKorshuk/huggingartists) | How to generate lyrics in the style of your favorite artist by fine-tuning a GPT-2 model |  [Aleksey Korshuk](https://github.com/AlekseyKorshuk) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/AlekseyKorshuk/huggingartists/blob/master/huggingartists-demo.ipynb) |
-| [Train T5 in Tensorflow 2 ](https://github.com/snapthat/TF-T5-text-to-text) | How to train T5 for any task using Tensorflow 2. This notebook demonstrates a Question & Answer task implemented in Tensorflow 2 using SQUAD | [Muhammad Harris](https://github.com/HarrisDePerceptron) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/snapthat/TF-T5-text-to-text/blob/master/snapthatT5/notebooks/TF-T5-Datasets%20Training.ipynb) |
-| [Train T5 on TPU](https://github.com/patil-suraj/exploring-T5/blob/master/T5_on_TPU.ipynb)  | How to train T5 on SQUAD with Transformers and Nlp | [Suraj Patil](https://github.com/patil-suraj) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/patil-suraj/exploring-T5/blob/master/T5_on_TPU.ipynb#scrollTo=QLGiFCDqvuil) |
-| [Fine-tune T5 for Classification and Multiple Choice](https://github.com/patil-suraj/exploring-T5/blob/master/t5_fine_tuning.ipynb)  | How to fine-tune T5 for classification and multiple choice tasks using a text-to-text format with PyTorch Lightning |  [Suraj Patil](https://github.com/patil-suraj) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/patil-suraj/exploring-T5/blob/master/t5_fine_tuning.ipynb) |
-| [Fine-tune DialoGPT on New Datasets and Languages](https://github.com/ncoop57/i-am-a-nerd/blob/master/_notebooks/2020-05-12-chatbot-part-1.ipynb)  | How to fine-tune the DialoGPT model on a new dataset for open-dialog conversational chatbots |  [Nathan Cooper](https://github.com/ncoop57) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/ncoop57/i-am-a-nerd/blob/master/_notebooks/2020-05-12-chatbot-part-1.ipynb) |
-| [Long Sequence Modeling with Reformer](https://github.com/patrickvonplaten/notebooks/blob/master/PyTorch_Reformer.ipynb)  | How to train on sequences as long as 500,000 tokens with Reformer |  [Patrick von Platen](https://github.com/patrickvonplaten) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/PyTorch_Reformer.ipynb)  |
-| [Fine-tune BART for Summarization](https://github.com/ohmeow/ohmeow_website/blob/master/_notebooks/2020-05-23-text-generation-with-blurr.ipynb) | How to fine-tune BART for summarization with fastai using blurr | [Wayde Gilliam](https://ohmeow.com/) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/ohmeow/ohmeow_website/blob/master/_notebooks/2020-05-23-text-generation-with-blurr.ipynb) |
-| [Fine-tune a pre-trained Transformer on anyone's tweets](https://colab.research.google.com/github/borisdayma/huggingtweets/blob/master/huggingtweets-demo.ipynb) | How to generate tweets in the style of your favorite Twitter account by fine-tuning a GPT-2 model |  [Boris Dayma](https://github.com/borisdayma) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/borisdayma/huggingtweets/blob/master/huggingtweets-demo.ipynb) |
-| [Optimize 🤗 Hugging Face models with Weights & Biases](https://colab.research.google.com/github/wandb/examples/blob/master/colabs/huggingface/Optimize_Hugging_Face_models_with_Weights_%26_Biases.ipynb) | A complete tutorial showcasing W&B integration with Hugging Face | [Boris Dayma](https://github.com/borisdayma) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/wandb/examples/blob/master/colabs/huggingface/Optimize_Hugging_Face_models_with_Weights_%26_Biases.ipynb) |
-| [Pretrain Longformer](https://github.com/allenai/longformer/blob/master/scripts/convert_model_to_long.ipynb)  | How to build a "long" version of existing pretrained models |  [Iz Beltagy](https://beltagy.net) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/allenai/longformer/blob/master/scripts/convert_model_to_long.ipynb) |
-| [Fine-tune Longformer for QA](https://github.com/patil-suraj/Notebooks/blob/master/longformer_qa_training.ipynb) | How to fine-tune longformer model for QA task | [Suraj Patil](https://github.com/patil-suraj) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/patil-suraj/Notebooks/blob/master/longformer_qa_training.ipynb) |
-| [Evaluate Model with 🤗nlp](https://github.com/patrickvonplaten/notebooks/blob/master/How_to_evaluate_Longformer_on_TriviaQA_using_NLP.ipynb) | How to evaluate longformer on TriviaQA with `nlp` | [Patrick von Platen](https://github.com/patrickvonplaten) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1m7eTGlPmLRgoPkkA7rkhQdZ9ydpmsdLE?usp=sharing) |
-| [Fine-tune T5 for Sentiment Span Extraction](https://github.com/enzoampil/t5-intro/blob/master/t5_qa_training_pytorch_span_extraction.ipynb)  | How to fine-tune T5 for sentiment span extraction using a text-to-text format with PyTorch Lightning |  [Lorenzo Ampil](https://github.com/enzoampil) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/enzoampil/t5-intro/blob/master/t5_qa_training_pytorch_span_extraction.ipynb) |
-| [Fine-tune DistilBert for Multiclass Classification](https://github.com/abhimishra91/transformers-tutorials/blob/master/transformers_multiclass_classification.ipynb) | How to fine-tune DistilBert for multiclass classification with PyTorch | [Abhishek Kumar Mishra](https://github.com/abhimishra91) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/abhimishra91/transformers-tutorials/blob/master/transformers_multiclass_classification.ipynb)|
-|[Fine-tune BERT for Multi-label Classification](https://github.com/abhimishra91/transformers-tutorials/blob/master/transformers_multi_label_classification.ipynb)|How to fine-tune BERT for multi-label classification using PyTorch|[Abhishek Kumar Mishra](https://github.com/abhimishra91) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/abhimishra91/transformers-tutorials/blob/master/transformers_multi_label_classification.ipynb)|
-|[Fine-tune T5 for Summarization](https://github.com/abhimishra91/transformers-tutorials/blob/master/transformers_summarization_wandb.ipynb)|How to fine-tune T5 for summarization in PyTorch and track experiments with WandB|[Abhishek Kumar Mishra](https://github.com/abhimishra91) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/abhimishra91/transformers-tutorials/blob/master/transformers_summarization_wandb.ipynb)|
-|[Speed up Fine-Tuning in Transformers with Dynamic Padding / Bucketing](https://github.com/ELS-RD/transformers-notebook/blob/master/Divide_Hugging_Face_Transformers_training_time_by_2_or_more.ipynb)|How to speed up fine-tuning by a factor of 2 using dynamic padding / bucketing|[Michael Benesty](https://github.com/pommedeterresautee) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1CBfRU1zbfu7-ijiOqAAQUA-RJaxfcJoO?usp=sharing)|
-|[Pretrain Reformer for Masked Language Modeling](https://github.com/patrickvonplaten/notebooks/blob/master/Reformer_For_Masked_LM.ipynb)| How to train a Reformer model with bi-directional self-attention layers | [Patrick von Platen](https://github.com/patrickvonplaten) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1tzzh0i8PgDQGV3SMFUGxM7_gGae3K-uW?usp=sharing)|
-|[Expand and Fine Tune Sci-BERT](https://github.com/lordtt13/word-embeddings/blob/master/COVID-19%20Research%20Data/COVID-SciBERT.ipynb)| How to increase vocabulary of a pretrained SciBERT model from AllenAI on the CORD dataset and pipeline it. | [Tanmay Thakur](https://github.com/lordtt13) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1rqAR40goxbAfez1xvF3hBJphSCsvXmh8)|
-|[Fine Tune BlenderBotSmall for Summarization using the Trainer API](https://github.com/lordtt13/transformers-experiments/blob/master/Custom%20Tasks/fine-tune-blenderbot_small-for-summarization.ipynb)| How to fine tune BlenderBotSmall for summarization on a custom dataset, using the Trainer API. | [Tanmay Thakur](https://github.com/lordtt13) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/19Wmupuls7mykSGyRN_Qo6lPQhgp56ymq?usp=sharing)|
-|[Fine-tune Electra and interpret with Integrated Gradients](https://github.com/elsanns/xai-nlp-notebooks/blob/master/electra_fine_tune_interpret_captum_ig.ipynb) | How to fine-tune Electra for sentiment analysis and interpret predictions with Captum Integrated Gradients | [Eliza Szczechla](https://elsanns.github.io) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/elsanns/xai-nlp-notebooks/blob/master/electra_fine_tune_interpret_captum_ig.ipynb)|
-|[fine-tune a non-English GPT-2 Model with Trainer class](https://github.com/philschmid/fine-tune-GPT-2/blob/master/Fine_tune_a_non_English_GPT_2_Model_with_Huggingface.ipynb) | How to fine-tune a non-English GPT-2 Model with Trainer class | [Philipp Schmid](https://www.philschmid.de) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/philschmid/fine-tune-GPT-2/blob/master/Fine_tune_a_non_English_GPT_2_Model_with_Huggingface.ipynb)|
-|[Fine-tune a DistilBERT Model for Multi Label Classification task](https://github.com/DhavalTaunk08/Transformers_scripts/blob/master/Transformers_multilabel_distilbert.ipynb) | How to fine-tune a DistilBERT Model for Multi Label Classification task | [Dhaval Taunk](https://github.com/DhavalTaunk08) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/DhavalTaunk08/Transformers_scripts/blob/master/Transformers_multilabel_distilbert.ipynb)|
-|[Fine-tune ALBERT for sentence-pair classification](https://github.com/NadirEM/nlp-notebooks/blob/master/Fine_tune_ALBERT_sentence_pair_classification.ipynb) | How to fine-tune an ALBERT model or another BERT-based model for the sentence-pair classification task | [Nadir El Manouzi](https://github.com/NadirEM) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/NadirEM/nlp-notebooks/blob/master/Fine_tune_ALBERT_sentence_pair_classification.ipynb)|
-|[Fine-tune Roberta for sentiment analysis](https://github.com/DhavalTaunk08/NLP_scripts/blob/master/sentiment_analysis_using_roberta.ipynb) | How to fine-tune a Roberta model for sentiment analysis | [Dhaval Taunk](https://github.com/DhavalTaunk08) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/DhavalTaunk08/NLP_scripts/blob/master/sentiment_analysis_using_roberta.ipynb)|
-|[Evaluating Question Generation Models](https://github.com/flexudy-pipe/qugeev) | How accurate are the answers to questions generated by your seq2seq transformer model? | [Pascal Zoleko](https://github.com/zolekode) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1bpsSqCQU-iw_5nNoRm_crPq6FRuJthq_?usp=sharing)|
-|[Classify text with DistilBERT and Tensorflow](https://github.com/peterbayerle/huggingface_notebook/blob/main/distilbert_tf.ipynb) | How to fine-tune DistilBERT for text classification in TensorFlow | [Peter Bayerle](https://github.com/peterbayerle) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/peterbayerle/huggingface_notebook/blob/main/distilbert_tf.ipynb)|
-|[Leverage BERT for Encoder-Decoder Summarization on CNN/Dailymail](https://github.com/patrickvonplaten/notebooks/blob/master/BERT2BERT_for_CNN_Dailymail.ipynb) | How to warm-start a *EncoderDecoderModel* with a *bert-base-uncased* checkpoint for summarization on CNN/Dailymail | [Patrick von Platen](https://github.com/patrickvonplaten) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/BERT2BERT_for_CNN_Dailymail.ipynb)|
-|[Leverage RoBERTa for Encoder-Decoder Summarization on BBC XSum](https://github.com/patrickvonplaten/notebooks/blob/master/RoBERTaShared_for_BBC_XSum.ipynb) | How to warm-start a shared *EncoderDecoderModel* with a *roberta-base* checkpoint for summarization on BBC/XSum | [Patrick von Platen](https://github.com/patrickvonplaten) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/RoBERTaShared_for_BBC_XSum.ipynb)|
-|[Fine-tune TAPAS on Sequential Question Answering (SQA)](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/TAPAS/Fine_tuning_TapasForQuestionAnswering_on_SQA.ipynb) | How to fine-tune *TapasForQuestionAnswering* with a *tapas-base* checkpoint on the Sequential Question Answering (SQA) dataset | [Niels Rogge](https://github.com/nielsrogge) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/TAPAS/Fine_tuning_TapasForQuestionAnswering_on_SQA.ipynb)|
-|[Evaluate TAPAS on Table Fact Checking (TabFact)](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/TAPAS/Evaluating_TAPAS_on_the_Tabfact_test_set.ipynb) | How to evaluate a fine-tuned *TapasForSequenceClassification* with a *tapas-base-finetuned-tabfact* checkpoint using a combination of the 🤗 datasets and 🤗 transformers libraries | [Niels Rogge](https://github.com/nielsrogge) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/TAPAS/Evaluating_TAPAS_on_the_Tabfact_test_set.ipynb)|
-|[Fine-tuning mBART for translation](https://colab.research.google.com/github/vasudevgupta7/huggingface-tutorials/blob/main/translation_training.ipynb) | How to fine-tune mBART using Seq2SeqTrainer for Hindi to English translation | [Vasudev Gupta](https://github.com/vasudevgupta7) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/vasudevgupta7/huggingface-tutorials/blob/main/translation_training.ipynb)|
-|[Fine-tune LayoutLM on FUNSD (a form understanding dataset)](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/LayoutLM/Fine_tuning_LayoutLMForTokenClassification_on_FUNSD.ipynb) | How to fine-tune *LayoutLMForTokenClassification* on the FUNSD dataset for information extraction from scanned documents | [Niels Rogge](https://github.com/nielsrogge) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/LayoutLM/Fine_tuning_LayoutLMForTokenClassification_on_FUNSD.ipynb)|
-|[Fine-Tune DistilGPT2 and Generate Text](https://colab.research.google.com/github/tripathiaakash/DistilGPT2-Tutorial/blob/main/distilgpt2_fine_tuning.ipynb) | How to fine-tune DistilGPT2 and generate text | [Aakash Tripathi](https://github.com/tripathiaakash) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/tripathiaakash/DistilGPT2-Tutorial/blob/main/distilgpt2_fine_tuning.ipynb)|
-|[Fine-Tune LED on up to 8K tokens](https://github.com/patrickvonplaten/notebooks/blob/master/Fine_tune_Longformer_Encoder_Decoder_(LED)_for_Summarization_on_pubmed.ipynb) | How to fine-tune LED on pubmed for long-range summarization | [Patrick von Platen](https://github.com/patrickvonplaten) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/Fine_tune_Longformer_Encoder_Decoder_(LED)_for_Summarization_on_pubmed.ipynb)|
-|[Evaluate LED on Arxiv](https://github.com/patrickvonplaten/notebooks/blob/master/LED_on_Arxiv.ipynb) | How to effectively evaluate LED on long-range summarization | [Patrick von Platen](https://github.com/patrickvonplaten) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/LED_on_Arxiv.ipynb)|
-|[Fine-tune LayoutLM on RVL-CDIP (a document image classification dataset)](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/LayoutLM/Fine_tuning_LayoutLMForSequenceClassification_on_RVL_CDIP.ipynb) | How to fine-tune *LayoutLMForSequenceClassification* on the RVL-CDIP dataset for scanned document classification | [Niels Rogge](https://github.com/nielsrogge) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/LayoutLM/Fine_tuning_LayoutLMForSequenceClassification_on_RVL_CDIP.ipynb)|
-|[Wav2Vec2 CTC decoding with GPT2 adjustment](https://github.com/voidful/huggingface_notebook/blob/main/xlsr_gpt.ipynb) | How to decode CTC sequence with language model adjustment | [Eric Lam](https://github.com/voidful) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1e_z5jQHYbO2YKEaUgzb1ww1WwiAyydAj?usp=sharing)|
-|[Fine-tune BART for summarization in two languages with Trainer class](https://github.com/elsanns/xai-nlp-notebooks/blob/master/fine_tune_bart_summarization_two_langs.ipynb) | How to fine-tune BART for summarization in two languages with Trainer class | [Eliza Szczechla](https://github.com/elsanns) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/elsanns/xai-nlp-notebooks/blob/master/fine_tune_bart_summarization_two_langs.ipynb)|
-|[Evaluate Big Bird on Trivia QA](https://github.com/patrickvonplaten/notebooks/blob/master/Evaluating_Big_Bird_on_TriviaQA.ipynb) | How to evaluate BigBird on long document question answering on Trivia QA | [Patrick von Platen](https://github.com/patrickvonplaten) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/Evaluating_Big_Bird_on_TriviaQA.ipynb)|
-| [Create video captions using Wav2Vec2](https://github.com/Muennighoff/ytclipcc/blob/main/wav2vec_youtube_captions.ipynb) | How to create YouTube captions from any video by transcribing the audio with Wav2Vec | [Niklas Muennighoff](https://github.com/Muennighoff) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/Muennighoff/ytclipcc/blob/main/wav2vec_youtube_captions.ipynb) |
-| [Fine-tune the Vision Transformer on CIFAR-10 using PyTorch Lightning](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/VisionTransformer/Fine_tuning_the_Vision_Transformer_on_CIFAR_10_with_PyTorch_Lightning.ipynb) | How to fine-tune the Vision Transformer (ViT) on CIFAR-10 using HuggingFace Transformers, Datasets and PyTorch Lightning | [Niels Rogge](https://github.com/nielsrogge) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/VisionTransformer/Fine_tuning_the_Vision_Transformer_on_CIFAR_10_with_PyTorch_Lightning.ipynb) |
-| [Fine-tune the Vision Transformer on CIFAR-10 using the 🤗 Trainer](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/VisionTransformer/Fine_tuning_the_Vision_Transformer_on_CIFAR_10_with_the_%F0%9F%A4%97_Trainer.ipynb) | How to fine-tune the Vision Transformer (ViT) on CIFAR-10 using HuggingFace Transformers, Datasets and the 🤗 Trainer | [Niels Rogge](https://github.com/nielsrogge) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/VisionTransformer/Fine_tuning_the_Vision_Transformer_on_CIFAR_10_with_the_%F0%9F%A4%97_Trainer.ipynb) |
-| [Evaluate LUKE on Open Entity, an entity typing dataset](https://github.com/studio-ousia/luke/blob/master/notebooks/huggingface_open_entity.ipynb) | How to evaluate *LukeForEntityClassification* on the Open Entity dataset | [Ikuya Yamada](https://github.com/ikuyamada) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/studio-ousia/luke/blob/master/notebooks/huggingface_open_entity.ipynb) |
-| [Evaluate LUKE on TACRED, a relation extraction dataset](https://github.com/studio-ousia/luke/blob/master/notebooks/huggingface_tacred.ipynb) | How to evaluate *LukeForEntityPairClassification* on the TACRED dataset | [Ikuya Yamada](https://github.com/ikuyamada) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/studio-ousia/luke/blob/master/notebooks/huggingface_tacred.ipynb) |
-| [Evaluate LUKE on CoNLL-2003, an important NER benchmark](https://github.com/studio-ousia/luke/blob/master/notebooks/huggingface_conll_2003.ipynb) | How to evaluate *LukeForEntitySpanClassification* on the CoNLL-2003 dataset | [Ikuya Yamada](https://github.com/ikuyamada) |[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/studio-ousia/luke/blob/master/notebooks/huggingface_conll_2003.ipynb) |
-| [Evaluate BigBird-Pegasus on PubMed dataset](https://github.com/vasudevgupta7/bigbird/blob/main/notebooks/bigbird_pegasus_evaluation.ipynb) | How to evaluate *BigBirdPegasusForConditionalGeneration* on PubMed dataset | [Vasudev Gupta](https://github.com/vasudevgupta7) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/vasudevgupta7/bigbird/blob/main/notebooks/bigbird_pegasus_evaluation.ipynb) |
-| [Speech Emotion Classification with Wav2Vec2](https://github/m3hrdadfi/soxan/blob/main/notebooks/Emotion_recognition_in_Greek_speech_using_Wav2Vec2.ipynb) | How to leverage a pretrained Wav2Vec2 model for Emotion Classification on the MEGA dataset | [Mehrdad Farahani](https://github.com/m3hrdadfi) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/m3hrdadfi/soxan/blob/main/notebooks/Emotion_recognition_in_Greek_speech_using_Wav2Vec2.ipynb) |
-| [Detect objects in an image with DETR](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/DETR/DETR_minimal_example_(with_DetrFeatureExtractor).ipynb) | How to use a trained *DetrForObjectDetection* model to detect objects in an image and visualize attention | [Niels Rogge](https://github.com/NielsRogge) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/DETR/DETR_minimal_example_(with_DetrFeatureExtractor).ipynb) |
-| [Fine-tune DETR on a custom object detection dataset](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/DETR/Fine_tuning_DetrForObjectDetection_on_custom_dataset_(balloon).ipynb) | How to fine-tune *DetrForObjectDetection* on a custom object detection dataset | [Niels Rogge](https://github.com/NielsRogge) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/DETR/Fine_tuning_DetrForObjectDetection_on_custom_dataset_(balloon).ipynb) |
-| [Finetune T5 for Named Entity Recognition](https://github.com/ToluClassics/Notebooks/blob/main/T5_Ner_Finetuning.ipynb) | How to fine-tune *T5* on a Named Entity Recognition Task | [Ogundepo Odunayo](https://github.com/ToluClassics) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1obr78FY_cBmWY5ODViCmzdY6O1KB65Vc?usp=sharing) |

docs/source/en/contributing.md

@@ -1 +0,0 @@
-../../../CONTRIBUTING.md

docs/source/en/converting_tensorflow_models.mdx

@@ -1,162 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Converting Tensorflow Checkpoints
-
-A command-line interface is provided to convert original Bert/GPT/GPT-2/Transformer-XL/XLNet/XLM checkpoints to models
-that can be loaded using the `from_pretrained` methods of the library.
-
-<Tip>
-
-Since 2.3.0 the conversion script is now part of the transformers CLI (**transformers-cli**) available in any
-transformers >= 2.3.0 installation.
-
-The documentation below reflects the **transformers-cli convert** command format.
-
-</Tip>
-
-## BERT
-
-You can convert any TensorFlow checkpoint for BERT (in particular [the pre-trained models released by Google](https://github.com/google-research/bert#pre-trained-models)) in a PyTorch save file by using the
-[convert_bert_original_tf_checkpoint_to_pytorch.py](https://github.com/huggingface/transformers/tree/main/src/transformers/models/bert/convert_bert_original_tf_checkpoint_to_pytorch.py) script.
-
-This CLI takes as input a TensorFlow checkpoint (three files starting with `bert_model.ckpt`) and the associated
-configuration file (`bert_config.json`), and creates a PyTorch model for this configuration, loads the weights from
-the TensorFlow checkpoint in the PyTorch model and saves the resulting model in a standard PyTorch save file that can
-be imported using `from_pretrained()` (see example in [quicktour](quicktour) , [run_glue.py](https://github.com/huggingface/transformers/tree/main/examples/pytorch/text-classification/run_glue.py) ).
-
-You only need to run this conversion script **once** to get a PyTorch model. You can then disregard the TensorFlow
-checkpoint (the three files starting with `bert_model.ckpt`) but be sure to keep the configuration file (\
-`bert_config.json`) and the vocabulary file (`vocab.txt`) as these are needed for the PyTorch model too.
-
-To run this specific conversion script you will need to have TensorFlow and PyTorch installed (`pip install tensorflow`). The rest of the repository only requires PyTorch.
-
-Here is an example of the conversion process for a pre-trained `BERT-Base Uncased` model:
-
-```bash
-export BERT_BASE_DIR=/path/to/bert/uncased_L-12_H-768_A-12
-
-transformers-cli convert --model_type bert \
-  --tf_checkpoint $BERT_BASE_DIR/bert_model.ckpt \
-  --config $BERT_BASE_DIR/bert_config.json \
-  --pytorch_dump_output $BERT_BASE_DIR/pytorch_model.bin
-```
-
-You can download Google's pre-trained models for the conversion [here](https://github.com/google-research/bert#pre-trained-models).
-
-## ALBERT
-
-Convert TensorFlow model checkpoints of ALBERT to PyTorch using the
-[convert_albert_original_tf_checkpoint_to_pytorch.py](https://github.com/huggingface/transformers/tree/main/src/transformers/models/albert/convert_albert_original_tf_checkpoint_to_pytorch.py) script.
-
-The CLI takes as input a TensorFlow checkpoint (three files starting with `model.ckpt-best`) and the accompanying
-configuration file (`albert_config.json`), then creates and saves a PyTorch model. To run this conversion you will
-need to have TensorFlow and PyTorch installed.
-
-Here is an example of the conversion process for the pre-trained `ALBERT Base` model:
-
-```bash
-export ALBERT_BASE_DIR=/path/to/albert/albert_base
-
-transformers-cli convert --model_type albert \
-  --tf_checkpoint $ALBERT_BASE_DIR/model.ckpt-best \
-  --config $ALBERT_BASE_DIR/albert_config.json \
-  --pytorch_dump_output $ALBERT_BASE_DIR/pytorch_model.bin
-```
-
-You can download Google's pre-trained models for the conversion [here](https://github.com/google-research/albert#pre-trained-models).
-
-## OpenAI GPT
-
-Here is an example of the conversion process for a pre-trained OpenAI GPT model, assuming that your NumPy checkpoint
-save as the same format than OpenAI pretrained model (see [here](https://github.com/openai/finetune-transformer-lm)\
-)
-
-```bash
-export OPENAI_GPT_CHECKPOINT_FOLDER_PATH=/path/to/openai/pretrained/numpy/weights
-
-transformers-cli convert --model_type gpt \
-  --tf_checkpoint $OPENAI_GPT_CHECKPOINT_FOLDER_PATH \
-  --pytorch_dump_output $PYTORCH_DUMP_OUTPUT \
-  [--config OPENAI_GPT_CONFIG] \
-  [--finetuning_task_name OPENAI_GPT_FINETUNED_TASK] \
-```
-
-## OpenAI GPT-2
-
-Here is an example of the conversion process for a pre-trained OpenAI GPT-2 model (see [here](https://github.com/openai/gpt-2))
-
-```bash
-export OPENAI_GPT2_CHECKPOINT_PATH=/path/to/gpt2/pretrained/weights
-
-transformers-cli convert --model_type gpt2 \
-  --tf_checkpoint $OPENAI_GPT2_CHECKPOINT_PATH \
-  --pytorch_dump_output $PYTORCH_DUMP_OUTPUT \
-  [--config OPENAI_GPT2_CONFIG] \
-  [--finetuning_task_name OPENAI_GPT2_FINETUNED_TASK]
-```
-
-## Transformer-XL
-
-Here is an example of the conversion process for a pre-trained Transformer-XL model (see [here](https://github.com/kimiyoung/transformer-xl/tree/master/tf#obtain-and-evaluate-pretrained-sota-models))
-
-```bash
-export TRANSFO_XL_CHECKPOINT_FOLDER_PATH=/path/to/transfo/xl/checkpoint
-
-transformers-cli convert --model_type transfo_xl \
-  --tf_checkpoint $TRANSFO_XL_CHECKPOINT_FOLDER_PATH \
-  --pytorch_dump_output $PYTORCH_DUMP_OUTPUT \
-  [--config TRANSFO_XL_CONFIG] \
-  [--finetuning_task_name TRANSFO_XL_FINETUNED_TASK]
-```
-
-## XLNet
-
-Here is an example of the conversion process for a pre-trained XLNet model:
-
-```bash
-export TRANSFO_XL_CHECKPOINT_PATH=/path/to/xlnet/checkpoint
-export TRANSFO_XL_CONFIG_PATH=/path/to/xlnet/config
-
-transformers-cli convert --model_type xlnet \
-  --tf_checkpoint $TRANSFO_XL_CHECKPOINT_PATH \
-  --config $TRANSFO_XL_CONFIG_PATH \
-  --pytorch_dump_output $PYTORCH_DUMP_OUTPUT \
-  [--finetuning_task_name XLNET_FINETUNED_TASK] \
-```
-
-## XLM
-
-Here is an example of the conversion process for a pre-trained XLM model:
-
-```bash
-export XLM_CHECKPOINT_PATH=/path/to/xlm/checkpoint
-
-transformers-cli convert --model_type xlm \
-  --tf_checkpoint $XLM_CHECKPOINT_PATH \
-  --pytorch_dump_output $PYTORCH_DUMP_OUTPUT
- [--config XML_CONFIG] \
- [--finetuning_task_name XML_FINETUNED_TASK]
-```
-
-## T5
-
-Here is an example of the conversion process for a pre-trained T5 model:
-
-```bash
-export T5=/path/to/t5/uncased_L-12_H-768_A-12
-
-transformers-cli convert --model_type t5 \
-  --tf_checkpoint $T5/t5_model.ckpt \
-  --config $T5/t5_config.json \
-  --pytorch_dump_output $T5/pytorch_model.bin
-```

docs/source/en/create_a_model.mdx

@@ -1,355 +0,0 @@
-<!--Copyright 2022 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Create a custom architecture
-
-An [`AutoClass`](model_doc/auto) automatically infers the model architecture and downloads pretrained configuration and weights. Generally, we recommend using an `AutoClass` to produce checkpoint-agnostic code. But users who want more control over specific model parameters can create a custom 🤗 Transformers model from just a few base classes. This could be particularly useful for anyone who is interested in studying, training or experimenting with a 🤗 Transformers model. In this guide, dive deeper into creating a custom model without an `AutoClass`. Learn how to:
-
-- Load and customize a model configuration.
-- Create a model architecture.
-- Create a slow and fast tokenizer for text.
-- Create a feature extractor for audio or image tasks.
-- Create a processor for multimodal tasks.
-
-## Configuration
-
-A [configuration](main_classes/configuration) refers to a model's specific attributes. Each model configuration has different attributes; for instance, all NLP models have the `hidden_size`, `num_attention_heads`, `num_hidden_layers` and `vocab_size` attributes in common. These attributes specify the number of attention heads or hidden layers to construct a model with.
-
-Get a closer look at [DistilBERT](model_doc/distilbert) by accessing [`DistilBertConfig`] to inspect it's attributes:
-
-```py
->>> from transformers import DistilBertConfig
-
->>> config = DistilBertConfig()
->>> print(config)
-DistilBertConfig {
-  "activation": "gelu",
-  "attention_dropout": 0.1,
-  "dim": 768,
-  "dropout": 0.1,
-  "hidden_dim": 3072,
-  "initializer_range": 0.02,
-  "max_position_embeddings": 512,
-  "model_type": "distilbert",
-  "n_heads": 12,
-  "n_layers": 6,
-  "pad_token_id": 0,
-  "qa_dropout": 0.1,
-  "seq_classif_dropout": 0.2,
-  "sinusoidal_pos_embds": false,
-  "transformers_version": "4.16.2",
-  "vocab_size": 30522
-}
-```
-
-[`DistilBertConfig`] displays all the default attributes used to build a base [`DistilBertModel`]. All attributes are customizable, creating space for experimentation. For example, you can customize a default model to:
-
-- Try a different activation function with the `activation` parameter.
-- Use a higher dropout ratio for the attention probabilities with the `attention_dropout` parameter.
-
-```py
->>> my_config = DistilBertConfig(activation="relu", attention_dropout=0.4)
->>> print(my_config)
-DistilBertConfig {
-  "activation": "relu",
-  "attention_dropout": 0.4,
-  "dim": 768,
-  "dropout": 0.1,
-  "hidden_dim": 3072,
-  "initializer_range": 0.02,
-  "max_position_embeddings": 512,
-  "model_type": "distilbert",
-  "n_heads": 12,
-  "n_layers": 6,
-  "pad_token_id": 0,
-  "qa_dropout": 0.1,
-  "seq_classif_dropout": 0.2,
-  "sinusoidal_pos_embds": false,
-  "transformers_version": "4.16.2",
-  "vocab_size": 30522
-}
-```
-
-Pretrained model attributes can be modified in the [`~PretrainedConfig.from_pretrained`] function:
-
-```py
->>> my_config = DistilBertConfig.from_pretrained("distilbert-base-uncased", activation="relu", attention_dropout=0.4)
-```
-
-Once you are satisfied with your model configuration, you can save it with [`~PretrainedConfig.save_pretrained`]. Your configuration file is stored as a JSON file in the specified save directory:
-
-```py
->>> my_config.save_pretrained(save_directory="./your_model_save_path")
-```
-
-To reuse the configuration file, load it with [`~PretrainedConfig.from_pretrained`]:
-
-```py
->>> my_config = DistilBertConfig.from_pretrained("./your_model_save_path/my_config.json")
-```
-
-<Tip>
-
-You can also save your configuration file as a dictionary or even just the difference between your custom configuration attributes and the default configuration attributes! See the [configuration](main_classes/configuration) documentation for more details.
-
-</Tip>
-
-## Model
-
-The next step is to create a [model](main_classes/models). The model - also loosely referred to as the architecture - defines what each layer is doing and what operations are happening. Attributes like `num_hidden_layers` from the configuration are used to define the architecture. Every model shares the base class [`PreTrainedModel`] and a few common methods like resizing input embeddings and pruning self-attention heads. In addition, all models are also either a [`torch.nn.Module`](https://pytorch.org/docs/stable/generated/torch.nn.Module.html), [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) or [`flax.linen.Module`](https://flax.readthedocs.io/en/latest/flax.linen.html#module) subclass. This means models are compatible with each of their respective framework's usage.
-
-<frameworkcontent>
-<pt>
-Load your custom configuration attributes into the model:
-
-```py
->>> from transformers import DistilBertModel
-
->>> my_config = DistilBertConfig.from_pretrained("./your_model_save_path/my_config.json")
->>> model = DistilBertModel(my_config)
-```
-
-This creates a model with random values instead of pretrained weights. You won't be able to use this model for anything useful yet until you train it. Training is a costly and time-consuming process. It is generally better to use a pretrained model to obtain better results faster, while using only a fraction of the resources required for training.
-
-Create a pretrained model with [`~PreTrainedModel.from_pretrained`]:
-
-```py
->>> model = DistilBertModel.from_pretrained("distilbert-base-uncased")
-```
-
-When you load pretrained weights, the default model configuration is automatically loaded if the model is provided by 🤗 Transformers. However, you can still replace - some or all of - the default model configuration attributes with your own if you'd like:
-
-```py
->>> model = DistilBertModel.from_pretrained("distilbert-base-uncased", config=my_config)
-```
-</pt>
-<tf>
-Load your custom configuration attributes into the model:
-
-```py
->>> from transformers import TFDistilBertModel
-
->>> my_config = DistilBertConfig.from_pretrained("./your_model_save_path/my_config.json")
->>> tf_model = TFDistilBertModel(my_config)
-```
-
-This creates a model with random values instead of pretrained weights. You won't be able to use this model for anything useful yet until you train it. Training is a costly and time-consuming process. It is generally better to use a pretrained model to obtain better results faster, while using only a fraction of the resources required for training.
-
-Create a pretrained model with [`~TFPreTrainedModel.from_pretrained`]:
-
-```py
->>> tf_model = TFDistilBertModel.from_pretrained("distilbert-base-uncased")
-```
-
-When you load pretrained weights, the default model configuration is automatically loaded if the model is provided by 🤗 Transformers. However, you can still replace - some or all of - the default model configuration attributes with your own if you'd like:
-
-```py
->>> tf_model = TFDistilBertModel.from_pretrained("distilbert-base-uncased", config=my_config)
-```
-</tf>
-</frameworkcontent>
-
-### Model heads
-
-At this point, you have a base DistilBERT model which outputs the *hidden states*. The hidden states are passed as inputs to a model head to produce the final output. 🤗 Transformers provides a different model head for each task as long as a model supports the task (i.e., you can't use DistilBERT for a sequence-to-sequence task like translation).
-
-<frameworkcontent>
-<pt>
-For example, [`DistilBertForSequenceClassification`] is a base DistilBERT model with a sequence classification head. The sequence classification head is a linear layer on top of the pooled outputs.
-
-```py
->>> from transformers import DistilBertForSequenceClassification
-
->>> model = DistilBertForSequenceClassification.from_pretrained("distilbert-base-uncased")
-```
-
-Easily reuse this checkpoint for another task by switching to a different model head. For a question answering task, you would use the [`DistilBertForQuestionAnswering`] model head. The question answering head is similar to the sequence classification head except it is a linear layer on top of the hidden states output.
-
-```py
->>> from transformers import DistilBertForQuestionAnswering
-
->>> model = DistilBertForQuestionAnswering.from_pretrained("distilbert-base-uncased")
-```
-</pt>
-<tf>
-For example, [`TFDistilBertForSequenceClassification`] is a base DistilBERT model with a sequence classification head. The sequence classification head is a linear layer on top of the pooled outputs.
-
-```py
->>> from transformers import TFDistilBertForSequenceClassification
-
->>> tf_model = TFDistilBertForSequenceClassification.from_pretrained("distilbert-base-uncased")
-```
-
-Easily reuse this checkpoint for another task by switching to a different model head. For a question answering task, you would use the [`TFDistilBertForQuestionAnswering`] model head. The question answering head is similar to the sequence classification head except it is a linear layer on top of the hidden states output.
-
-```py
->>> from transformers import TFDistilBertForQuestionAnswering
-
->>> tf_model = TFDistilBertForQuestionAnswering.from_pretrained("distilbert-base-uncased")
-```
-</tf>
-</frameworkcontent>
-
-## Tokenizer
-
-The last base class you need before using a model for textual data is a [tokenizer](main_classes/tokenizer) to convert raw text to tensors. There are two types of tokenizers you can use with 🤗 Transformers:
-
-- [`PreTrainedTokenizer`]: a Python implementation of a tokenizer.
-- [`PreTrainedTokenizerFast`]: a tokenizer from our Rust-based [🤗 Tokenizer](https://huggingface.co/docs/tokenizers/python/latest/) library. This tokenizer type is significantly faster - especially during batch tokenization - due to it's Rust implementation. The fast tokenizer also offers additional methods like *offset mapping* which maps tokens to their original words or characters.
-
-Both tokenizers support common methods such as encoding and decoding, adding new tokens, and managing special tokens.
-
-<Tip warning={true}>
-
-Not every model supports a fast tokenizer. Take a look at this [table](index#supported-frameworks) to check if a model has fast tokenizer support.
-
-</Tip>
-
-If you trained your own tokenizer, you can create one from your *vocabulary* file:
-
-```py
->>> from transformers import DistilBertTokenizer
-
->>> my_tokenizer = DistilBertTokenizer(vocab_file="my_vocab_file.txt", do_lower_case=False, padding_side="left")
-```
-
-It is important to remember the vocabulary from a custom tokenizer will be different from the vocabulary generated by a pretrained model's tokenizer. You need to use a pretrained model's vocabulary if you are using a pretrained model, otherwise the inputs won't make sense. Create a tokenizer with a pretrained model's vocabulary with the [`DistilBertTokenizer`] class:
-
-```py
->>> from transformers import DistilBertTokenizer
-
->>> slow_tokenizer = DistilBertTokenizer.from_pretrained("distilbert-base-uncased")
-```
-
-Create a fast tokenizer with the [`DistilBertTokenizerFast`] class:
-
-```py
->>> from transformers import DistilBertTokenizerFast
-
->>> fast_tokenizer = DistilBertTokenizerFast.from_pretrained("distilbert-base-uncased")
-```
-
-<Tip>
-
-By default, [`AutoTokenizer`] will try to load a fast tokenizer. You can disable this behavior by setting `use_fast=False` in `from_pretrained`.
-
-</Tip>
-
-## Feature Extractor
-
-A feature extractor processes audio or image inputs. It inherits from the base [`~feature_extraction_utils.FeatureExtractionMixin`] class, and may also inherit from the [`ImageFeatureExtractionMixin`] class for processing image features or the [`SequenceFeatureExtractor`] class for processing audio inputs.
-
-Depending on whether you are working on an audio or vision task, create a feature extractor associated with the model you're using. For example, create a default [`ViTFeatureExtractor`] if you are using [ViT](model_doc/vit) for image classification:
-
-```py
->>> from transformers import ViTFeatureExtractor
-
->>> vit_extractor = ViTFeatureExtractor()
->>> print(vit_extractor)
-ViTFeatureExtractor {
-  "do_normalize": true,
-  "do_resize": true,
-  "feature_extractor_type": "ViTFeatureExtractor",
-  "image_mean": [
-    0.5,
-    0.5,
-    0.5
-  ],
-  "image_std": [
-    0.5,
-    0.5,
-    0.5
-  ],
-  "resample": 2,
-  "size": 224
-}
-```
-
-<Tip>
-
-If you aren't looking for any customization, just use the `from_pretrained` method to load a model's default feature extractor parameters.
-
-</Tip>
-
-Modify any of the [`ViTFeatureExtractor`] parameters to create your custom feature extractor:
-
-```py
->>> from transformers import ViTFeatureExtractor
-
->>> my_vit_extractor = ViTFeatureExtractor(resample="PIL.Image.BOX", do_normalize=False, image_mean=[0.3, 0.3, 0.3])
->>> print(my_vit_extractor)
-ViTFeatureExtractor {
-  "do_normalize": false,
-  "do_resize": true,
-  "feature_extractor_type": "ViTFeatureExtractor",
-  "image_mean": [
-    0.3,
-    0.3,
-    0.3
-  ],
-  "image_std": [
-    0.5,
-    0.5,
-    0.5
-  ],
-  "resample": "PIL.Image.BOX",
-  "size": 224
-}
-```
-
-For audio inputs, you can create a [`Wav2Vec2FeatureExtractor`] and customize the parameters in a similar way:
-
-```py
->>> from transformers import Wav2Vec2FeatureExtractor
-
->>> w2v2_extractor = Wav2Vec2FeatureExtractor()
->>> print(w2v2_extractor)
-Wav2Vec2FeatureExtractor {
-  "do_normalize": true,
-  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
-  "feature_size": 1,
-  "padding_side": "right",
-  "padding_value": 0.0,
-  "return_attention_mask": false,
-  "sampling_rate": 16000
-}
-```
-
-## Processor
-
-For models that support multimodal tasks, 🤗 Transformers offers a processor class that conveniently wraps a feature extractor and tokenizer into a single object. For example, let's use the [`Wav2Vec2Processor`] for an automatic speech recognition task (ASR). ASR transcribes audio to text, so you will need a feature extractor and a tokenizer.
-
-Create a feature extractor to handle the audio inputs:
-
-```py
->>> from transformers import Wav2Vec2FeatureExtractor
-
->>> feature_extractor = Wav2Vec2FeatureExtractor(padding_value=1.0, do_normalize=True)
-```
-
-Create a tokenizer to handle the text inputs:
-
-```py
->>> from transformers import Wav2Vec2CTCTokenizer
-
->>> tokenizer = Wav2Vec2CTCTokenizer(vocab_file="my_vocab_file.txt")
-```
-
-Combine the feature extractor and tokenizer in [`Wav2Vec2Processor`]:
-
-```py
->>> from transformers import Wav2Vec2Processor
-
->>> processor = Wav2Vec2Processor(feature_extractor=feature_extractor, tokenizer=tokenizer)
-```
-
-With two basic classes - configuration and model - and an additional preprocessing class (tokenizer, feature extractor, or processor), you can create any of the models supported by 🤗 Transformers. Each of these base classes are configurable, allowing you to use the specific attributes you want. You can easily setup a model for training or modify an existing pretrained model to fine-tune.

docs/source/en/custom_models.mdx

@@ -1,349 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Sharing custom models
-
-The 🤗 Transformers library is designed to be easily extensible. Every model is fully coded in a given subfolder
-of the repository with no abstraction, so you can easily copy a modeling file and tweak it to your needs.
-
-If you are writing a brand new model, it might be easier to start from scratch. In this tutorial, we will show you
-how to write a custom model and its configuration so it can be used inside Transformers, and how you can share it
-with the community (with the code it relies on) so that anyone can use it, even if it's not present in the 🤗
-Transformers library.
-
-We will illustrate all of this on a ResNet model, by wrapping the ResNet class of the
-[timm library](https://github.com/rwightman/pytorch-image-models/tree/master/timm) into a [`PreTrainedModel`].
-
-## Writing a custom configuration
-
-Before we dive into the model, let's first write its configuration. The configuration of a model is an object that
-will contain all the necessary information to build the model. As we will see in the next section, the model can only
-take a `config` to be initialized, so we really need that object to be as complete as possible.
-
-In our example, we will take a couple of arguments of the ResNet class that we might want to tweak. Different
-configurations will then give us the different types of ResNets that are possible. We then just store those arguments,
-after checking the validity of a few of them.
-
-```python
-from transformers import PretrainedConfig
-from typing import List
-
-
-class ResnetConfig(PretrainedConfig):
-    model_type = "resnet"
-
-    def __init__(
-        self,
-        block_type="bottleneck",
-        layers: List[int] = [3, 4, 6, 3],
-        num_classes: int = 1000,
-        input_channels: int = 3,
-        cardinality: int = 1,
-        base_width: int = 64,
-        stem_width: int = 64,
-        stem_type: str = "",
-        avg_down: bool = False,
-        **kwargs,
-    ):
-        if block_type not in ["basic", "bottleneck"]:
-            raise ValueError(f"`block` must be 'basic' or bottleneck', got {block}.")
-        if stem_type not in ["", "deep", "deep-tiered"]:
-            raise ValueError(f"`stem_type` must be '', 'deep' or 'deep-tiered', got {block}.")
-
-        self.block_type = block_type
-        self.layers = layers
-        self.num_classes = num_classes
-        self.input_channels = input_channels
-        self.cardinality = cardinality
-        self.base_width = base_width
-        self.stem_width = stem_width
-        self.stem_type = stem_type
-        self.avg_down = avg_down
-        super().__init__(**kwargs)
-```
-
-The three important things to remember when writing you own configuration are the following:
-- you have to inherit from `PretrainedConfig`,
-- the `__init__` of your `PretrainedConfig` must accept any kwargs,
-- those `kwargs` need to be passed to the superclass `__init__`.
-
-The inheritance is to make sure you get all the functionality from the 🤗 Transformers library, while the two other
-constraints come from the fact a `PretrainedConfig` has more fields than the ones you are setting. When reloading a
-config with the `from_pretrained` method, those fields need to be accepted by your config and then sent to the
-superclass.
-
-Defining a `model_type` for your configuration (here `model_type="resnet"`) is not mandatory, unless you want to
-register your model with the auto classes (see last section).
-
-With this done, you can easily create and save your configuration like you would do with any other model config of the
-library. Here is how we can create a resnet50d config and save it:
-
-```py
-resnet50d_config = ResnetConfig(block_type="bottleneck", stem_width=32, stem_type="deep", avg_down=True)
-resnet50d_config.save_pretrained("custom-resnet")
-```
-
-This will save a file named `config.json` inside the folder `custom-resnet`. You can then reload your config with the
-`from_pretrained` method:
-
-```py
-resnet50d_config = ResnetConfig.from_pretrained("custom-resnet")
-```
-
-You can also use any other method of the [`PretrainedConfig`] class, like [`~PretrainedConfig.push_to_hub`] to
-directly upload your config to the Hub.
-
-## Writing a custom model
-
-Now that we have our ResNet configuration, we can go on writing the model. We will actually write two: one that
-extracts the hidden features from a batch of images (like [`BertModel`]) and one that is suitable for image
-classification (like [`BertForSequenceClassification`]).
-
-As we mentioned before, we'll only write a loose wrapper of the model to keep it simple for this example. The only
-thing we need to do before writing this class is a map between the block types and actual block classes. Then the
-model is defined from the configuration by passing everything to the `ResNet` class:
-
-```py
-from transformers import PreTrainedModel
-from timm.models.resnet import BasicBlock, Bottleneck, ResNet
-from .configuration_resnet import ResnetConfig
-
-
-BLOCK_MAPPING = {"basic": BasicBlock, "bottleneck": Bottleneck}
-
-
-class ResnetModel(PreTrainedModel):
-    config_class = ResnetConfig
-
-    def __init__(self, config):
-        super().__init__(config)
-        block_layer = BLOCK_MAPPING[config.block_type]
-        self.model = ResNet(
-            block_layer,
-            config.layers,
-            num_classes=config.num_classes,
-            in_chans=config.input_channels,
-            cardinality=config.cardinality,
-            base_width=config.base_width,
-            stem_width=config.stem_width,
-            stem_type=config.stem_type,
-            avg_down=config.avg_down,
-        )
-
-    def forward(self, tensor):
-        return self.model.forward_features(tensor)
-```
-
-For the model that will classify images, we just change the forward method:
-
-```py
-class ResnetModelForImageClassification(PreTrainedModel):
-    config_class = ResnetConfig
-
-    def __init__(self, config):
-        super().__init__(config)
-        block_layer = BLOCK_MAPPING[config.block_type]
-        self.model = ResNet(
-            block_layer,
-            config.layers,
-            num_classes=config.num_classes,
-            in_chans=config.input_channels,
-            cardinality=config.cardinality,
-            base_width=config.base_width,
-            stem_width=config.stem_width,
-            stem_type=config.stem_type,
-            avg_down=config.avg_down,
-        )
-
-    def forward(self, tensor, labels=None):
-        logits = self.model(tensor)
-        if labels is not None:
-            loss = torch.nn.cross_entropy(logits, labels)
-            return {"loss": loss, "logits": logits}
-        return {"logits": logits}
-```
-
-In both cases, notice how we inherit from `PreTrainedModel` and call the superclass initialization with the `config`
-(a bit like when you write a regular `torch.nn.Module`). The line that sets the `config_class` is not mandatory, unless
-you want to register your model with the auto classes (see last section).
-
-<Tip>
-
-If your model is very similar to a model inside the library, you can re-use the same configuration as this model.
-
-</Tip>
-
-You can have your model return anything you want, but returning a dictionary like we did for
-`ResnetModelForImageClassification`, with the loss included when labels are passed, will make your model directly
-usable inside the [`Trainer`] class. Using another output format is fine as long as you are planning on using your own
-training loop or another library for training.
-
-Now that we have our model class, let's create one:
-
-```py
-resnet50d = ResnetModelForImageClassification(resnet50d_config)
-```
-
-Again, you can use any of the methods of [`PreTrainedModel`], like [`~PreTrainedModel.save_pretrained`] or
-[`~PreTrainedModel.push_to_hub`]. We will use the second in the next section, and see how to push the model weights
-with the code of our model. But first, let's load some pretrained weights inside our model.
-
-In your own use case, you will probably be training your custom model on your own data. To go fast for this tutorial,
-we will use the pretrained version of the resnet50d. Since our model is just a wrapper around it, it's going to be
-easy to transfer those weights:
-
-```py
-import timm
-
-pretrained_model = timm.create_model("resnet50d", pretrained=True)
-resnet50d.model.load_state_dict(pretrained_model.state_dict())
-```
-
-Now let's see how to make sure that when we do [`~PreTrainedModel.save_pretrained`] or [`~PreTrainedModel.push_to_hub`], the
-code of the model is saved.
-
-## Sending the code to the Hub
-
-<Tip warning={true}>
-
-This API is experimental and may have some slight breaking changes in the next releases.
-
-</Tip>
-
-First, make sure your model is fully defined in a `.py` file. It can rely on relative imports to some other files as
-long as all the files are in the same directory (we don't support submodules for this feature yet). For our example,
-we'll define a `modeling_resnet.py` file and a `configuration_resnet.py` file in a folder of the current working
-directory named `resnet_model`. The configuration file contains the code for `ResnetConfig` and the modeling file
-contains the code of `ResnetModel` and `ResnetModelForImageClassification`.
-
-```
-.
-└── resnet_model
-    ├── __init__.py
-    ├── configuration_resnet.py
-    └── modeling_resnet.py
-```
-
-The `__init__.py` can be empty, it's just there so that Python detects `resnet_model` can be use as a module.
-
-<Tip warning={true}>
-
-If copying a modeling files from the library, you will need to replace all the relative imports at the top of the file
-to import from the `transformers` package.
-
-</Tip>
-
-Note that you can re-use (or subclass) an existing configuration/model.
-
-To share your model with the community, follow those steps: first import the ResNet model and config from the newly
-created files:
-
-```py
-from resnet_model.configuration_resnet import ResnetConfig
-from resnet_model.modeling_resnet import ResnetModel, ResnetModelForImageClassification
-```
-
-Then you have to tell the library you want to copy the code files of those objects when using the `save_pretrained`
-method and properly register them with a given Auto class (especially for models), just run:
-
-```py
-ResnetConfig.register_for_auto_class()
-ResnetModel.register_for_auto_class("AutoModel")
-ResnetModelForImageClassification.register_for_auto_class("AutoModelForImageClassification")
-```
-
-Note that there is no need to specify an auto class for the configuration (there is only one auto class for them,
-[`AutoConfig`]) but it's different for models. Your custom model could be suitable for many different tasks, so you
-have to specify which one of the auto classes is the correct one for your model.
-
-Next, let's create the config and models as we did before:
-
-```py
-resnet50d_config = ResnetConfig(block_type="bottleneck", stem_width=32, stem_type="deep", avg_down=True)
-resnet50d = ResnetModelForImageClassification(resnet50d_config)
-
-pretrained_model = timm.create_model("resnet50d", pretrained=True)
-resnet50d.model.load_state_dict(pretrained_model.state_dict())
-```
-
-Now to send the model to the Hub, make sure you are logged in. Either run in your terminal:
-
-```bash
-huggingface-cli login
-```
-
-or from a notebook:
-
-```py
-from huggingface_hub import notebook_login
-
-notebook_login()
-```
-
-You can then push to your own namespace (or an organization you are a member of) like this:
-
-```py
-resnet50d.push_to_hub("custom-resnet50d")
-```
-
-On top of the modeling weights and the configuration in json format, this also copied the modeling and
-configuration `.py` files in the folder `custom-resnet50d` and uploaded the result to the Hub. You can check the result
-in this [model repo](https://huggingface.co/sgugger/custom-resnet50d).
-
-See the [sharing tutorial](model_sharing) for more information on the push to Hub method.
-
-## Using a model with custom code
-
-You can use any configuration, model or tokenizer with custom code files in its repository with the auto-classes and
-the `from_pretrained` method. All files and code uploaded to the Hub are scanned for malware (refer to the [Hub security](https://huggingface.co/docs/hub/security#malware-scanning) documentation for more information), but you should still 
-review the model code and author to avoid executing malicious code on your machine. Set `trust_remote_code=True` to use
-a model with custom code:
-
-```py
-from transformers import AutoModelForImageClassification
-
-model = AutoModelForImageClassification.from_pretrained("sgugger/custom-resnet50d", trust_remote_code=True)
-```
-
-It is also strongly encouraged to pass a commit hash as a `revision` to make sure the author of the models did not
-update the code with some malicious new lines (unless you fully trust the authors of the models).
-
-```py
-commit_hash = "ed94a7c6247d8aedce4647f00f20de6875b5b292"
-model = AutoModelForImageClassification.from_pretrained(
-    "sgugger/custom-resnet50d", trust_remote_code=True, revision=commit_hash
-)
-```
-
-Note that when browsing the commit history of the model repo on the Hub, there is a button to easily copy the commit
-hash of any commit.
-
-## Registering a model with custom code to the auto classes
-
-If you are writing a library that extends 🤗 Transformers, you may want to extend the auto classes to include your own
-model. This is different from pushing the code to the Hub in the sense that users will need to import your library to
-get the custom models (contrarily to automatically downloading the model code from the Hub).
-
-As long as your config has a `model_type` attribute that is different from existing model types, and that your model
-classes have the right `config_class` attributes, you can just add them to the auto classes likes this:
-
-```py
-from transformers import AutoConfig, AutoModel, AutoModelForImageClassification
-
-AutoConfig.register("resnet", ResnetConfig)
-AutoModel.register(ResnetConfig, ResnetModel)
-AutoModelForImageClassification.register(ResnetConfig, ResnetModelForImageClassification)
-```
-
-Note that the first argument used when registering your custom config to [`AutoConfig`] needs to match the `model_type`
-of your custom config, and the first argument used when registering your custom models to any auto model class needs
-to match the `config_class` of those models.

docs/source/en/debugging.mdx

@@ -1,335 +0,0 @@
-<!--Copyright 2021 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Debugging
-
-## Multi-GPU Network Issues Debug
-
-When training or inferencing with `DistributedDataParallel` and multiple GPU, if you run into issue of inter-communication between processes and/or nodes, you can use the following script to diagnose network issues.
-
-```bash
-wget https://raw.githubusercontent.com/huggingface/transformers/main/scripts/distributed/torch-distributed-gpu-test.py
-```
-
-For example to test how 2 GPUs interact do:
-
-```bash
-python -m torch.distributed.run --nproc_per_node 2 --nnodes 1 torch-distributed-gpu-test.py
-```
-If both processes can talk to each and allocate GPU memory each will print an OK status.
-
-For more GPUs or nodes adjust the arguments in the script.
-
-You will find a lot more details inside the diagnostics script and even a recipe to how you could run it in a SLURM environment.
-
-An additional level of debug is to add `NCCL_DEBUG=INFO` environment variable as follows:
-
-```bash
-NCCL_DEBUG=INFO python -m torch.distributed.run --nproc_per_node 2 --nnodes 1 torch-distributed-gpu-test.py
-```
-
-This will dump a lot of NCCL-related debug information, which you can then search online if you find that some problems are reported. Or if you're not sure how to interpret the output you can share the log file in an Issue.
-
-
-
-## Underflow and Overflow Detection
-
-<Tip>
-
-This feature is currently available for PyTorch-only.
-
-</Tip>
-
-<Tip>
-
-For multi-GPU training it requires DDP (`torch.distributed.launch`).
-
-</Tip>
-
-<Tip>
-
-This feature can be used with any `nn.Module`-based model.
-
-</Tip>
-
-If you start getting `loss=NaN` or the model inhibits some other abnormal behavior due to `inf` or `nan` in
-activations or weights one needs to discover where the first underflow or overflow happens and what led to it. Luckily
-you can accomplish that easily by activating a special module that will do the detection automatically.
-
-If you're using [`Trainer`], you just need to add:
-
-```bash
---debug underflow_overflow
-```
-
-to the normal command line arguments, or pass `debug="underflow_overflow"` when creating the
-[`TrainingArguments`] object.
-
-If you're using your own training loop or another Trainer you can accomplish the same with:
-
-```python
-from .debug_utils import DebugUnderflowOverflow
-
-debug_overflow = DebugUnderflowOverflow(model)
-```
-
-[`~debug_utils.DebugUnderflowOverflow`] inserts hooks into the model that immediately after each
-forward call will test input and output variables and also the corresponding module's weights. As soon as `inf` or
-`nan` is detected in at least one element of the activations or weights, the program will assert and print a report
-like this (this was caught with `google/mt5-small` under fp16 mixed precision):
-
-```
-Detected inf/nan during batch_number=0
-Last 21 forward frames:
-abs min  abs max  metadata
-                  encoder.block.1.layer.1.DenseReluDense.dropout Dropout
-0.00e+00 2.57e+02 input[0]
-0.00e+00 2.85e+02 output
-[...]
-                  encoder.block.2.layer.0 T5LayerSelfAttention
-6.78e-04 3.15e+03 input[0]
-2.65e-04 3.42e+03 output[0]
-             None output[1]
-2.25e-01 1.00e+04 output[2]
-                  encoder.block.2.layer.1.layer_norm T5LayerNorm
-8.69e-02 4.18e-01 weight
-2.65e-04 3.42e+03 input[0]
-1.79e-06 4.65e+00 output
-                  encoder.block.2.layer.1.DenseReluDense.wi_0 Linear
-2.17e-07 4.50e+00 weight
-1.79e-06 4.65e+00 input[0]
-2.68e-06 3.70e+01 output
-                  encoder.block.2.layer.1.DenseReluDense.wi_1 Linear
-8.08e-07 2.66e+01 weight
-1.79e-06 4.65e+00 input[0]
-1.27e-04 2.37e+02 output
-                  encoder.block.2.layer.1.DenseReluDense.dropout Dropout
-0.00e+00 8.76e+03 input[0]
-0.00e+00 9.74e+03 output
-                  encoder.block.2.layer.1.DenseReluDense.wo Linear
-1.01e-06 6.44e+00 weight
-0.00e+00 9.74e+03 input[0]
-3.18e-04 6.27e+04 output
-                  encoder.block.2.layer.1.DenseReluDense T5DenseGatedGeluDense
-1.79e-06 4.65e+00 input[0]
-3.18e-04 6.27e+04 output
-                  encoder.block.2.layer.1.dropout Dropout
-3.18e-04 6.27e+04 input[0]
-0.00e+00      inf output
-```
-
-The example output has been trimmed in the middle for brevity.
-
-The second column shows the value of the absolute largest element, so if you have a closer look at the last few frames,
-the inputs and outputs were in the range of `1e4`. So when this training was done under fp16 mixed precision the very
-last step overflowed (since under `fp16` the largest number before `inf` is `64e3`). To avoid overflows under
-`fp16` the activations must remain way below `1e4`, because `1e4 * 1e4 = 1e8` so any matrix multiplication with
-large activations is going to lead to a numerical overflow condition.
-
-At the very start of the trace you can discover at which batch number the problem occurred (here `Detected inf/nan during batch_number=0` means the problem occurred on the first batch).
-
-Each reported frame starts by declaring the fully qualified entry for the corresponding module this frame is reporting
-for. If we look just at this frame:
-
-```
-                  encoder.block.2.layer.1.layer_norm T5LayerNorm
-8.69e-02 4.18e-01 weight
-2.65e-04 3.42e+03 input[0]
-1.79e-06 4.65e+00 output
-```
-
-Here, `encoder.block.2.layer.1.layer_norm` indicates that it was a layer norm for the first layer, of the second
-block of the encoder. And the specific calls of the `forward` is `T5LayerNorm`.
-
-Let's look at the last few frames of that report:
-
-```
-Detected inf/nan during batch_number=0
-Last 21 forward frames:
-abs min  abs max  metadata
-[...]
-                  encoder.block.2.layer.1.DenseReluDense.wi_0 Linear
-2.17e-07 4.50e+00 weight
-1.79e-06 4.65e+00 input[0]
-2.68e-06 3.70e+01 output
-                  encoder.block.2.layer.1.DenseReluDense.wi_1 Linear
-8.08e-07 2.66e+01 weight
-1.79e-06 4.65e+00 input[0]
-1.27e-04 2.37e+02 output
-                  encoder.block.2.layer.1.DenseReluDense.wo Linear
-1.01e-06 6.44e+00 weight
-0.00e+00 9.74e+03 input[0]
-3.18e-04 6.27e+04 output
-                  encoder.block.2.layer.1.DenseReluDense T5DenseGatedGeluDense
-1.79e-06 4.65e+00 input[0]
-3.18e-04 6.27e+04 output
-                  encoder.block.2.layer.1.dropout Dropout
-3.18e-04 6.27e+04 input[0]
-0.00e+00      inf output
-```
-
-The last frame reports for `Dropout.forward` function with the first entry for the only input and the second for the
-only output. You can see that it was called from an attribute `dropout` inside `DenseReluDense` class. We can see
-that it happened during the first layer, of the 2nd block, during the very first batch. Finally, the absolute largest
-input elements was `6.27e+04` and same for the output was `inf`.
-
-You can see here, that `T5DenseGatedGeluDense.forward` resulted in output activations, whose absolute max value was
-around 62.7K, which is very close to fp16's top limit of 64K. In the next frame we have `Dropout` which renormalizes
-the weights, after it zeroed some of the elements, which pushes the absolute max value to more than 64K, and we get an
-overflow (`inf`).
-
-As you can see it's the previous frames that we need to look into when the numbers start going into very large for fp16
-numbers.
-
-Let's match the report to the code from `models/t5/modeling_t5.py`:
-
-```python
-class T5DenseGatedGeluDense(nn.Module):
-    def __init__(self, config):
-        super().__init__()
-        self.wi_0 = nn.Linear(config.d_model, config.d_ff, bias=False)
-        self.wi_1 = nn.Linear(config.d_model, config.d_ff, bias=False)
-        self.wo = nn.Linear(config.d_ff, config.d_model, bias=False)
-        self.dropout = nn.Dropout(config.dropout_rate)
-        self.gelu_act = ACT2FN["gelu_new"]
-
-    def forward(self, hidden_states):
-        hidden_gelu = self.gelu_act(self.wi_0(hidden_states))
-        hidden_linear = self.wi_1(hidden_states)
-        hidden_states = hidden_gelu * hidden_linear
-        hidden_states = self.dropout(hidden_states)
-        hidden_states = self.wo(hidden_states)
-        return hidden_states
-```
-
-Now it's easy to see the `dropout` call, and all the previous calls as well.
-
-Since the detection is happening in a forward hook, these reports are printed immediately after each `forward`
-returns.
-
-Going back to the full report, to act on it and to fix the problem, we need to go a few frames up where the numbers
-started to go up and most likely switch to the `fp32` mode here, so that the numbers don't overflow when multiplied
-or summed up. Of course, there might be other solutions. For example, we could turn off `amp` temporarily if it's
-enabled, after moving the original `forward` into a helper wrapper, like so:
-
-```python
-def _forward(self, hidden_states):
-    hidden_gelu = self.gelu_act(self.wi_0(hidden_states))
-    hidden_linear = self.wi_1(hidden_states)
-    hidden_states = hidden_gelu * hidden_linear
-    hidden_states = self.dropout(hidden_states)
-    hidden_states = self.wo(hidden_states)
-    return hidden_states
-
-
-import torch
-
-
-def forward(self, hidden_states):
-    if torch.is_autocast_enabled():
-        with torch.cuda.amp.autocast(enabled=False):
-            return self._forward(hidden_states)
-    else:
-        return self._forward(hidden_states)
-```
-
-Since the automatic detector only reports on inputs and outputs of full frames, once you know where to look, you may
-want to analyse the intermediary stages of any specific `forward` function as well. In such a case you can use the
-`detect_overflow` helper function to inject the detector where you want it, for example:
-
-```python
-from debug_utils import detect_overflow
-
-
-class T5LayerFF(nn.Module):
-    [...]
-
-    def forward(self, hidden_states):
-        forwarded_states = self.layer_norm(hidden_states)
-        detect_overflow(forwarded_states, "after layer_norm")
-        forwarded_states = self.DenseReluDense(forwarded_states)
-        detect_overflow(forwarded_states, "after DenseReluDense")
-        return hidden_states + self.dropout(forwarded_states)
-```
-
-You can see that we added 2 of these and now we track if `inf` or `nan` for `forwarded_states` was detected
-somewhere in between.
-
-Actually, the detector already reports these because each of the calls in the example above is a `nn.Module`, but
-let's say if you had some local direct calculations this is how you'd do that.
-
-Additionally, if you're instantiating the debugger in your own code, you can adjust the number of frames printed from
-its default, e.g.:
-
-```python
-from .debug_utils import DebugUnderflowOverflow
-
-debug_overflow = DebugUnderflowOverflow(model, max_frames_to_save=100)
-```
-
-### Specific batch absolute mix and max value tracing
-
-The same debugging class can be used for per-batch tracing with the underflow/overflow detection feature turned off.
-
-Let's say you want to watch the absolute min and max values for all the ingredients of each `forward` call of a given
-batch, and only do that for batches 1 and 3. Then you instantiate this class as:
-
-```python
-debug_overflow = DebugUnderflowOverflow(model, trace_batch_nums=[1, 3])
-```
-
-And now full batches 1 and 3 will be traced using the same format as the underflow/overflow detector does.
-
-Batches are 0-indexed.
-
-This is helpful if you know that the program starts misbehaving after a certain batch number, so you can fast-forward
-right to that area. Here is a sample truncated output for such configuration:
-
-```
-                  *** Starting batch number=1 ***
-abs min  abs max  metadata
-                  shared Embedding
-1.01e-06 7.92e+02 weight
-0.00e+00 2.47e+04 input[0]
-5.36e-05 7.92e+02 output
-[...]
-                  decoder.dropout Dropout
-1.60e-07 2.27e+01 input[0]
-0.00e+00 2.52e+01 output
-                  decoder T5Stack
-     not a tensor output
-                  lm_head Linear
-1.01e-06 7.92e+02 weight
-0.00e+00 1.11e+00 input[0]
-6.06e-02 8.39e+01 output
-                   T5ForConditionalGeneration
-     not a tensor output
-
-                  *** Starting batch number=3 ***
-abs min  abs max  metadata
-                  shared Embedding
-1.01e-06 7.92e+02 weight
-0.00e+00 2.78e+04 input[0]
-5.36e-05 7.92e+02 output
-[...]
-```
-
-Here you will get a huge number of frames dumped - as many as there were forward calls in your model, so it may or may
-not what you want, but sometimes it can be easier to use for debugging purposes than a normal debugger. For example, if
-a problem starts happening at batch number 150. So you can dump traces for batches 149 and 150 and compare where
-numbers started to diverge.
-
-You can also specify the batch number after which to stop the training, with:
-
-```python
-debug_overflow = DebugUnderflowOverflow(model, trace_batch_nums=[1, 3], abort_after_batch_num=3)
-```

docs/source/en/fast_tokenizers.mdx

@@ -1,70 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Use tokenizers from 🤗 Tokenizers
-
-The [`PreTrainedTokenizerFast`] depends on the [🤗 Tokenizers](https://huggingface.co/docs/tokenizers) library. The tokenizers obtained from the 🤗 Tokenizers library can be
-loaded very simply into 🤗 Transformers.
-
-Before getting in the specifics, let's first start by creating a dummy tokenizer in a few lines:
-
-```python
->>> from tokenizers import Tokenizer
->>> from tokenizers.models import BPE
->>> from tokenizers.trainers import BpeTrainer
->>> from tokenizers.pre_tokenizers import Whitespace
-
->>> tokenizer = Tokenizer(BPE(unk_token="[UNK]"))
->>> trainer = BpeTrainer(special_tokens=["[UNK]", "[CLS]", "[SEP]", "[PAD]", "[MASK]"])
-
->>> tokenizer.pre_tokenizer = Whitespace()
->>> files = [...]
->>> tokenizer.train(files, trainer)
-```
-
-We now have a tokenizer trained on the files we defined. We can either continue using it in that runtime, or save it to
-a JSON file for future re-use.
-
-## Loading directly from the tokenizer object
-
-Let's see how to leverage this tokenizer object in the 🤗 Transformers library. The
-[`PreTrainedTokenizerFast`] class allows for easy instantiation, by accepting the instantiated
-*tokenizer* object as an argument:
-
-```python
->>> from transformers import PreTrainedTokenizerFast
-
->>> fast_tokenizer = PreTrainedTokenizerFast(tokenizer_object=tokenizer)
-```
-
-This object can now be used with all the methods shared by the 🤗 Transformers tokenizers! Head to [the tokenizer
-page](main_classes/tokenizer) for more information.
-
-## Loading from a JSON file
-
-In order to load a tokenizer from a JSON file, let's first start by saving our tokenizer:
-
-```python
->>> tokenizer.save("tokenizer.json")
-```
-
-The path to which we saved this file can be passed to the [`PreTrainedTokenizerFast`] initialization
-method using the `tokenizer_file` parameter:
-
-```python
->>> from transformers import PreTrainedTokenizerFast
-
->>> fast_tokenizer = PreTrainedTokenizerFast(tokenizer_file="tokenizer.json")
-```
-
-This object can now be used with all the methods shared by the 🤗 Transformers tokenizers! Head to [the tokenizer
-page](main_classes/tokenizer) for more information.

docs/source/en/glossary.mdx

@@ -1,300 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Glossary
-
-## General terms
-
-- autoencoding models: see MLM
-- autoregressive models: see CLM
-- CLM: causal language modeling, a pretraining task where the model reads the texts in order and has to predict the
-  next word. It's usually done by reading the whole sentence but using a mask inside the model to hide the future
-  tokens at a certain timestep.
-- deep learning: machine learning algorithms which uses neural networks with several layers.
-- MLM: masked language modeling, a pretraining task where the model sees a corrupted version of the texts, usually done
-  by masking some tokens randomly, and has to predict the original text.
-- multimodal: a task that combines texts with another kind of inputs (for instance images).
-- NLG: natural language generation, all tasks related to generating text (for instance talk with transformers,
-  translation).
-- NLP: natural language processing, a generic way to say "deal with texts".
-- NLU: natural language understanding, all tasks related to understanding what is in a text (for instance classifying
-  the whole text, individual words).
-- pretrained model: a model that has been pretrained on some data (for instance all of Wikipedia). Pretraining methods
-  involve a self-supervised objective, which can be reading the text and trying to predict the next word (see CLM) or
-  masking some words and trying to predict them (see MLM).
-- RNN: recurrent neural network, a type of model that uses a loop over a layer to process texts.
-- self-attention: each element of the input finds out which other elements of the input they should attend to.
-- seq2seq or sequence-to-sequence: models that generate a new sequence from an input, like translation models, or
-  summarization models (such as [Bart](model_doc/bart) or [T5](model_doc/t5)).
-- token: a part of a sentence, usually a word, but can also be a subword (non-common words are often split in subwords)
-  or a punctuation symbol.
-- transformer: self-attention based deep learning model architecture.
-
-## Model inputs
-
-Every model is different yet bears similarities with the others. Therefore most models use the same inputs, which are
-detailed here alongside usage examples.
-
-<a id='input-ids'></a>
-
-### Input IDs
-
-The input ids are often the only required parameters to be passed to the model as input. *They are token indices,
-numerical representations of tokens building the sequences that will be used as input by the model*.
-
-<Youtube id="VFp38yj8h3A"/>
-
-Each tokenizer works differently but the underlying mechanism remains the same. Here's an example using the BERT
-tokenizer, which is a [WordPiece](https://arxiv.org/pdf/1609.08144.pdf) tokenizer:
-
-```python
->>> from transformers import BertTokenizer
-
->>> tokenizer = BertTokenizer.from_pretrained("bert-base-cased")
-
->>> sequence = "A Titan RTX has 24GB of VRAM"
-```
-
-The tokenizer takes care of splitting the sequence into tokens available in the tokenizer vocabulary.
-
-```python
->>> tokenized_sequence = tokenizer.tokenize(sequence)
-```
-
-The tokens are either words or subwords. Here for instance, "VRAM" wasn't in the model vocabulary, so it's been split
-in "V", "RA" and "M". To indicate those tokens are not separate words but parts of the same word, a double-hash prefix
-is added for "RA" and "M":
-
-```python
->>> print(tokenized_sequence)
-['A', 'Titan', 'R', '##T', '##X', 'has', '24', '##GB', 'of', 'V', '##RA', '##M']
-```
-
-These tokens can then be converted into IDs which are understandable by the model. This can be done by directly feeding
-the sentence to the tokenizer, which leverages the Rust implementation of [🤗 Tokenizers](https://github.com/huggingface/tokenizers) for peak performance.
-
-```python
->>> inputs = tokenizer(sequence)
-```
-
-The tokenizer returns a dictionary with all the arguments necessary for its corresponding model to work properly. The
-token indices are under the key "input_ids":
-
-```python
->>> encoded_sequence = inputs["input_ids"]
->>> print(encoded_sequence)
-[101, 138, 18696, 155, 1942, 3190, 1144, 1572, 13745, 1104, 159, 9664, 2107, 102]
-```
-
-Note that the tokenizer automatically adds "special tokens" (if the associated model relies on them) which are special
-IDs the model sometimes uses.
-
-If we decode the previous sequence of ids,
-
-```python
->>> decoded_sequence = tokenizer.decode(encoded_sequence)
-```
-
-we will see
-
-```python
->>> print(decoded_sequence)
-[CLS] A Titan RTX has 24GB of VRAM [SEP]
-```
-
-because this is the way a [`BertModel`] is going to expect its inputs.
-
-<a id='attention-mask'></a>
-
-### Attention mask
-
-The attention mask is an optional argument used when batching sequences together.
-
-<Youtube id="M6adb1j2jPI"/>
-
-This argument indicates to the model which tokens should be attended to, and which should not.
-
-For example, consider these two sequences:
-
-```python
->>> from transformers import BertTokenizer
-
->>> tokenizer = BertTokenizer.from_pretrained("bert-base-cased")
-
->>> sequence_a = "This is a short sequence."
->>> sequence_b = "This is a rather long sequence. It is at least longer than the sequence A."
-
->>> encoded_sequence_a = tokenizer(sequence_a)["input_ids"]
->>> encoded_sequence_b = tokenizer(sequence_b)["input_ids"]
-```
-
-The encoded versions have different lengths:
-
-```python
->>> len(encoded_sequence_a), len(encoded_sequence_b)
-(8, 19)
-```
-
-Therefore, we can't put them together in the same tensor as-is. The first sequence needs to be padded up to the length
-of the second one, or the second one needs to be truncated down to the length of the first one.
-
-In the first case, the list of IDs will be extended by the padding indices. We can pass a list to the tokenizer and ask
-it to pad like this:
-
-```python
->>> padded_sequences = tokenizer([sequence_a, sequence_b], padding=True)
-```
-
-We can see that 0s have been added on the right of the first sentence to make it the same length as the second one:
-
-```python
->>> padded_sequences["input_ids"]
-[[101, 1188, 1110, 170, 1603, 4954, 119, 102, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [101, 1188, 1110, 170, 1897, 1263, 4954, 119, 1135, 1110, 1120, 1655, 2039, 1190, 1103, 4954, 138, 119, 102]]
-```
-
-This can then be converted into a tensor in PyTorch or TensorFlow. The attention mask is a binary tensor indicating the
-position of the padded indices so that the model does not attend to them. For the [`BertTokenizer`],
-`1` indicates a value that should be attended to, while `0` indicates a padded value. This attention mask is
-in the dictionary returned by the tokenizer under the key "attention_mask":
-
-```python
->>> padded_sequences["attention_mask"]
-[[1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]]
-```
-
-<a id='token-type-ids'></a>
-
-### Token Type IDs
-
-Some models' purpose is to do classification on pairs of sentences or question answering.
-
-<Youtube id="0u3ioSwev3s"/>
-
-These require two different sequences to be joined in a single "input_ids" entry, which usually is performed with the
-help of special tokens, such as the classifier (`[CLS]`) and separator (`[SEP]`) tokens. For example, the BERT
-model builds its two sequence input as such:
-
-```python
->>> # [CLS] SEQUENCE_A [SEP] SEQUENCE_B [SEP]
-```
-
-We can use our tokenizer to automatically generate such a sentence by passing the two sequences to `tokenizer` as two
-arguments (and not a list, like before) like this:
-
-```python
->>> from transformers import BertTokenizer
-
->>> tokenizer = BertTokenizer.from_pretrained("bert-base-cased")
->>> sequence_a = "HuggingFace is based in NYC"
->>> sequence_b = "Where is HuggingFace based?"
-
->>> encoded_dict = tokenizer(sequence_a, sequence_b)
->>> decoded = tokenizer.decode(encoded_dict["input_ids"])
-```
-
-which will return:
-
-```python
->>> print(decoded)
-[CLS] HuggingFace is based in NYC [SEP] Where is HuggingFace based? [SEP]
-```
-
-This is enough for some models to understand where one sequence ends and where another begins. However, other models,
-such as BERT, also deploy token type IDs (also called segment IDs). They are represented as a binary mask identifying
-the two types of sequence in the model.
-
-The tokenizer returns this mask as the "token_type_ids" entry:
-
-```python
->>> encoded_dict["token_type_ids"]
-[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1]
-```
-
-The first sequence, the "context" used for the question, has all its tokens represented by a `0`, whereas the
-second sequence, corresponding to the "question", has all its tokens represented by a `1`.
-
-Some models, like [`XLNetModel`] use an additional token represented by a `2`.
-
-<a id='position-ids'></a>
-
-### Position IDs
-
-Contrary to RNNs that have the position of each token embedded within them, transformers are unaware of the position of
-each token. Therefore, the position IDs (`position_ids`) are used by the model to identify each token's position in
-the list of tokens.
-
-They are an optional parameter. If no `position_ids` are passed to the model, the IDs are automatically created as
-absolute positional embeddings.
-
-Absolute positional embeddings are selected in the range `[0, config.max_position_embeddings - 1]`. Some models use
-other types of positional embeddings, such as sinusoidal position embeddings or relative position embeddings.
-
-<a id='labels'></a>
-
-### Labels
-
-The labels are an optional argument which can be passed in order for the model to compute the loss itself. These labels
-should be the expected prediction of the model: it will use the standard loss in order to compute the loss between its
-predictions and the expected value (the label).
-
-These labels are different according to the model head, for example:
-
-- For sequence classification models (e.g., [`BertForSequenceClassification`]), the model expects a
-  tensor of dimension `(batch_size)` with each value of the batch corresponding to the expected label of the
-  entire sequence.
-- For token classification models (e.g., [`BertForTokenClassification`]), the model expects a tensor
-  of dimension `(batch_size, seq_length)` with each value corresponding to the expected label of each individual
-  token.
-- For masked language modeling (e.g., [`BertForMaskedLM`]), the model expects a tensor of dimension
-  `(batch_size, seq_length)` with each value corresponding to the expected label of each individual token: the
-  labels being the token ID for the masked token, and values to be ignored for the rest (usually -100).
-- For sequence to sequence tasks,(e.g., [`BartForConditionalGeneration`],
-  [`MBartForConditionalGeneration`]), the model expects a tensor of dimension `(batch_size, tgt_seq_length)` with each value corresponding to the target sequences associated with each input sequence. During
-  training, both *BART* and *T5* will make the appropriate *decoder_input_ids* and decoder attention masks internally.
-  They usually do not need to be supplied. This does not apply to models leveraging the Encoder-Decoder framework. See
-  the documentation of each model for more information on each specific model's labels.
-
-The base models (e.g., [`BertModel`]) do not accept labels, as these are the base transformer
-models, simply outputting features.
-
-<a id='decoder-input-ids'></a>
-
-### Decoder input IDs
-
-This input is specific to encoder-decoder models, and contains the input IDs that will be fed to the decoder. These
-inputs should be used for sequence to sequence tasks, such as translation or summarization, and are usually built in a
-way specific to each model.
-
-Most encoder-decoder models (BART, T5) create their `decoder_input_ids` on their own from the `labels`. In
-such models, passing the `labels` is the preferred way to handle training.
-
-Please check each model's docs to see how they handle these input IDs for sequence to sequence training.
-
-<a id='feed-forward-chunking'></a>
-
-### Feed Forward Chunking
-
-In each residual attention block in transformers the self-attention layer is usually followed by 2 feed forward layers.
-The intermediate embedding size of the feed forward layers is often bigger than the hidden size of the model (e.g., for
-`bert-base-uncased`).
-
-For an input of size `[batch_size, sequence_length]`, the memory required to store the intermediate feed forward
-embeddings `[batch_size, sequence_length, config.intermediate_size]` can account for a large fraction of the memory
-use. The authors of [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) noticed that since the
-computation is independent of the `sequence_length` dimension, it is mathematically equivalent to compute the output
-embeddings of both feed forward layers `[batch_size, config.hidden_size]_0, ..., [batch_size, config.hidden_size]_n`
-individually and concat them afterward to `[batch_size, sequence_length, config.hidden_size]` with `n = sequence_length`, which trades increased computation time against reduced memory use, but yields a mathematically
-**equivalent** result.
-
-For models employing the function [`apply_chunking_to_forward`], the `chunk_size` defines the
-number of output embeddings that are computed in parallel and thus defines the trade-off between memory and time
-complexity. If `chunk_size` is set to 0, no feed forward chunking is done.

docs/source/en/index.mdx

@@ -1,322 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# 🤗 Transformers
-
-State-of-the-art Machine Learning for PyTorch, TensorFlow and JAX.
-
-🤗 Transformers provides APIs to easily download and train state-of-the-art pretrained models. Using pretrained models can reduce your compute costs, carbon footprint, and save you time from training a model from scratch. The models can be used across different modalities such as:
-
-* 📝 Text: text classification, information extraction, question answering, summarization, translation, and text generation in over 100 languages.
-* 🖼️ Images: image classification, object detection, and segmentation.
-* 🗣️ Audio: speech recognition and audio classification.
-* 🐙 Multimodal: table question answering, optical character recognition, information extraction from scanned documents, video classification, and visual question answering.
-
-Our library supports seamless integration between three of the most popular deep learning libraries: [PyTorch](https://pytorch.org/), [TensorFlow](https://www.tensorflow.org/) and [JAX](https://jax.readthedocs.io/en/latest/). Train your model in three lines of code in one framework, and load it for inference with another.
-
-Each 🤗 Transformers architecture is defined in a standalone Python module so they can be easily customized for research and experiments.
-
-## If you are looking for custom support from the Hugging Face team
-
-<a target="_blank" href="https://huggingface.co/support">
-    <img alt="HuggingFace Expert Acceleration Program" src="https://cdn-media.huggingface.co/marketing/transformers/new-support-improved.png" style="max-width: 600px; border: 1px solid #eee; border-radius: 4px; box-shadow: 0 1px 2px 0 rgba(0, 0, 0, 0.05);">
-</a><br>
-
-## Contents
-
-The documentation is organized in five parts:
-
-- **GET STARTED** contains a quick tour and installation instructions to get up and running with 🤗 Transformers.
-- **TUTORIALS** are a great place to begin if you are new to our library. This section will help you gain the basic skills you need to start using 🤗 Transformers.
-- **HOW-TO GUIDES** will show you how to achieve a specific goal like fine-tuning a pretrained model for language modeling or how to create a custom model head.
-- **CONCEPTUAL GUIDES** provides more discussion and explanation of the underlying concepts and ideas behind models, tasks, and the design philosophy of 🤗 Transformers. 
-- **API** describes each class and function, grouped in:
-
-  - **MAIN CLASSES** for the main classes exposing the important APIs of the library.
-  - **MODELS** for the classes and functions related to each model implemented in the library.
-  - **INTERNAL HELPERS** for the classes and functions we use internally.
-
-The library currently contains JAX, PyTorch and TensorFlow implementations, pretrained model weights, usage scripts and conversion utilities for the following models.
-
-### Supported models
-
-<!--This list is updated automatically from the README with _make fix-copies_. Do not update manually! -->
-
-1. **[ALBERT](model_doc/albert)** (from Google Research and the Toyota Technological Institute at Chicago) released with the paper [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://arxiv.org/abs/1909.11942), by Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut.
-1. **[BART](model_doc/bart)** (from Facebook) released with the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/abs/1910.13461) by Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Ves Stoyanov and Luke Zettlemoyer.
-1. **[BARThez](model_doc/barthez)** (from École polytechnique) released with the paper [BARThez: a Skilled Pretrained French Sequence-to-Sequence Model](https://arxiv.org/abs/2010.12321) by Moussa Kamal Eddine, Antoine J.-P. Tixier, Michalis Vazirgiannis.
-1. **[BARTpho](model_doc/bartpho)** (from VinAI Research) released with the paper [BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese](https://arxiv.org/abs/2109.09701) by Nguyen Luong Tran, Duong Minh Le and Dat Quoc Nguyen.
-1. **[BEiT](model_doc/beit)** (from Microsoft) released with the paper [BEiT: BERT Pre-Training of Image Transformers](https://arxiv.org/abs/2106.08254) by Hangbo Bao, Li Dong, Furu Wei.
-1. **[BERT](model_doc/bert)** (from Google) released with the paper [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://arxiv.org/abs/1810.04805) by Jacob Devlin, Ming-Wei Chang, Kenton Lee and Kristina Toutanova.
-1. **[BERT For Sequence Generation](model_doc/bert-generation)** (from Google) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
-1. **[BERTweet](model_doc/bertweet)** (from VinAI Research) released with the paper [BERTweet: A pre-trained language model for English Tweets](https://aclanthology.org/2020.emnlp-demos.2/) by Dat Quoc Nguyen, Thanh Vu and Anh Tuan Nguyen.
-1. **[BigBird-Pegasus](model_doc/bigbird_pegasus)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
-1. **[BigBird-RoBERTa](model_doc/big_bird)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
-1. **[Blenderbot](model_doc/blenderbot)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
-1. **[BlenderbotSmall](model_doc/blenderbot-small)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
-1. **[BLOOM](model_doc/bloom)** (from BigScience workshop) released by the [BigSicence Workshop](https://bigscience.huggingface.co/).
-1. **[BORT](model_doc/bort)** (from Alexa) released with the paper [Optimal Subarchitecture Extraction For BERT](https://arxiv.org/abs/2010.10499) by Adrian de Wynter and Daniel J. Perry.
-1. **[ByT5](model_doc/byt5)** (from Google Research) released with the paper [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) by Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel.
-1. **[CamemBERT](model_doc/camembert)** (from Inria/Facebook/Sorbonne) released with the paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) by Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suárez*, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah and Benoît Sagot.
-1. **[CANINE](model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
-1. **[CLIP](model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
-1. **[CodeGen](model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
-1. **[ConvBERT](model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
-1. **[ConvNeXT](model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
-1. **[CPM](model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
-1. **[CTRL](model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
-1. **[CvT](model_doc/cvt)** (from Microsoft) released with the paper [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) by Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang.
-1. **[Data2Vec](model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
-1. **[DeBERTa](model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
-1. **[DeBERTa-v2](model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
-1. **[Decision Transformer](model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
-1. **[DeiT](model_doc/deit)** (from Facebook) released with the paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) by Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou.
-1. **[DETR](model_doc/detr)** (from Facebook) released with the paper [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) by Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko.
-1. **[DialoGPT](model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
-1. **[DistilBERT](model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation) and a German version of DistilBERT.
-1. **[DiT](model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
-1. **[DPR](model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
-1. **[DPT](master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by René Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
-1. **[ELECTRA](model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
-1. **[EncoderDecoder](model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
-1. **[FlauBERT](model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab.
-1. **[FLAVA](model_doc/flava)** (from Facebook AI) released with the paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) by Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela.
-1. **[FNet](model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
-1. **[Funnel Transformer](model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
-1. **[GLPN](model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
-1. **[GPT](model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
-1. **[GPT Neo](model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
-1. **[GPT NeoX](model_doc/gpt_neox)** (from EleutherAI) released with the paper [GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://arxiv.org/abs/2204.06745) by Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, Samuel Weinbach
-1. **[GPT-2](model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
-1. **[GPT-J](model_doc/gptj)** (from EleutherAI) released in the repository [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
-1. **[GroupViT](model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
-1. **[Hubert](model_doc/hubert)** (from Facebook) released with the paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) by Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
-1. **[I-BERT](model_doc/ibert)** (from Berkeley) released with the paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) by Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
-1. **[ImageGPT](model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
-1. **[LayoutLM](model_doc/layoutlm)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
-1. **[LayoutLMv2](model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) by Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
-1. **[LayoutLMv3](model_doc/layoutlmv3)** (from Microsoft Research Asia) released with the paper [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) by Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei.
-1. **[LayoutXLM](model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
-1. **[LED](model_doc/led)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
-1. **[LeViT](model_doc/levit)** (from Meta AI) released with the paper [LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference](https://arxiv.org/abs/2104.01136) by Ben Graham, Alaaeldin El-Nouby, Hugo Touvron, Pierre Stock, Armand Joulin, Hervé Jégou, Matthijs Douze.
-1. **[Longformer](model_doc/longformer)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
-1. **[LongT5](model_doc/longt5)** (from Google AI) released with the paper [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://arxiv.org/abs/2112.07916) by Mandy Guo, Joshua Ainslie, David Uthus, Santiago Ontanon, Jianmo Ni, Yun-Hsuan Sung, Yinfei Yang.
-1. **[LUKE](model_doc/luke)** (from Studio Ousia) released with the paper [LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention](https://arxiv.org/abs/2010.01057) by Ikuya Yamada, Akari Asai, Hiroyuki Shindo, Hideaki Takeda, Yuji Matsumoto.
-1. **[LXMERT](model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
-1. **[M-CTC-T](model_doc/mctct)** (from Facebook) released with the paper [Pseudo-Labeling For Massively Multilingual Speech Recognition](https://arxiv.org/abs/2111.00161) by Loren Lugosch, Tatiana Likhomanenko, Gabriel Synnaeve, and Ronan Collobert.
-1. **[M2M100](model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
-1. **[MarianMT](model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jörg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
-1. **[MaskFormer](model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
-1. **[mBART](model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
-1. **[mBART-50](model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
-1. **[Megatron-BERT](model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
-1. **[Megatron-GPT2](model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
-1. **[mLUKE](model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
-1. **[MobileBERT](model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
-1. **[MobileViT](model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
-1. **[MPNet](model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
-1. **[MT5](model_doc/mt5)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
-1. **[MVP](model_doc/mvp)** (from RUC AI Box) released with the paper [MVP: Multi-task Supervised Pre-training for Natural Language Generation](https://arxiv.org/abs/2206.12131) by Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen.
-1. **[Nezha](model_doc/nezha)** (from Huawei Noah’s Ark Lab) released with the paper [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://arxiv.org/abs/1909.00204) by Junqiu Wei, Xiaozhe Ren, Xiaoguang Li, Wenyong Huang, Yi Liao, Yasheng Wang, Jiashu Lin, Xin Jiang, Xiao Chen and Qun Liu.
-1. **[NLLB](model_doc/nllb)** (from Meta) released with the paper [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) by the NLLB team.
-1. **[Nyströmformer](model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
-1. **[OPT](master/model_doc/opt)** (from Meta AI) released with the paper [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) by Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al.
-1. **[OWL-ViT](model_doc/owlvit)** (from Google AI) released with the paper [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) by Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby.
-1. **[Pegasus](model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
-1. **[Perceiver IO](model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira.
-1. **[PhoBERT](model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
-1. **[PLBart](model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
-1. **[PoolFormer](model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
-1. **[ProphetNet](model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
-1. **[QDQBert](model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
-1. **[RAG](model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
-1. **[REALM](model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
-1. **[Reformer](model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
-1. **[RegNet](model_doc/regnet)** (from META Platforms) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
-1. **[RemBERT](model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
-1. **[ResNet](model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun.
-1. **[RoBERTa](model_doc/roberta)** (from Facebook), released together with the paper [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
-1. **[RoFormer](model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/abs/2104.09864) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
-1. **[SegFormer](model_doc/segformer)** (from NVIDIA) released with the paper [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) by Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo.
-1. **[SEW](model_doc/sew)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
-1. **[SEW-D](model_doc/sew_d)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
-1. **[SpeechToTextTransformer](model_doc/speech_to_text)** (from Facebook), released together with the paper [fairseq S2T: Fast Speech-to-Text Modeling with fairseq](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Dmytro Okhonko, Juan Pino.
-1. **[SpeechToTextTransformer2](model_doc/speech_to_text_2)** (from Facebook), released together with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
-1. **[Splinter](model_doc/splinter)** (from Tel Aviv University), released together with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
-1. **[SqueezeBERT](model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
-1. **[Swin Transformer](model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
-1. **[Swin Transformer V2](model_doc/swinv2)** (from Microsoft) released with the paper [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) by Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo.
-1. **[T5](model_doc/t5)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
-1. **[T5v1.1](model_doc/t5v1.1)** (from Google AI) released in the repository [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
-1. **[TAPAS](model_doc/tapas)** (from Google AI) released with the paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) by Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos.
-1. **[TAPEX](model_doc/tapex)** (from Microsoft Research) released with the paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) by Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
-1. **[Trajectory Transformer](model_doc/trajectory_transformers)** (from the University of California at Berkeley) released with the paper [Offline Reinforcement Learning as One Big Sequence Modeling Problem](https://arxiv.org/abs/2106.02039) by Michael Janner, Qiyang Li, Sergey Levine
-1. **[Transformer-XL](model_doc/transfo-xl)** (from Google/CMU) released with the paper [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) by Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov.
-1. **[TrOCR](model_doc/trocr)** (from Microsoft), released together with the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
-1. **[UL2](model_doc/ul2)** (from Google Research) released with the paper [Unifying Language Learning Paradigms](https://arxiv.org/abs/2205.05131v1) by Yi Tay, Mostafa Dehghani, Vinh Q. Tran, Xavier Garcia, Dara Bahri, Tal Schuster, Huaixiu Steven Zheng, Neil Houlsby, Donald Metzler
-1. **[UniSpeech](model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
-1. **[UniSpeechSat](model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
-1. **[VAN](model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/abs/2202.09741) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
-1. **[VideoMAE](model_doc/videomae)** (from Multimedia Computing Group, Nanjing University) released with the paper [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://arxiv.org/abs/2203.12602) by Zhan Tong, Yibing Song, Jue Wang, Limin Wang.
-1. **[ViLT](model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
-1. **[Vision Transformer (ViT)](model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby.
-1. **[VisualBERT](model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
-1. **[ViTMAE](model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
-1. **[Wav2Vec2](model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
-1. **[Wav2Vec2-Conformer](model_doc/wav2vec2-conformer)** (from Facebook AI) released with the paper [FAIRSEQ S2T: Fast Speech-to-Text Modeling with FAIRSEQ](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Sravya Popuri, Dmytro Okhonko, Juan Pino.
-1. **[Wav2Vec2Phoneme](model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
-1. **[WavLM](model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
-1. **[XGLM](model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li.
-1. **[XLM](model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
-1. **[XLM-ProphetNet](model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
-1. **[XLM-RoBERTa](model_doc/xlm-roberta)** (from Facebook AI), released together with the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov.
-1. **[XLM-RoBERTa-XL](model_doc/xlm-roberta-xl)** (from Facebook AI), released together with the paper [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) by Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau.
-1. **[XLNet](model_doc/xlnet)** (from Google/CMU) released with the paper [​XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
-1. **[XLS-R](model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
-1. **[XLSR-Wav2Vec2](model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
-1. **[YOLOS](model_doc/yolos)** (from Huazhong University of Science & Technology) released with the paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) by Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
-1. **[YOSO](model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) Once: Linear Cost Self-Attention Via Bernoulli Sampling](https://arxiv.org/abs/2111.09714) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
-
-
-### Supported frameworks
-
-The table below represents the current support in the library for each of those models, whether they have a Python
-tokenizer (called "slow"). A "fast" tokenizer backed by the 🤗 Tokenizers library, whether they have support in Jax (via
-Flax), PyTorch, and/or TensorFlow.
-
-<!--This table is updated automatically from the auto modules with _make fix-copies_. Do not update manually!-->
-
-|            Model            | Tokenizer slow | Tokenizer fast | PyTorch support | TensorFlow support | Flax Support |
-|:---------------------------:|:--------------:|:--------------:|:---------------:|:------------------:|:------------:|
-|           ALBERT            |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|            BART             |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|            BEiT             |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
-|            BERT             |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|       Bert Generation       |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
-|           BigBird           |       ✅       |       ✅       |       ✅        |         ❌         |      ✅      |
-|       BigBird-Pegasus       |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|         Blenderbot          |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|       BlenderbotSmall       |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|            BLOOM            |       ❌       |       ✅       |       ✅        |         ❌         |      ❌      |
-|          CamemBERT          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|           CANINE            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
-|            CLIP             |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|           CodeGen           |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
-|          ConvBERT           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|          ConvNeXT           |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
-|            CTRL             |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
-|             CvT             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|        Data2VecAudio        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|        Data2VecText         |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|       Data2VecVision        |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
-|           DeBERTa           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|         DeBERTa-v2          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|    Decision Transformer     |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|            DeiT             |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
-|            DETR             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|         DistilBERT          |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|             DPR             |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|             DPT             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|           ELECTRA           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|       Encoder decoder       |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
-| FairSeq Machine-Translation |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
-|          FlauBERT           |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
-|            FLAVA            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|            FNet             |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
-|     Funnel Transformer      |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|            GLPN             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|           GPT Neo           |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
-|          GPT NeoX           |       ❌       |       ✅       |       ✅        |         ❌         |      ❌      |
-|            GPT-J            |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
-|          GroupViT           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|           Hubert            |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
-|           I-BERT            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|          ImageGPT           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|          LayoutLM           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|         LayoutLMv2          |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
-|         LayoutLMv3          |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
-|             LED             |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|            LeViT            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|         Longformer          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|           LongT5            |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
-|            LUKE             |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
-|           LXMERT            |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|           M-CTC-T           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|           M2M100            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
-|           Marian            |       ✅       |       ❌       |       ✅        |         ✅         |      ✅      |
-|         MaskFormer          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|            mBART            |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|        Megatron-BERT        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|         MobileBERT          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|          MobileViT          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|            MPNet            |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|             MT5             |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|             MVP             |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
-|            Nezha            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|        Nyströmformer        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|         OpenAI GPT          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|        OpenAI GPT-2         |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|             OPT             |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
-|           OWL-ViT           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|           Pegasus           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|          Perceiver          |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
-|           PLBart            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
-|         PoolFormer          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|         ProphetNet          |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
-|           QDQBert           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|             RAG             |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
-|            REALM            |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
-|          Reformer           |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
-|           RegNet            |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
-|           RemBERT           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|           ResNet            |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
-|          RetriBERT          |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
-|           RoBERTa           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|          RoFormer           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|          SegFormer          |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
-|             SEW             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|            SEW-D            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|   Speech Encoder decoder    |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
-|         Speech2Text         |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
-|        Speech2Text2         |       ✅       |       ❌       |       ❌        |         ❌         |      ❌      |
-|          Splinter           |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
-|         SqueezeBERT         |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
-|      Swin Transformer       |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
-|     Swin Transformer V2     |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|             T5              |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|            TAPAS            |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
-|   Trajectory Transformer    |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|       Transformer-XL        |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
-|            TrOCR            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|          UniSpeech          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|        UniSpeechSat         |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|             VAN             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|          VideoMAE           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|            ViLT             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|   Vision Encoder decoder    |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
-|    VisionTextDualEncoder    |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
-|         VisualBERT          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|             ViT             |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
-|           ViTMAE            |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
-|          Wav2Vec2           |       ✅       |       ❌       |       ✅        |         ✅         |      ✅      |
-|     Wav2Vec2-Conformer      |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|            WavLM            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|            XGLM             |       ✅       |       ✅       |       ✅        |         ❌         |      ✅      |
-|             XLM             |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
-|       XLM-ProphetNet        |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
-|         XLM-RoBERTa         |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|       XLM-RoBERTa-XL        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|            XLNet            |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|            YOLOS            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|            YOSO             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-
-<!-- End table-->

docs/source/en/installation.mdx

@@ -1,240 +0,0 @@
-<!---
-Copyright 2022 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
--->
-
-# Installation
-
-Install 🤗 Transformers for whichever deep learning library you're working with, setup your cache, and optionally configure 🤗 Transformers to run offline.
-
-🤗 Transformers is tested on Python 3.6+, PyTorch 1.1.0+, TensorFlow 2.0+, and Flax. Follow the installation instructions below for the deep learning library you are using:
-
-* [PyTorch](https://pytorch.org/get-started/locally/) installation instructions.
-* [TensorFlow 2.0](https://www.tensorflow.org/install/pip) installation instructions.
-* [Flax](https://flax.readthedocs.io/en/latest/) installation instructions.
-
-## Install with pip
-
-You should install 🤗 Transformers in a [virtual environment](https://docs.python.org/3/library/venv.html). If you're unfamiliar with Python virtual environments, take a look at this [guide](https://packaging.python.org/guides/installing-using-pip-and-virtual-environments/). A virtual environment makes it easier to manage different projects, and avoid compatibility issues between dependencies.
-
-Start by creating a virtual environment in your project directory:
-
-```bash
-python -m venv .env
-```
-
-Activate the virtual environment. On Linux and MacOs:
-
-```bash
-source .env/bin/activate
-```
-Activate Virtual environment on Windows
-
-```bash
-.env/Scripts/activate
-```
-
-Now you're ready to install 🤗 Transformers with the following command:
-
-```bash
-pip install transformers
-```
-
-For CPU-support only, you can conveniently install 🤗 Transformers and a deep learning library in one line. For example, install 🤗 Transformers and PyTorch with:
-
-```bash
-pip install transformers[torch]
-```
-
-🤗 Transformers and TensorFlow 2.0:
-
-```bash
-pip install transformers[tf-cpu]
-```
-
-🤗 Transformers and Flax:
-
-```bash
-pip install transformers[flax]
-```
-
-Finally, check if 🤗 Transformers has been properly installed by running the following command. It will download a pretrained model:
-
-```bash
-python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('we love you'))"
-```
-
-Then print out the label and score:
-
-```bash
-[{'label': 'POSITIVE', 'score': 0.9998704791069031}]
-```
-
-## Install from source
-
-Install 🤗 Transformers from source with the following command:
-
-```bash
-pip install git+https://github.com/huggingface/transformers
-```
-
-This command installs the bleeding edge `main` version rather than the latest `stable` version. The `main` version is useful for staying up-to-date with the latest developments. For instance, if a bug has been fixed since the last official release but a new release hasn't been rolled out yet. However, this means the `main` version may not always be stable. We strive to keep the `main` version operational, and most issues are usually resolved within a few hours or a day. If you run into a problem, please open an [Issue](https://github.com/huggingface/transformers/issues) so we can fix it even sooner!
-
-Check if 🤗 Transformers has been properly installed by running the following command:
-
-```bash
-python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('I love you'))"
-```
-
-## Editable install
-
-You will need an editable install if you'd like to:
-
-* Use the `main` version of the source code.
-* Contribute to 🤗 Transformers and need to test changes in the code.
-
-Clone the repository and install 🤗 Transformers with the following commands:
-
-```bash
-git clone https://github.com/huggingface/transformers.git
-cd transformers
-pip install -e .
-```
-
-These commands will link the folder you cloned the repository to and your Python library paths. Python will now look inside the folder you cloned to in addition to the normal library paths. For example, if your Python packages are typically installed in `~/anaconda3/envs/main/lib/python3.7/site-packages/`, Python will also search the folder you cloned to: `~/transformers/`.
-
-<Tip warning={true}>
-
-You must keep the `transformers` folder if you want to keep using the library.
-
-</Tip>
-
-Now you can easily update your clone to the latest version of 🤗 Transformers with the following command:
-
-```bash
-cd ~/transformers/
-git pull
-```
-
-Your Python environment will find the `main` version of 🤗 Transformers on the next run.
-
-## Install with conda
-
-Install from the conda channel `huggingface`:
-
-```bash
-conda install -c huggingface transformers
-```
-
-## Cache setup
-
-Pretrained models are downloaded and locally cached at: `~/.cache/huggingface/hub`. This is the default directory given by the shell environment variable `TRANSFORMERS_CACHE`. On Windows, the default directory is given by `C:\Users\username\.cache\huggingface\hub`. You can change the shell environment variables shown below - in order of priority - to specify a different cache directory:
-
-1. Shell environment variable (default): `HUGGINGFACE_HUB_CACHE` or `TRANSFORMERS_CACHE`.
-2. Shell environment variable: `HF_HOME`.
-3. Shell environment variable: `XDG_CACHE_HOME` + `/huggingface`.
-
-<Tip>
-
-🤗 Transformers will use the shell environment variables `PYTORCH_TRANSFORMERS_CACHE` or `PYTORCH_PRETRAINED_BERT_CACHE` if you are coming from an earlier iteration of this library and have set those environment variables, unless you specify the shell environment variable `TRANSFORMERS_CACHE`.
-
-</Tip>
-
-## Offline mode
-
-🤗 Transformers is able to run in a firewalled or offline environment by only using local files. Set the environment variable `TRANSFORMERS_OFFLINE=1` to enable this behavior.
-
-<Tip>
-
-Add [🤗 Datasets](https://huggingface.co/docs/datasets/) to your offline training workflow by setting the environment variable `HF_DATASETS_OFFLINE=1`.
-
-</Tip>
-
-For example, you would typically run a program on a normal network firewalled to external instances with the following command:
-
-```bash
-python examples/pytorch/translation/run_translation.py --model_name_or_path t5-small --dataset_name wmt16 --dataset_config ro-en ...
-```
-
-Run this same program in an offline instance with:
-
-```bash
-HF_DATASETS_OFFLINE=1 TRANSFORMERS_OFFLINE=1 \
-python examples/pytorch/translation/run_translation.py --model_name_or_path t5-small --dataset_name wmt16 --dataset_config ro-en ...
-```
-
-The script should now run without hanging or waiting to timeout because it knows it should only look for local files.
-
-### Fetch models and tokenizers to use offline
-
-Another option for using 🤗 Transformers offline is to download the files ahead of time, and then point to their local path when you need to use them offline. There are three ways to do this:
-
-* Download a file through the user interface on the [Model Hub](https://huggingface.co/models) by clicking on the ↓ icon.
-
-    ![download-icon](https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/download-icon.png)
-
-* Use the [`PreTrainedModel.from_pretrained`] and [`PreTrainedModel.save_pretrained`] workflow:
-
-    1. Download your files ahead of time with [`PreTrainedModel.from_pretrained`]:
-
-    ```py
-    >>> from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
-
-    >>> tokenizer = AutoTokenizer.from_pretrained("bigscience/T0_3B")
-    >>> model = AutoModelForSeq2SeqLM.from_pretrained("bigscience/T0_3B")
-    ```
-
-    2. Save your files to a specified directory with [`PreTrainedModel.save_pretrained`]:
-
-    ```py
-    >>> tokenizer.save_pretrained("./your/path/bigscience_t0")
-    >>> model.save_pretrained("./your/path/bigscience_t0")
-    ```
-
-    3. Now when you're offline, reload your files with [`PreTrainedModel.from_pretrained`] from the specified directory:
-
-    ```py
-    >>> tokenizer = AutoTokenizer.from_pretrained("./your/path/bigscience_t0")
-    >>> model = AutoModel.from_pretrained("./your/path/bigscience_t0")
-    ```
-
-* Programmatically download files with the [huggingface_hub](https://github.com/huggingface/huggingface_hub/tree/main/src/huggingface_hub) library:
-
-    1. Install the `huggingface_hub` library in your virtual environment:
-
-    ```bash
-    python -m pip install huggingface_hub
-    ```
-
-    2. Use the [`hf_hub_download`](https://huggingface.co/docs/hub/adding-a-library#download-files-from-the-hub) function to download a file to a specific path. For example, the following command downloads the `config.json` file from the [T0](https://huggingface.co/bigscience/T0_3B) model to your desired path:
-
-    ```py
-    >>> from huggingface_hub import hf_hub_download
-
-    >>> hf_hub_download(repo_id="bigscience/T0_3B", filename="config.json", cache_dir="./your/path/bigscience_t0")
-    ```
-
-Once your file is downloaded and locally cached, specify it's local path to load and use it:
-
-```py
->>> from transformers import AutoConfig
-
->>> config = AutoConfig.from_pretrained("./your/path/bigscience_t0/config.json")
-```
-
-<Tip>
-
-See the [How to download files from the Hub](https://huggingface.co/docs/hub/how-to-downstream) section for more details on downloading files stored on the Hub.
-
-</Tip>

docs/source/en/internal/file_utils.mdx

@@ -1,46 +0,0 @@
-<!--Copyright 2021 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# General Utilities
-
-This page lists all of Transformers general utility functions that are found in the file `utils.py`.
-
-Most of those are only useful if you are studying the general code in the library.
-
-
-## Enums and namedtuples
-
-[[autodoc]] utils.ExplicitEnum
-
-[[autodoc]] utils.PaddingStrategy
-
-[[autodoc]] utils.TensorType
-
-## Special Decorators
-
-[[autodoc]] utils.add_start_docstrings
-
-[[autodoc]] utils.add_start_docstrings_to_model_forward
-
-[[autodoc]] utils.add_end_docstrings
-
-[[autodoc]] utils.add_code_sample_docstrings
-
-[[autodoc]] utils.replace_return_docstrings
-
-## Special Properties
-
-[[autodoc]] utils.cached_property
-
-## Other Utilities
-
-[[autodoc]] utils._LazyModule

docs/source/en/internal/generation_utils.mdx

@@ -1,263 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Utilities for Generation
-
-This page lists all the utility functions used by [`~generation_utils.GenerationMixin.generate`],
-[`~generation_utils.GenerationMixin.greedy_search`],
-[`~generation_utils.GenerationMixin.sample`],
-[`~generation_utils.GenerationMixin.beam_search`],
-[`~generation_utils.GenerationMixin.beam_sample`],
-[`~generation_utils.GenerationMixin.group_beam_search`], and
-[`~generation_utils.GenerationMixin.constrained_beam_search`].
-
-Most of those are only useful if you are studying the code of the generate methods in the library.
-
-## Generate Outputs
-
-The output of [`~generation_utils.GenerationMixin.generate`] is an instance of a subclass of
-[`~utils.ModelOutput`]. This output is a data structure containing all the information returned
-by [`~generation_utils.GenerationMixin.generate`], but that can also be used as tuple or dictionary.
-
-Here's an example:
-
-```python
-from transformers import GPT2Tokenizer, GPT2LMHeadModel
-
-tokenizer = GPT2Tokenizer.from_pretrained("gpt2")
-model = GPT2LMHeadModel.from_pretrained("gpt2")
-
-inputs = tokenizer("Hello, my dog is cute and ", return_tensors="pt")
-generation_output = model.generate(**inputs, return_dict_in_generate=True, output_scores=True)
-```
-
-The `generation_output` object is a [`~generation_utils.GreedySearchDecoderOnlyOutput`], as we can
-see in the documentation of that class below, it means it has the following attributes:
-
-- `sequences`: the generated sequences of tokens
-- `scores` (optional): the prediction scores of the language modelling head, for each generation step
-- `hidden_states` (optional): the hidden states of the model, for each generation step
-- `attentions` (optional): the attention weights of the model, for each generation step
-
-Here we have the `scores` since we passed along `output_scores=True`, but we don't have `hidden_states` and
-`attentions` because we didn't pass `output_hidden_states=True` or `output_attentions=True`.
-
-You can access each attribute as you would usually do, and if that attribute has not been returned by the model, you
-will get `None`. Here for instance `generation_output.scores` are all the generated prediction scores of the
-language modeling head, and `generation_output.attentions` is `None`.
-
-When using our `generation_output` object as a tuple, it only keeps the attributes that don't have `None` values.
-Here, for instance, it has two elements, `loss` then `logits`, so
-
-```python
-generation_output[:2]
-```
-
-will return the tuple `(generation_output.sequences, generation_output.scores)` for instance.
-
-When using our `generation_output` object as a dictionary, it only keeps the attributes that don't have `None`
-values. Here, for instance, it has two keys that are `sequences` and `scores`.
-
-We document here all output types.
-
-
-### GreedySearchOutput
-
-[[autodoc]] generation_utils.GreedySearchDecoderOnlyOutput
-
-[[autodoc]] generation_utils.GreedySearchEncoderDecoderOutput
-
-[[autodoc]] generation_flax_utils.FlaxGreedySearchOutput
-
-### SampleOutput
-
-[[autodoc]] generation_utils.SampleDecoderOnlyOutput
-
-[[autodoc]] generation_utils.SampleEncoderDecoderOutput
-
-[[autodoc]] generation_flax_utils.FlaxSampleOutput
-
-### BeamSearchOutput
-
-[[autodoc]] generation_utils.BeamSearchDecoderOnlyOutput
-
-[[autodoc]] generation_utils.BeamSearchEncoderDecoderOutput
-
-### BeamSampleOutput
-
-[[autodoc]] generation_utils.BeamSampleDecoderOnlyOutput
-
-[[autodoc]] generation_utils.BeamSampleEncoderDecoderOutput
-
-## LogitsProcessor
-
-A [`LogitsProcessor`] can be used to modify the prediction scores of a language model head for
-generation.
-
-[[autodoc]] LogitsProcessor
-    - __call__
-
-[[autodoc]] LogitsProcessorList
-    - __call__
-
-[[autodoc]] LogitsWarper
-    - __call__
-
-[[autodoc]] MinLengthLogitsProcessor
-    - __call__
-
-[[autodoc]] TemperatureLogitsWarper
-    - __call__
-
-[[autodoc]] RepetitionPenaltyLogitsProcessor
-    - __call__
-
-[[autodoc]] TopPLogitsWarper
-    - __call__
-
-[[autodoc]] TopKLogitsWarper
-    - __call__
-
-[[autodoc]] TypicalLogitsWarper
-    - __call__
-
-[[autodoc]] NoRepeatNGramLogitsProcessor
-    - __call__
-
-[[autodoc]] NoBadWordsLogitsProcessor
-    - __call__
-
-[[autodoc]] PrefixConstrainedLogitsProcessor
-    - __call__
-
-[[autodoc]] HammingDiversityLogitsProcessor
-    - __call__
-
-[[autodoc]] ForcedBOSTokenLogitsProcessor
-    - __call__
-
-[[autodoc]] ForcedEOSTokenLogitsProcessor
-    - __call__
-
-[[autodoc]] InfNanRemoveLogitsProcessor
-    - __call__
-
-[[autodoc]] TFLogitsProcessor
-    - __call__
-
-[[autodoc]] TFLogitsProcessorList
-    - __call__
-
-[[autodoc]] TFLogitsWarper
-    - __call__
-
-[[autodoc]] TFTemperatureLogitsWarper
-    - __call__
-
-[[autodoc]] TFTopPLogitsWarper
-    - __call__
-
-[[autodoc]] TFTopKLogitsWarper
-    - __call__
-
-[[autodoc]] TFMinLengthLogitsProcessor
-    - __call__
-
-[[autodoc]] TFNoBadWordsLogitsProcessor
-    - __call__
-
-[[autodoc]] TFNoRepeatNGramLogitsProcessor
-    - __call__
-
-[[autodoc]] TFRepetitionPenaltyLogitsProcessor
-    - __call__
-
-[[autodoc]] TFForcedBOSTokenLogitsProcessor
-    - __call__
-
-[[autodoc]] TFForcedEOSTokenLogitsProcessor
-    - __call__
-
-[[autodoc]] FlaxLogitsProcessor
-    - __call__
-
-[[autodoc]] FlaxLogitsProcessorList
-    - __call__
-
-[[autodoc]] FlaxLogitsWarper
-    - __call__
-
-[[autodoc]] FlaxTemperatureLogitsWarper
-    - __call__
-
-[[autodoc]] FlaxTopPLogitsWarper
-    - __call__
-
-[[autodoc]] FlaxTopKLogitsWarper
-    - __call__
-
-[[autodoc]] FlaxForcedBOSTokenLogitsProcessor
-    - __call__
-
-[[autodoc]] FlaxForcedEOSTokenLogitsProcessor
-    - __call__
-
-[[autodoc]] FlaxMinLengthLogitsProcessor
-    - __call__
-
-## StoppingCriteria
-
-A [`StoppingCriteria`] can be used to change when to stop generation (other than EOS token).
-
-[[autodoc]] StoppingCriteria
-    - __call__
-
-[[autodoc]] StoppingCriteriaList
-    - __call__
-
-[[autodoc]] MaxLengthCriteria
-    - __call__
-
-[[autodoc]] MaxTimeCriteria
-    - __call__
-
-## Constraints
-
-A [`Constraint`] can be used to force the generation to include specific tokens or sequences in the output.
-
-[[autodoc]] Constraint
-
-[[autodoc]] PhrasalConstraint
-
-[[autodoc]] DisjunctiveConstraint
-
-[[autodoc]] ConstraintListState
-
-## BeamSearch
-
-[[autodoc]] BeamScorer
-    - process
-    - finalize
-
-[[autodoc]] BeamSearchScorer
-    - process
-    - finalize
-
-[[autodoc]] ConstrainedBeamSearchScorer
-    - process
-    - finalize
-
-## Utilities
-
-[[autodoc]] top_k_top_p_filtering
-
-[[autodoc]] tf_top_k_top_p_filtering

docs/source/en/internal/modeling_utils.mdx

@@ -1,82 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Custom Layers and Utilities
-
-This page lists all the custom layers used by the library, as well as the utility functions it provides for modeling.
-
-Most of those are only useful if you are studying the code of the models in the library.
-
-
-## Pytorch custom modules
-
-[[autodoc]] pytorch_utils.Conv1D
-
-[[autodoc]] modeling_utils.PoolerStartLogits
-    - forward
-
-[[autodoc]] modeling_utils.PoolerEndLogits
-    - forward
-
-[[autodoc]] modeling_utils.PoolerAnswerClass
-    - forward
-
-[[autodoc]] modeling_utils.SquadHeadOutput
-
-[[autodoc]] modeling_utils.SQuADHead
-    - forward
-
-[[autodoc]] modeling_utils.SequenceSummary
-    - forward
-
-## PyTorch Helper Functions
-
-[[autodoc]] pytorch_utils.apply_chunking_to_forward
-
-[[autodoc]] pytorch_utils.find_pruneable_heads_and_indices
-
-[[autodoc]] pytorch_utils.prune_layer
-
-[[autodoc]] pytorch_utils.prune_conv1d_layer
-
-[[autodoc]] pytorch_utils.prune_linear_layer
-
-## TensorFlow custom layers
-
-[[autodoc]] modeling_tf_utils.TFConv1D
-
-[[autodoc]] modeling_tf_utils.TFSharedEmbeddings
-    - call
-
-[[autodoc]] modeling_tf_utils.TFSequenceSummary
-
-## TensorFlow loss functions
-
-[[autodoc]] modeling_tf_utils.TFCausalLanguageModelingLoss
-
-[[autodoc]] modeling_tf_utils.TFMaskedLanguageModelingLoss
-
-[[autodoc]] modeling_tf_utils.TFMultipleChoiceLoss
-
-[[autodoc]] modeling_tf_utils.TFQuestionAnsweringLoss
-
-[[autodoc]] modeling_tf_utils.TFSequenceClassificationLoss
-
-[[autodoc]] modeling_tf_utils.TFTokenClassificationLoss
-
-## TensorFlow Helper Functions
-
-[[autodoc]] modeling_tf_utils.get_initializer
-
-[[autodoc]] modeling_tf_utils.keras_serializable
-
-[[autodoc]] modeling_tf_utils.shape_list

docs/source/en/internal/pipelines_utils.mdx

@@ -1,40 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Utilities for pipelines
-
-This page lists all the utility functions the library provides for pipelines.
-
-Most of those are only useful if you are studying the code of the models in the library.
-
-
-## Argument handling
-
-[[autodoc]] pipelines.ArgumentHandler
-
-[[autodoc]] pipelines.ZeroShotClassificationArgumentHandler
-
-[[autodoc]] pipelines.QuestionAnsweringArgumentHandler
-
-## Data format
-
-[[autodoc]] pipelines.PipelineDataFormat
-
-[[autodoc]] pipelines.CsvPipelineDataFormat
-
-[[autodoc]] pipelines.JsonPipelineDataFormat
-
-[[autodoc]] pipelines.PipedPipelineDataFormat
-
-## Utilities
-
-[[autodoc]] pipelines.PipelineException

docs/source/en/internal/tokenization_utils.mdx

@@ -1,38 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Utilities for Tokenizers
-
-This page lists all the utility functions used by the tokenizers, mainly the class
-[`~tokenization_utils_base.PreTrainedTokenizerBase`] that implements the common methods between
-[`PreTrainedTokenizer`] and [`PreTrainedTokenizerFast`] and the mixin
-[`~tokenization_utils_base.SpecialTokensMixin`].
-
-Most of those are only useful if you are studying the code of the tokenizers in the library.
-
-## PreTrainedTokenizerBase
-
-[[autodoc]] tokenization_utils_base.PreTrainedTokenizerBase
-    - __call__
-    - all
-
-## SpecialTokensMixin
-
-[[autodoc]] tokenization_utils_base.SpecialTokensMixin
-
-## Enums and namedtuples
-
-[[autodoc]] tokenization_utils_base.TruncationStrategy
-
-[[autodoc]] tokenization_utils_base.CharSpan
-
-[[autodoc]] tokenization_utils_base.TokenSpan

docs/source/en/internal/trainer_utils.mdx

@@ -1,45 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# Utilities for Trainer
-
-This page lists all the utility functions used by [`Trainer`].
-
-Most of those are only useful if you are studying the code of the Trainer in the library.
-
-## Utilities
-
-[[autodoc]] EvalPrediction
-
-[[autodoc]] IntervalStrategy
-
-[[autodoc]] enable_full_determinism
-
-[[autodoc]] set_seed
-
-[[autodoc]] torch_distributed_zero_first
-
-## Callbacks internals
-
-[[autodoc]] trainer_callback.CallbackHandler
-
-## Distributed Evaluation
-
-[[autodoc]] trainer_pt_utils.DistributedTensorGatherer
-
-## Distributed Evaluation
-
-[[autodoc]] HfArgumentParser
-
-## Debug Utilities
-
-[[autodoc]] debug_utils.DebugUnderflowOverflow