Render custom tool docs a bit better (#23269)

* Try on a couple of blocks to see

* Build the doc please

* Build the doc please

* Build the doc please

* add more

* Finish with all

* Style

.circleci/create_circleci_config.py

@@ -51,8 +51,6 @@ class CircleCIJob:
     resource_class: Optional[str] = "xlarge"
     tests_to_run: Optional[List[str]] = None
     working_directory: str = "~/transformers"
-    # This should be only used for doctest job!
-    command_timeout: Optional[int] = None
 
     def __post_init__(self):
         # Deal with defaults for mutable attributes.
@@ -109,15 +107,11 @@ class CircleCIJob:
         steps.append({"store_artifacts": {"path": "~/transformers/installed.txt"}})
 
         all_options = {**COMMON_PYTEST_OPTIONS, **self.pytest_options}
-        pytest_flags = [f"--{key}={value}" if (value is not None or key in ["doctest-modules"]) else f"-{key}" for key, value in all_options.items()]
+        pytest_flags = [f"--{key}={value}" if value is not None else f"-{key}" for key, value in all_options.items()]
         pytest_flags.append(
             f"--make-reports={self.name}" if "examples" in self.name else f"--make-reports=tests_{self.name}"
         )
-        test_command = ""
-        if self.command_timeout:
-            test_command = f"timeout {self.command_timeout} "
-        test_command += f"python -m pytest -n {self.pytest_num_workers} " + " ".join(pytest_flags)
-        
+        test_command = f"python -m pytest -n {self.pytest_num_workers} " + " ".join(pytest_flags)
         if self.parallelism == 1:
             if self.tests_to_run is None:
                 test_command += " << pipeline.parameters.tests_to_run >>"
@@ -167,37 +161,12 @@ class CircleCIJob:
             steps.append({"store_artifacts": {"path": "~/transformers/tests.txt"}})
             steps.append({"store_artifacts": {"path": "~/transformers/splitted_tests.txt"}})
 
-            test_command = ""
-            if self.timeout:
-                test_command = f"timeout {self.timeout} "
-            test_command += f"python -m pytest -n {self.pytest_num_workers} " + " ".join(pytest_flags)
+            test_command = f"python -m pytest -n {self.pytest_num_workers} " + " ".join(pytest_flags)
             test_command += " $(cat splitted_tests.txt)"
         if self.marker is not None:
             test_command += f" -m {self.marker}"
-
-        if self.name == "pr_documentation_tests":
-            # can't use ` | tee tee tests_output.txt` as usual
-            test_command += " > tests_output.txt"
-            # Save the return code, so we can check if it is timeout in the next step.
-            test_command += '; touch "$?".txt'
-            # Never fail the test step for the doctest job. We will check the results in the next step, and fail that
-            # step instead if the actual test failures are found. This is to avoid the timeout being reported as test
-            # failure.
-            test_command = f"({test_command}) || true"
-        else:
-            test_command += " | tee tests_output.txt"
+        test_command += " | tee tests_output.txt"
         steps.append({"run": {"name": "Run tests", "command": test_command}})
-
-        # return code `124` means the previous (pytest run) step is timeout
-        if self.name == "pr_documentation_tests":
-            checkout_doctest_command = 'if [ -s reports/tests_pr_documentation_tests/failures_short.txt ]; '
-            checkout_doctest_command += 'then echo "some test failed"; '
-            checkout_doctest_command += 'cat reports/tests_pr_documentation_tests/failures_short.txt; '
-            checkout_doctest_command += 'cat reports/tests_pr_documentation_tests/summary_short.txt; exit -1; '
-            checkout_doctest_command += 'elif [ -s reports/tests_pr_documentation_tests/stats.txt ]; then echo "All tests pass!"; '
-            checkout_doctest_command += 'elif [ -f 124.txt ]; then echo "doctest timeout!"; else echo "other fatal error)"; exit -1; fi;'
-            steps.append({"run": {"name": "Check doctest results", "command": checkout_doctest_command}})
-
         steps.append({"store_artifacts": {"path": "~/transformers/tests_output.txt"}})
         steps.append({"store_artifacts": {"path": "~/transformers/reports"}})
         job["steps"] = steps
@@ -217,7 +186,7 @@ torch_and_tf_job = CircleCIJob(
         "git lfs install",
         "pip install --upgrade pip",
         "pip install .[sklearn,tf-cpu,torch,testing,sentencepiece,torch-speech,vision]",
-        "pip install tensorflow_probability",
+        'pip install "tensorflow_probability<0.20"',
         "pip install git+https://github.com/huggingface/accelerate",
     ],
     marker="is_pt_tf_cross_test",
@@ -258,7 +227,7 @@ tf_job = CircleCIJob(
         "sudo apt-get -y update && sudo apt-get install -y libsndfile1-dev espeak-ng cmake",
         "pip install --upgrade pip",
         "pip install .[sklearn,tf-cpu,testing,sentencepiece,tf-speech,vision]",
-        "pip install tensorflow_probability",
+        'pip install "tensorflow_probability<0.20"',
     ],
     parallelism=1,
     pytest_options={"rA": None},
@@ -297,7 +266,7 @@ pipelines_tf_job = CircleCIJob(
         "sudo apt-get -y update && sudo apt-get install -y cmake",
         "pip install --upgrade pip",
         "pip install .[sklearn,tf-cpu,testing,sentencepiece,vision]",
-        "pip install tensorflow_probability",
+        'pip install "tensorflow_probability<0.20"',
     ],
     pytest_options={"rA": None},
     marker="is_pipeline_test",
@@ -432,47 +401,6 @@ repo_utils_job = CircleCIJob(
     tests_to_run="tests/repo_utils",
 )
 
-
-# We also include a `dummy.py` file in the files to be doc-tested to prevent edge case failure. Otherwise, the pytest
-# hangs forever during test collection while showing `collecting 0 items / 21 errors`. (To see this, we have to remove
-# the bash output redirection.)
-py_command = 'from utils.tests_fetcher import get_doctest_files; to_test = get_doctest_files() + ["dummy.py"]; to_test = " ".join(to_test); print(to_test)'
-py_command = f"$(python3 -c '{py_command}')"
-command = f'echo "{py_command}" > pr_documentation_tests_temp.txt'
-doc_test_job = CircleCIJob(
-    "pr_documentation_tests",
-    additional_env={"TRANSFORMERS_VERBOSITY": "error", "DATASETS_VERBOSITY": "error", "SKIP_CUDA_DOCTEST": "1"},
-    install_steps=[
-        "sudo apt-get -y update && sudo apt-get install -y libsndfile1-dev espeak-ng time",
-        "pip install --upgrade pip",
-        "pip install -e .[dev]",
-        "pip install git+https://github.com/huggingface/accelerate",
-        "pip install --upgrade pytest pytest-sugar",
-        "find -name __pycache__ -delete",
-        "find . -name \*.pyc -delete",
-        # Add an empty file to keep the test step running correctly even no file is selected to be tested.
-        "touch dummy.py",
-        {
-            "name": "Get files to test",
-            "command": command,
-        },
-        {
-            "name": "Show information in `Get files to test`",
-            "command":
-                "cat pr_documentation_tests_temp.txt"
-        },
-        {
-            "name": "Get the last line in `pr_documentation_tests.txt`",
-            "command":
-                "tail -n1 pr_documentation_tests_temp.txt | tee pr_documentation_tests.txt"
-        },
-    ],
-    tests_to_run="$(cat pr_documentation_tests.txt)",  # noqa
-    pytest_options={"-doctest-modules": None, "doctest-glob": "*.mdx", "dist": "loadfile", "rvsA": None},
-    command_timeout=1200,  # test cannot run longer than 1200 seconds
-    pytest_num_workers=1,
-)
-
 REGULAR_TESTS = [
     torch_and_tf_job,
     torch_and_flax_job,
@@ -483,7 +411,6 @@ REGULAR_TESTS = [
     hub_job,
     onnx_job,
     exotic_models_job,
-    doc_test_job
 ]
 EXAMPLES_TESTS = [
     examples_torch_job,

.github/workflows/doctests.yml

@@ -25,17 +25,11 @@ jobs:
       image: huggingface/transformers-all-latest-gpu
       options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
     steps:
-      - name: uninstall transformers (installed during docker image build)
-        run: python3 -m pip uninstall -y transformers
-
       - uses: actions/checkout@v3
       - name: NVIDIA-SMI
         run: |
           nvidia-smi
 
-      - name: Install transformers in edit mode
-        run: python3 -m pip install -e .
-
       - name: GPU visibility
         run: |
           python3 utils/print_env.py
@@ -43,10 +37,18 @@ jobs:
       - name: Show installed libraries and their versions
         run: pip freeze
 
+      - 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

Makefile

@@ -47,10 +47,10 @@ repo-consistency:
 # this target runs checks on all files
 
 quality:
-	black --check $(check_dirs) setup.py conftest.py
+	black --check $(check_dirs) setup.py
 	python utils/custom_init_isort.py --check_only
 	python utils/sort_auto_mappings.py --check_only
-	ruff $(check_dirs) setup.py conftest.py
+	ruff $(check_dirs) setup.py
 	doc-builder style src/transformers docs/source --max_len 119 --check_only --path_to_docs docs/source
 	python utils/check_doc_toc.py
 
@@ -65,8 +65,8 @@ extra_style_checks:
 # this target runs checks on all files and potentially modifies some of them
 
 style:
-	black $(check_dirs) setup.py conftest.py
-	ruff $(check_dirs) setup.py conftest.py --fix
+	black $(check_dirs) setup.py
+	ruff $(check_dirs) setup.py --fix
 	${MAKE} autogenerate_code
 	${MAKE} extra_style_checks
 

README.md

@@ -432,7 +432,6 @@ Current number of checkpoints: ![](https://img.shields.io/endpoint?url=https://h
 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. **[SwiftFormer](https://huggingface.co/docs/transformers/main/model_doc/swiftformer)** (from MBZUAI) released with the paper [SwiftFormer: Efficient Additive Attention for Transformer-based Real-time Mobile Vision Applications](https://arxiv.org/abs/2303.15446) by Abdelrahman Shaker, Muhammad Maaz, Hanoona Rasheed, Salman Khan, Ming-Hsuan Yang, Fahad Shahbaz Khan.
 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/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. **[Swin2SR](https://huggingface.co/docs/transformers/model_doc/swin2sr)** (from University of Würzburg) released with the paper [Swin2SR: SwinV2 Transformer for Compressed Image Super-Resolution and Restoration](https://arxiv.org/abs/2209.11345) by Marcos V. Conde, Ui-Jin Choi, Maxime Burchi, Radu Timofte.

README_es.md

@@ -329,7 +329,7 @@ Número actual de puntos de control: ![](https://img.shields.io/endpoint?url=htt
 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. **[FocalNet](https://huggingface.co/docs/transformers/model_doc/focalnet)** (from Microsoft Research) released with the paper [Focal Modulation Networks](https://arxiv.org/abs/2203.11926) by Jianwei Yang, Chunyuan Li, Xiyang Dai, Lu Yuan, Jianfeng Gao.
+1. **[FocalNet](https://huggingface.co/docs/transformers/main/model_doc/focalnet)** (from Microsoft Research) released with the paper [Focal Modulation Networks](https://arxiv.org/abs/2203.11926) by Jianwei Yang, Chunyuan Li, Xiyang Dai, Lu Yuan, Jianfeng Gao.
 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. **[GIT](https://huggingface.co/docs/transformers/model_doc/git)** (from Microsoft Research) released with the paper [GIT: A Generative Image-to-text Transformer for Vision and Language](https://arxiv.org/abs/2205.14100) by Jianfeng Wang, Zhengyuan Yang, Xiaowei Hu, Linjie Li, Kevin Lin, Zhe Gan, Zicheng Liu, Ce Liu, Lijuan Wang.
 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.
@@ -388,7 +388,7 @@ Número actual de puntos de control: ![](https://img.shields.io/endpoint?url=htt
 1. **[NLLB-MOE](https://huggingface.co/docs/transformers/model_doc/nllb-moe)** (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. **[OneFormer](https://huggingface.co/docs/transformers/model_doc/oneformer)** (from SHI Labs) released with the paper [OneFormer: One Transformer to Rule Universal Image Segmentation](https://arxiv.org/abs/2211.06220) by Jitesh Jain, Jiachen Li, MangTik Chiu, Ali Hassani, Nikita Orlov, Humphrey Shi.
-1. **[OpenLlama](https://huggingface.co/docs/transformers/model_doc/open-llama)** (from [s-JoL](https://huggingface.co/s-JoL)) released in [Open-Llama](https://github.com/s-JoL/Open-Llama).
+1. **[OpenLlama](https://huggingface.co/docs/transformers/main/model_doc/open-llama)** (from [s-JoL](https://huggingface.co/s-JoL)) released in [Open-Llama](https://github.com/s-JoL/Open-Llama).
 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.
@@ -410,9 +410,9 @@ Número actual de puntos de control: ![](https://img.shields.io/endpoint?url=htt
 1. **[RoBERTa-PreLayerNorm](https://huggingface.co/docs/transformers/model_doc/roberta-prelayernorm)** (from Facebook) released with the paper [fairseq: A Fast, Extensible Toolkit for Sequence Modeling](https://arxiv.org/abs/1904.01038) by Myle Ott, Sergey Edunov, Alexei Baevski, Angela Fan, Sam Gross, Nathan Ng, David Grangier, Michael Auli.
 1. **[RoCBert](https://huggingface.co/docs/transformers/model_doc/roc_bert)** (from WeChatAI) released with the paper [RoCBert: Robust Chinese Bert with Multimodal Contrastive Pretraining](https://aclanthology.org/2022.acl-long.65.pdf) by HuiSu, WeiweiShi, XiaoyuShen, XiaoZhou, TuoJi, JiaruiFang, JieZhou.
 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. **[RWKV](https://huggingface.co/docs/transformers/model_doc/rwkv)** (from Bo Peng) released with the paper [this repo](https://github.com/BlinkDL/RWKV-LM) by Bo Peng.
+1. **[RWKV](https://huggingface.co/docs/transformers/main/model_doc/rwkv)** (from Bo Peng) released with the paper [this repo](https://github.com/BlinkDL/RWKV-LM) by Bo Peng.
 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. **[Segment Anything](https://huggingface.co/docs/transformers/model_doc/sam)** (from Meta AI) released with the paper [Segment Anything](https://arxiv.org/pdf/2304.02643v1.pdf) by Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alex Berg, Wan-Yen Lo, Piotr Dollar, Ross Girshick.
+1. **[Segment Anything](https://huggingface.co/docs/transformers/main/model_doc/sam)** (from Meta AI) released with the paper [Segment Anything](https://arxiv.org/pdf/2304.02643v1.pdf) by Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alex Berg, Wan-Yen Lo, Piotr Dollar, Ross Girshick.
 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. **[SpeechT5](https://huggingface.co/docs/transformers/model_doc/speecht5)** (from Microsoft Research) released with the paper [SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing](https://arxiv.org/abs/2110.07205) by Junyi Ao, Rui Wang, Long Zhou, Chengyi Wang, Shuo Ren, Yu Wu, Shujie Liu, Tom Ko, Qing Li, Yu Zhang, Zhihua Wei, Yao Qian, Jinyu Li, Furu Wei.
@@ -420,7 +420,6 @@ Número actual de puntos de control: ![](https://img.shields.io/endpoint?url=htt
 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. **[SwiftFormer](https://huggingface.co/docs/transformers/main/model_doc/swiftformer)** (from MBZUAI) released with the paper [SwiftFormer: Efficient Additive Attention for Transformer-based Real-time Mobile Vision Applications](https://arxiv.org/abs/2303.15446) by Abdelrahman Shaker, Muhammad Maaz, Hanoona Rasheed, Salman Khan, Ming-Hsuan Yang, Fahad Shahbaz Khan.
 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/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. **[Swin2SR](https://huggingface.co/docs/transformers/model_doc/swin2sr)** (from University of Würzburg) released with the paper [Swin2SR: SwinV2 Transformer for Compressed Image Super-Resolution and Restoration](https://arxiv.org/abs/2209.11345) by Marcos V. Conde, Ui-Jin Choi, Maxime Burchi, Radu Timofte.

README_hd.md

@@ -301,7 +301,7 @@ conda install -c huggingface transformers
 1. **[FlauBERT](https://huggingface.co/docs/transformers/model_doc/flaubert)** (CNRS से) साथ वाला पेपर [FlauBERT: Unsupervised Language Model Pre-training for फ़्रेंच](https://arxiv .org/abs/1912.05372) Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, बेंजामिन लेकोउटेक्स, अलेक्जेंड्रे अल्लाउज़ेन, बेनोइट क्रैबे, लॉरेंट बेसेसियर, डिडिएर श्वाब द्वारा।
 1. **[FLAVA](https://huggingface.co/docs/transformers/model_doc/flava)** (FLAVA: A फाउंडेशनल लैंग्वेज एंड विजन अलाइनमेंट मॉडल) (https://arxiv) साथ वाला पेपर .org/abs/2112.04482) अमनप्रीत सिंह, रोंगहांग हू, वेदानुज गोस्वामी, गुइल्यूम कुएरॉन, वोज्शिएक गालुबा, मार्कस रोहरबैक, और डौवे कीला द्वारा।
 1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (गूगल रिसर्च से) साथ वाला पेपर [FNet: मिक्सिंग टोकन विद फूरियर ट्रांसफॉर्म्स](https://arxiv.org /abs/2105.03824) जेम्स ली-थॉर्प, जोशुआ आइंस्ली, इल्या एकस्टीन, सैंटियागो ओंटानन द्वारा।
-1. **[FocalNet](https://huggingface.co/docs/transformers/model_doc/focalnet)** (Microsoft Research से) Jianwei Yang, Chunyuan Li, Xiyang Dai, Lu Yuan, Jianfeng Gao. द्वाराअनुसंधान पत्र [Focal Modulation Networks](https://arxiv.org/abs/2203.11926) के साथ जारी किया गया
+1. **[FocalNet](https://huggingface.co/docs/transformers/main/model_doc/focalnet)** (Microsoft Research से) Jianwei Yang, Chunyuan Li, Xiyang Dai, Lu Yuan, Jianfeng Gao. द्वाराअनुसंधान पत्र [Focal Modulation Networks](https://arxiv.org/abs/2203.11926) के साथ जारी किया गया
 1. **[Funnel Transformer](https://huggingface.co/docs/transformers/model_doc/funnel)** (सीएमयू/गूगल ब्रेन से) साथ में कागज [फ़नल-ट्रांसफॉर्मर: कुशल भाषा प्रसंस्करण के लिए अनुक्रमिक अतिरेक को छानना](https://arxiv.org/abs/2006.03236) जिहांग दाई, गुओकुन लाई, यिमिंग यांग, क्वोक वी. ले ​​द्वारा रिहाई।
 1. **[GIT](https://huggingface.co/docs/transformers/model_doc/git)** (from Microsoft Research) released with the paper [GIT: A Generative Image-to-text Transformer for Vision and Language](https://arxiv.org/abs/2205.14100) by Jianfeng Wang, Zhengyuan Yang, Xiaowei Hu, Linjie Li, Kevin Lin, Zhe Gan, Zicheng Liu, Ce Liu, Lijuan Wang.
 1. **[GLPN](https://huggingface.co/docs/transformers/model_doc/glpn)** (KAIST से) साथ वाला पेपर [वर्टिकल कटडेप्थ के साथ मोनोकुलर डेप्थ एस्टीमेशन के लिए ग्लोबल-लोकल पाथ नेटवर्क्स](https:/ /arxiv.org/abs/2201.07436) डोयोन किम, वूंगह्युन गा, प्युंगवान आह, डोंगग्यू जू, सेहवान चुन, जुनमो किम द्वारा।
@@ -360,7 +360,7 @@ conda install -c huggingface transformers
 1. **[NLLB-MOE](https://huggingface.co/docs/transformers/model_doc/nllb-moe)** (Meta से) the NLLB team. द्वाराअनुसंधान पत्र [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) के साथ जारी किया गया
 1. **[Nyströmformer](https://huggingface.co/docs/transformers/model_doc/nystromformer)** (विस्कॉन्सिन विश्वविद्यालय - मैडिसन से) साथ में कागज [Nyströmformer: A Nyström- आधारित एल्गोरिथम आत्म-ध्यान का अनुमान लगाने के लिए ](https://arxiv.org/abs/2102.03902) युनयांग ज़िओंग, झानपेंग ज़ेंग, रुद्रसिस चक्रवर्ती, मिंगक्सिंग टैन, ग्लेन फंग, यिन ली, विकास सिंह द्वारा पोस्ट किया गया।
 1. **[OneFormer](https://huggingface.co/docs/transformers/model_doc/oneformer)** (SHI Labs से) पेपर [OneFormer: One Transformer to Rule Universal Image Segmentation](https://arxiv.org/abs/2211.06220) जितेश जैन, जिआचेन ली, मांगटिक चिउ, अली हसनी, निकिता ओरलोव, हम्फ्री शि के द्वारा जारी किया गया है।
-1. **[OpenLlama](https://huggingface.co/docs/transformers/model_doc/open-llama)** (from [s-JoL](https://huggingface.co/s-JoL)) released in [Open-Llama](https://github.com/s-JoL/Open-Llama). 
+1. **[OpenLlama](https://huggingface.co/docs/transformers/main/model_doc/open-llama)** (from [s-JoL](https://huggingface.co/s-JoL)) released in [Open-Llama](https://github.com/s-JoL/Open-Llama). 
 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)** (Google AI से) साथ में कागज [विज़न ट्रांसफॉर्मर्स के साथ सिंपल ओपन-वोकैबुलरी ऑब्जेक्ट डिटेक्शन](https:/ /arxiv.org/abs/2205.06230) मैथियास मिंडरर, एलेक्सी ग्रिट्सेंको, ऑस्टिन स्टोन, मैक्सिम न्यूमैन, डिर्क वीसेनबोर्न, एलेक्सी डोसोवित्स्की, अरविंद महेंद्रन, अनुराग अर्नब, मुस्तफा देहघानी, ज़ुओरन शेन, जिओ वांग, ज़ियाओहुआ झाई, थॉमस किफ़, और नील हॉल्सबी द्वारा पोस्ट किया गया।
 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.
@@ -382,9 +382,9 @@ conda install -c huggingface transformers
 1. **[RoBERTa-PreLayerNorm](https://huggingface.co/docs/transformers/model_doc/roberta-prelayernorm)** (from Facebook) released with the paper [fairseq: A Fast, Extensible Toolkit for Sequence Modeling](https://arxiv.org/abs/1904.01038) by Myle Ott, Sergey Edunov, Alexei Baevski, Angela Fan, Sam Gross, Nathan Ng, David Grangier, Michael Auli.
 1. **[RoCBert](https://huggingface.co/docs/transformers/model_doc/roc_bert)** (from WeChatAI) released with the paper [RoCBert: Robust Chinese Bert with Multimodal Contrastive Pretraining](https://aclanthology.org/2022.acl-long.65.pdf) by HuiSu, WeiweiShi, XiaoyuShen, XiaoZhou, TuoJi, JiaruiFang, JieZhou.
 1. **[RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer)** (झुईई टेक्नोलॉजी से), साथ में पेपर [रोफॉर्मर: रोटरी पोजिशन एंबेडिंग के साथ एन्हांस्ड ट्रांसफॉर्मर] (https://arxiv.org/pdf/2104.09864v1.pdf) जियानलिन सु और यू लू और शेंगफेंग पैन और बो वेन और युनफेंग लियू द्वारा प्रकाशित।
-1. **[RWKV](https://huggingface.co/docs/transformers/model_doc/rwkv)** (Bo Peng से) Bo Peng. द्वाराअनुसंधान पत्र [this repo](https://github.com/BlinkDL/RWKV-LM) के साथ जारी किया गया
+1. **[RWKV](https://huggingface.co/docs/transformers/main/model_doc/rwkv)** (Bo Peng से) Bo Peng. द्वाराअनुसंधान पत्र [this repo](https://github.com/BlinkDL/RWKV-LM) के साथ जारी किया गया
 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. **[Segment Anything](https://huggingface.co/docs/transformers/model_doc/sam)** (Meta AI से) Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alex Berg, Wan-Yen Lo, Piotr Dollar, Ross Girshick. द्वाराअनुसंधान पत्र [Segment Anything](https://arxiv.org/pdf/2304.02643v1.pdf) के साथ जारी किया गया
+1. **[Segment Anything](https://huggingface.co/docs/transformers/main/model_doc/sam)** (Meta AI से) Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alex Berg, Wan-Yen Lo, Piotr Dollar, Ross Girshick. द्वाराअनुसंधान पत्र [Segment Anything](https://arxiv.org/pdf/2304.02643v1.pdf) के साथ जारी किया गया
 1. **[SEW](https://huggingface.co/docs/transformers/model_doc/sew)** (ASAPP से) साथ देने वाला पेपर [भाषण पहचान के लिए अनसुपरवाइज्ड प्री-ट्रेनिंग में परफॉर्मेंस-एफिशिएंसी ट्रेड-ऑफ्स](https ://arxiv.org/abs/2109.06870) फेलिक्स वू, क्वांगयुन किम, जिंग पैन, क्यू हान, किलियन क्यू. वेनबर्गर, योव आर्टज़ी द्वारा।
 1. **[SEW-D](https://huggingface.co/docs/transformers/model_doc/sew_d)** (ASAPP से) साथ में पेपर [भाषण पहचान के लिए अनसुपरवाइज्ड प्री-ट्रेनिंग में परफॉर्मेंस-एफिशिएंसी ट्रेड-ऑफ्स] (https://arxiv.org/abs/2109.06870) फेलिक्स वू, क्वांगयुन किम, जिंग पैन, क्यू हान, किलियन क्यू. वेनबर्गर, योआव आर्टज़ी द्वारा पोस्ट किया गया।
 1. **[SpeechT5](https://huggingface.co/docs/transformers/model_doc/speecht5)** (from Microsoft Research) released with the paper [SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing](https://arxiv.org/abs/2110.07205) by Junyi Ao, Rui Wang, Long Zhou, Chengyi Wang, Shuo Ren, Yu Wu, Shujie Liu, Tom Ko, Qing Li, Yu Zhang, Zhihua Wei, Yao Qian, Jinyu Li, Furu Wei.
@@ -392,7 +392,6 @@ conda install -c huggingface transformers
 1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (फेसबुक से) साथ में पेपर [लार्ज-स्केल सेल्फ- एंड सेमी-सुपरवाइज्ड लर्निंग फॉर स्पीच ट्रांसलेशन](https://arxiv.org/abs/2104.06678) चांगहान वांग, ऐनी वू, जुआन पिनो, एलेक्सी बेवस्की, माइकल औली, एलेक्सिस द्वारा Conneau द्वारा पोस्ट किया गया।
 1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (तेल अवीव यूनिवर्सिटी से) साथ में पेपर [स्पैन सिलेक्शन को प्री-ट्रेनिंग करके कुछ-शॉट क्वेश्चन आंसरिंग](https:// arxiv.org/abs/2101.00438) ओरि राम, युवल कर्स्टन, जोनाथन बेरेंट, अमीर ग्लोबर्सन, ओमर लेवी द्वारा।
 1. **[SqueezeBERT](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (बर्कले से) कागज के साथ [SqueezeBERT: कुशल तंत्रिका नेटवर्क के बारे में NLP को कंप्यूटर विज़न क्या सिखा सकता है?](https: //arxiv.org/abs/2006.11316) फॉरेस्ट एन. इनडोला, अल्बर्ट ई. शॉ, रवि कृष्णा, और कर्ट डब्ल्यू. केटज़र द्वारा।
-1. **[SwiftFormer](https://huggingface.co/docs/transformers/main/model_doc/swiftformer)** (MBZUAI से) Abdelrahman Shaker, Muhammad Maaz, Hanoona Rasheed, Salman Khan, Ming-Hsuan Yang, Fahad Shahbaz Khan. द्वाराअनुसंधान पत्र [SwiftFormer: Efficient Additive Attention for Transformer-based Real-time Mobile Vision Applications](https://arxiv.org/abs/2303.15446) के साथ जारी किया गया
 1. **[Swin Transformer](https://huggingface.co/docs/transformers/model_doc/swin)** (माइक्रोसॉफ्ट से) साथ में कागज [स्वाइन ट्रांसफॉर्मर: शिफ्टेड विंडोज का उपयोग कर पदानुक्रमित विजन ट्रांसफॉर्मर](https://arxiv .org/abs/2103.14030) ज़ी लियू, युटोंग लिन, यू काओ, हान हू, यिक्सुआन वेई, झेंग झांग, स्टीफन लिन, बैनिंग गुओ द्वारा।
 1. **[Swin Transformer V2](https://huggingface.co/docs/transformers/model_doc/swinv2)** (Microsoft से) साथ वाला पेपर [Swin Transformer V2: स्केलिंग अप कैपेसिटी एंड रेजोल्यूशन](https:// ज़ी लियू, हान हू, युटोंग लिन, ज़ुलिआंग याओ, ज़ेंडा ज़ी, यिक्सुआन वेई, जिया निंग, यू काओ, झेंग झांग, ली डोंग, फुरु वेई, बैनिंग गुओ द्वारा arxiv.org/abs/2111.09883।
 1. **[Swin2SR](https://huggingface.co/docs/transformers/model_doc/swin2sr)** (from University of Würzburg) released with the paper [Swin2SR: SwinV2 Transformer for Compressed Image Super-Resolution and Restoration](https://arxiv.org/abs/2209.11345) by Marcos V. Conde, Ui-Jin Choi, Maxime Burchi, Radu Timofte.

README_ja.md

@@ -363,7 +363,7 @@ Flax、PyTorch、TensorFlowをcondaでインストールする方法は、それ
 1. **[FlauBERT](https://huggingface.co/docs/transformers/model_doc/flaubert)** (CNRS から) Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab から公開された研究論文: [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372)
 1. **[FLAVA](https://huggingface.co/docs/transformers/model_doc/flava)** (Facebook AI から) Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela から公開された研究論文: [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482)
 1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (Google Research から) James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon から公開された研究論文: [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824)
-1. **[FocalNet](https://huggingface.co/docs/transformers/model_doc/focalnet)** (Microsoft Research から) Jianwei Yang, Chunyuan Li, Xiyang Dai, Lu Yuan, Jianfeng Gao. から公開された研究論文 [Focal Modulation Networks](https://arxiv.org/abs/2203.11926)
+1. **[FocalNet](https://huggingface.co/docs/transformers/main/model_doc/focalnet)** (Microsoft Research から) Jianwei Yang, Chunyuan Li, Xiyang Dai, Lu Yuan, Jianfeng Gao. から公開された研究論文 [Focal Modulation Networks](https://arxiv.org/abs/2203.11926)
 1. **[Funnel Transformer](https://huggingface.co/docs/transformers/model_doc/funnel)** (CMU/Google Brain から) Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le から公開された研究論文: [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236)
 1. **[GIT](https://huggingface.co/docs/transformers/model_doc/git)** (Microsoft Research から) Jianfeng Wang, Zhengyuan Yang, Xiaowei Hu, Linjie Li, Kevin Lin, Zhe Gan, Zicheng Liu, Ce Liu, Lijuan Wang. から公開された研究論文 [GIT: A Generative Image-to-text Transformer for Vision and Language](https://arxiv.org/abs/2205.14100)
 1. **[GLPN](https://huggingface.co/docs/transformers/model_doc/glpn)** (KAIST から) Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim から公開された研究論文: [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436)
@@ -422,7 +422,7 @@ Flax、PyTorch、TensorFlowをcondaでインストールする方法は、それ
 1. **[NLLB-MOE](https://huggingface.co/docs/transformers/model_doc/nllb-moe)** (Meta から) the NLLB team. から公開された研究論文 [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672)
 1. **[Nyströmformer](https://huggingface.co/docs/transformers/model_doc/nystromformer)** (the University of Wisconsin - Madison から) Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh から公開された研究論文: [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902)
 1. **[OneFormer](https://huggingface.co/docs/transformers/model_doc/oneformer)** (SHI Labs から) Jitesh Jain, Jiachen Li, MangTik Chiu, Ali Hassani, Nikita Orlov, Humphrey Shi から公開された研究論文: [OneFormer: One Transformer to Rule Universal Image Segmentation](https://arxiv.org/abs/2211.06220)
-1. **[OpenLlama](https://huggingface.co/docs/transformers/model_doc/open-llama)** (from [s-JoL](https://huggingface.co/s-JoL)) released in [Open-Llama](https://github.com/s-JoL/Open-Llama). 
+1. **[OpenLlama](https://huggingface.co/docs/transformers/main/model_doc/open-llama)** (from [s-JoL](https://huggingface.co/s-JoL)) released in [Open-Llama](https://github.com/s-JoL/Open-Llama). 
 1. **[OPT](https://huggingface.co/docs/transformers/master/model_doc/opt)** (Meta AI から) Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al から公開された研究論文: [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068)
 1. **[OWL-ViT](https://huggingface.co/docs/transformers/model_doc/owlvit)** (Google AI から) 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 から公開された研究論文: [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230)
 1. **[Pegasus](https://huggingface.co/docs/transformers/model_doc/pegasus)** (Google から) Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu から公開された研究論文: [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777)
@@ -444,9 +444,9 @@ Flax、PyTorch、TensorFlowをcondaでインストールする方法は、それ
 1. **[RoBERTa-PreLayerNorm](https://huggingface.co/docs/transformers/model_doc/roberta-prelayernorm)** (Facebook から) Myle Ott, Sergey Edunov, Alexei Baevski, Angela Fan, Sam Gross, Nathan Ng, David Grangier, Michael Auli から公開された研究論文: [fairseq: A Fast, Extensible Toolkit for Sequence Modeling](https://arxiv.org/abs/1904.01038)
 1. **[RoCBert](https://huggingface.co/docs/transformers/model_doc/roc_bert)** (WeChatAI から) HuiSu, WeiweiShi, XiaoyuShen, XiaoZhou, TuoJi, JiaruiFang, JieZhou から公開された研究論文: [RoCBert: Robust Chinese Bert with Multimodal Contrastive Pretraining](https://aclanthology.org/2022.acl-long.65.pdf)
 1. **[RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer)** (ZhuiyiTechnology から), Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu から公開された研究論文: [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/abs/2104.09864)
-1. **[RWKV](https://huggingface.co/docs/transformers/model_doc/rwkv)** (Bo Peng から) Bo Peng. から公開された研究論文 [this repo](https://github.com/BlinkDL/RWKV-LM)
+1. **[RWKV](https://huggingface.co/docs/transformers/main/model_doc/rwkv)** (Bo Peng から) Bo Peng. から公開された研究論文 [this repo](https://github.com/BlinkDL/RWKV-LM)
 1. **[SegFormer](https://huggingface.co/docs/transformers/model_doc/segformer)** (NVIDIA から) Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo から公開された研究論文: [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203)
-1. **[Segment Anything](https://huggingface.co/docs/transformers/model_doc/sam)** (Meta AI から) Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alex Berg, Wan-Yen Lo, Piotr Dollar, Ross Girshick. から公開された研究論文 [Segment Anything](https://arxiv.org/pdf/2304.02643v1.pdf)
+1. **[Segment Anything](https://huggingface.co/docs/transformers/main/model_doc/sam)** (Meta AI から) Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alex Berg, Wan-Yen Lo, Piotr Dollar, Ross Girshick. から公開された研究論文 [Segment Anything](https://arxiv.org/pdf/2304.02643v1.pdf)
 1. **[SEW](https://huggingface.co/docs/transformers/model_doc/sew)** (ASAPP から) Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi から公開された研究論文: [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870)
 1. **[SEW-D](https://huggingface.co/docs/transformers/model_doc/sew_d)** (ASAPP から) Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi から公開された研究論文: [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870)
 1. **[SpeechT5](https://huggingface.co/docs/transformers/model_doc/speecht5)** (Microsoft Research から) Junyi Ao, Rui Wang, Long Zhou, Chengyi Wang, Shuo Ren, Yu Wu, Shujie Liu, Tom Ko, Qing Li, Yu Zhang, Zhihua Wei, Yao Qian, Jinyu Li, Furu Wei. から公開された研究論文 [SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing](https://arxiv.org/abs/2110.07205)
@@ -454,7 +454,6 @@ Flax、PyTorch、TensorFlowをcondaでインストールする方法は、それ
 1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (Facebook から), Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau から公開された研究論文: [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678)
 1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (Tel Aviv University から), Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy から公開された研究論文: [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438)
 1. **[SqueezeBERT](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (Berkeley から) Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer から公開された研究論文: [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316)
-1. **[SwiftFormer](https://huggingface.co/docs/transformers/main/model_doc/swiftformer)** (MBZUAI から) Abdelrahman Shaker, Muhammad Maaz, Hanoona Rasheed, Salman Khan, Ming-Hsuan Yang, Fahad Shahbaz Khan. から公開された研究論文 [SwiftFormer: Efficient Additive Attention for Transformer-based Real-time Mobile Vision Applications](https://arxiv.org/abs/2303.15446)
 1. **[Swin Transformer](https://huggingface.co/docs/transformers/model_doc/swin)** (Microsoft から) Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo から公開された研究論文: [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030)
 1. **[Swin Transformer V2](https://huggingface.co/docs/transformers/model_doc/swinv2)** (Microsoft から) Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo から公開された研究論文: [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883)
 1. **[Swin2SR](https://huggingface.co/docs/transformers/model_doc/swin2sr)** (University of Würzburg から) Marcos V. Conde, Ui-Jin Choi, Maxime Burchi, Radu Timofte から公開された研究論文: [Swin2SR: SwinV2 Transformer for Compressed Image Super-Resolution and Restoration](https://arxiv.org/abs/2209.11345)

README_ko.md

@@ -278,7 +278,7 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 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. **[FocalNet](https://huggingface.co/docs/transformers/model_doc/focalnet)** (from Microsoft Research) released with the paper [Focal Modulation Networks](https://arxiv.org/abs/2203.11926) by Jianwei Yang, Chunyuan Li, Xiyang Dai, Lu Yuan, Jianfeng Gao.
+1. **[FocalNet](https://huggingface.co/docs/transformers/main/model_doc/focalnet)** (from Microsoft Research) released with the paper [Focal Modulation Networks](https://arxiv.org/abs/2203.11926) by Jianwei Yang, Chunyuan Li, Xiyang Dai, Lu Yuan, Jianfeng Gao.
 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. **[GIT](https://huggingface.co/docs/transformers/model_doc/git)** (from Microsoft Research) released with the paper [GIT: A Generative Image-to-text Transformer for Vision and Language](https://arxiv.org/abs/2205.14100) by Jianfeng Wang, Zhengyuan Yang, Xiaowei Hu, Linjie Li, Kevin Lin, Zhe Gan, Zicheng Liu, Ce Liu, Lijuan Wang.
 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.
@@ -337,7 +337,7 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[NLLB-MOE](https://huggingface.co/docs/transformers/model_doc/nllb-moe)** (Meta 에서 제공)은 the NLLB team.의 [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672)논문과 함께 발표했습니다.
 1. **[Nyströmformer](https://huggingface.co/docs/transformers/model_doc/nystromformer)** (the University of Wisconsin - Madison 에서) Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh 의 [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) 논문과 함께 발표했습니다.
 1. **[OneFormer](https://huggingface.co/docs/transformers/model_doc/oneformer)** (SHI Labs 에서) Jitesh Jain, Jiachen Li, MangTik Chiu, Ali Hassani, Nikita Orlov, Humphrey Shi 의 [OneFormer: One Transformer to Rule Universal Image Segmentation](https://arxiv.org/abs/2211.06220) 논문과 함께 발표했습니다.
-1. **[OpenLlama](https://huggingface.co/docs/transformers/model_doc/open-llama)** (from [s-JoL](https://huggingface.co/s-JoL)) released in [Open-Llama](https://github.com/s-JoL/Open-Llama). 
+1. **[OpenLlama](https://huggingface.co/docs/transformers/main/model_doc/open-llama)** (from [s-JoL](https://huggingface.co/s-JoL)) released in [Open-Llama](https://github.com/s-JoL/Open-Llama). 
 1. **[OPT](https://huggingface.co/docs/transformers/master/model_doc/opt)** (Meta AI 에서) Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al 의 [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) 논문과 함께 발표했습니다.
 1. **[OWL-ViT](https://huggingface.co/docs/transformers/model_doc/owlvit)** (Google AI 에서) 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 의 [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) 논문과 함께 발표했습니다.
 1. **[Pegasus](https://huggingface.co/docs/transformers/model_doc/pegasus)** (Google 에서) Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu 의 [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) 논문과 함께 발표했습니다.
@@ -359,9 +359,9 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[RoBERTa-PreLayerNorm](https://huggingface.co/docs/transformers/model_doc/roberta-prelayernorm)** (Facebook 에서) Myle Ott, Sergey Edunov, Alexei Baevski, Angela Fan, Sam Gross, Nathan Ng, David Grangier, Michael Auli 의 [fairseq: A Fast, Extensible Toolkit for Sequence Modeling](https://arxiv.org/abs/1904.01038) 논문과 함께 발표했습니다.
 1. **[RoCBert](https://huggingface.co/docs/transformers/model_doc/roc_bert)** (WeChatAI 에서) HuiSu, WeiweiShi, XiaoyuShen, XiaoZhou, TuoJi, JiaruiFang, JieZhou 의 [RoCBert: Robust Chinese Bert with Multimodal Contrastive Pretraining](https://aclanthology.org/2022.acl-long.65.pdf) 논문과 함께 발표했습니다.
 1. **[RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer)** (ZhuiyiTechnology 에서) Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu 의 a [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/pdf/2104.09864v1.pdf) 논문과 함께 발표했습니다.
-1. **[RWKV](https://huggingface.co/docs/transformers/model_doc/rwkv)** (Bo Peng 에서 제공)은 Bo Peng.의 [this repo](https://github.com/BlinkDL/RWKV-LM)논문과 함께 발표했습니다.
+1. **[RWKV](https://huggingface.co/docs/transformers/main/model_doc/rwkv)** (Bo Peng 에서 제공)은 Bo Peng.의 [this repo](https://github.com/BlinkDL/RWKV-LM)논문과 함께 발표했습니다.
 1. **[SegFormer](https://huggingface.co/docs/transformers/model_doc/segformer)** (NVIDIA 에서) Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo 의 [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) 논문과 함께 발표했습니다.
-1. **[Segment Anything](https://huggingface.co/docs/transformers/model_doc/sam)** (Meta AI 에서 제공)은 Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alex Berg, Wan-Yen Lo, Piotr Dollar, Ross Girshick.의 [Segment Anything](https://arxiv.org/pdf/2304.02643v1.pdf)논문과 함께 발표했습니다.
+1. **[Segment Anything](https://huggingface.co/docs/transformers/main/model_doc/sam)** (Meta AI 에서 제공)은 Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alex Berg, Wan-Yen Lo, Piotr Dollar, Ross Girshick.의 [Segment Anything](https://arxiv.org/pdf/2304.02643v1.pdf)논문과 함께 발표했습니다.
 1. **[SEW](https://huggingface.co/docs/transformers/model_doc/sew)** (ASAPP 에서) Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi 의 [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) 논문과 함께 발표했습니다.
 1. **[SEW-D](https://huggingface.co/docs/transformers/model_doc/sew_d)** (ASAPP 에서) Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi 의 [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) 논문과 함께 발표했습니다.
 1. **[SpeechT5](https://huggingface.co/docs/transformers/model_doc/speecht5)** (Microsoft Research 에서 제공)은 Junyi Ao, Rui Wang, Long Zhou, Chengyi Wang, Shuo Ren, Yu Wu, Shujie Liu, Tom Ko, Qing Li, Yu Zhang, Zhihua Wei, Yao Qian, Jinyu Li, Furu Wei.의 [SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing](https://arxiv.org/abs/2110.07205)논문과 함께 발표했습니다.
@@ -369,7 +369,6 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (Facebook 에서) Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau 의 [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) 논문과 함께 발표했습니다.
 1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (Tel Aviv University 에서) Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy 의 [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) 논문과 함께 발표했습니다.
 1. **[SqueezeBERT](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (Berkeley 에서) Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer 의 [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) 논문과 함께 발표했습니다.
-1. **[SwiftFormer](https://huggingface.co/docs/transformers/main/model_doc/swiftformer)** (MBZUAI 에서 제공)은 Abdelrahman Shaker, Muhammad Maaz, Hanoona Rasheed, Salman Khan, Ming-Hsuan Yang, Fahad Shahbaz Khan.의 [SwiftFormer: Efficient Additive Attention for Transformer-based Real-time Mobile Vision Applications](https://arxiv.org/abs/2303.15446)논문과 함께 발표했습니다.
 1. **[Swin Transformer](https://huggingface.co/docs/transformers/model_doc/swin)** (Microsoft 에서) Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo 의 [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) 논문과 함께 발표했습니다.
 1. **[Swin Transformer V2](https://huggingface.co/docs/transformers/model_doc/swinv2)** (Microsoft 에서) Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo 의 [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) 논문과 함께 발표했습니다.
 1. **[Swin2SR](https://huggingface.co/docs/transformers/model_doc/swin2sr)** (University of Würzburg 에서) Marcos V. Conde, Ui-Jin Choi, Maxime Burchi, Radu Timofte 의 [Swin2SR: SwinV2 Transformer for Compressed Image Super-Resolution and Restoration](https://arxiv.org/abs/2209.11345) 논문과 함께 발표했습니다.

README_zh-hans.md

@@ -302,7 +302,7 @@ conda install -c huggingface transformers
 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. **[FocalNet](https://huggingface.co/docs/transformers/model_doc/focalnet)** (来自 Microsoft Research) 伴随论文 [Focal Modulation Networks](https://arxiv.org/abs/2203.11926) 由 Jianwei Yang, Chunyuan Li, Xiyang Dai, Lu Yuan, Jianfeng Gao 发布。
+1. **[FocalNet](https://huggingface.co/docs/transformers/main/model_doc/focalnet)** (来自 Microsoft Research) 伴随论文 [Focal Modulation Networks](https://arxiv.org/abs/2203.11926) 由 Jianwei Yang, Chunyuan Li, Xiyang Dai, Lu Yuan, Jianfeng Gao 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. **[GIT](https://huggingface.co/docs/transformers/model_doc/git)** (来自 Microsoft Research) 伴随论文 [GIT: A Generative Image-to-text Transformer for Vision and Language](https://arxiv.org/abs/2205.14100) 由 Jianfeng Wang, Zhengyuan Yang, Xiaowei Hu, Linjie Li, Kevin Lin, Zhe Gan, Zicheng Liu, Ce Liu, Lijuan Wang 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。
@@ -361,7 +361,7 @@ conda install -c huggingface transformers
 1. **[NLLB-MOE](https://huggingface.co/docs/transformers/model_doc/nllb-moe)** (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. **[OneFormer](https://huggingface.co/docs/transformers/model_doc/oneformer)** (来自 SHI Labs)  伴随论文 [OneFormer: One Transformer to Rule Universal Image Segmentation](https://arxiv.org/abs/2211.06220) 由 Jitesh Jain, Jiachen Li, MangTik Chiu, Ali Hassani, Nikita Orlov, Humphrey Shi 发布。
-1. **[OpenLlama](https://huggingface.co/docs/transformers/model_doc/open-llama)** (来自 [s-JoL](https://huggingface.co/s-JoL)) 由 [Open-Llama](https://github.com/s-JoL/Open-Llama) 发布. 
+1. **[OpenLlama](https://huggingface.co/docs/transformers/main/model_doc/open-llama)** (来自 [s-JoL](https://huggingface.co/s-JoL)) 由 [Open-Llama](https://github.com/s-JoL/Open-Llama) 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。
@@ -383,9 +383,9 @@ conda install -c huggingface transformers
 1. **[RoBERTa-PreLayerNorm](https://huggingface.co/docs/transformers/model_doc/roberta-prelayernorm)** (来自 Facebook) 伴随论文 [fairseq: A Fast, Extensible Toolkit for Sequence Modeling](https://arxiv.org/abs/1904.01038) 由 Myle Ott, Sergey Edunov, Alexei Baevski, Angela Fan, Sam Gross, Nathan Ng, David Grangier, Michael Auli 发布。
 1. **[RoCBert](https://huggingface.co/docs/transformers/model_doc/roc_bert)** (来自 WeChatAI), 伴随论文 [RoCBert: Robust Chinese Bert with Multimodal Contrastive Pretraining](https://aclanthology.org/2022.acl-long.65.pdf) 由 HuiSu, WeiweiShi, XiaoyuShen, XiaoZhou, TuoJi, JiaruiFang, JieZhou 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. **[RWKV](https://huggingface.co/docs/transformers/model_doc/rwkv)** (来自 Bo Peng) 伴随论文 [this repo](https://github.com/BlinkDL/RWKV-LM) 由 Bo Peng 发布。
+1. **[RWKV](https://huggingface.co/docs/transformers/main/model_doc/rwkv)** (来自 Bo Peng) 伴随论文 [this repo](https://github.com/BlinkDL/RWKV-LM) 由 Bo Peng 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. **[Segment Anything](https://huggingface.co/docs/transformers/model_doc/sam)** (来自 Meta AI) 伴随论文 [Segment Anything](https://arxiv.org/pdf/2304.02643v1.pdf) 由 Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alex Berg, Wan-Yen Lo, Piotr Dollar, Ross Girshick 发布。
+1. **[Segment Anything](https://huggingface.co/docs/transformers/main/model_doc/sam)** (来自 Meta AI) 伴随论文 [Segment Anything](https://arxiv.org/pdf/2304.02643v1.pdf) 由 Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alex Berg, Wan-Yen Lo, Piotr Dollar, Ross Girshick 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. **[SpeechT5](https://huggingface.co/docs/transformers/model_doc/speecht5)** (来自 Microsoft Research) 伴随论文 [SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing](https://arxiv.org/abs/2110.07205) 由 Junyi Ao, Rui Wang, Long Zhou, Chengyi Wang, Shuo Ren, Yu Wu, Shujie Liu, Tom Ko, Qing Li, Yu Zhang, Zhihua Wei, Yao Qian, Jinyu Li, Furu Wei 发布。
@@ -393,7 +393,6 @@ conda install -c huggingface transformers
 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. **[SwiftFormer](https://huggingface.co/docs/transformers/main/model_doc/swiftformer)** (来自 MBZUAI) 伴随论文 [SwiftFormer: Efficient Additive Attention for Transformer-based Real-time Mobile Vision Applications](https://arxiv.org/abs/2303.15446) 由 Abdelrahman Shaker, Muhammad Maaz, Hanoona Rasheed, Salman Khan, Ming-Hsuan Yang, Fahad Shahbaz Khan 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/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. **[Swin2SR](https://huggingface.co/docs/transformers/model_doc/swin2sr)** (来自 University of Würzburg) 伴随论文 [Swin2SR: SwinV2 Transformer for Compressed Image Super-Resolution and Restoration](https://arxiv.org/abs/2209.11345) 由 Marcos V. Conde, Ui-Jin Choi, Maxime Burchi, Radu Timofte 发布。

README_zh-hant.md

@@ -314,7 +314,7 @@ conda install -c huggingface transformers
 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. **[FocalNet](https://huggingface.co/docs/transformers/model_doc/focalnet)** (from Microsoft Research) released with the paper [Focal Modulation Networks](https://arxiv.org/abs/2203.11926) by Jianwei Yang, Chunyuan Li, Xiyang Dai, Lu Yuan, Jianfeng Gao.
+1. **[FocalNet](https://huggingface.co/docs/transformers/main/model_doc/focalnet)** (from Microsoft Research) released with the paper [Focal Modulation Networks](https://arxiv.org/abs/2203.11926) by Jianwei Yang, Chunyuan Li, Xiyang Dai, Lu Yuan, Jianfeng Gao.
 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. **[GIT](https://huggingface.co/docs/transformers/model_doc/git)** (from Microsoft Research) released with the paper [GIT: A Generative Image-to-text Transformer for Vision and Language](https://arxiv.org/abs/2205.14100) by Jianfeng Wang, Zhengyuan Yang, Xiaowei Hu, Linjie Li, Kevin Lin, Zhe Gan, Zicheng Liu, Ce Liu, Lijuan Wang.
 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.
@@ -373,7 +373,7 @@ conda install -c huggingface transformers
 1. **[NLLB-MOE](https://huggingface.co/docs/transformers/model_doc/nllb-moe)** (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. **[OneFormer](https://huggingface.co/docs/transformers/model_doc/oneformer)** (from SHI Labs) released with the paper [OneFormer: One Transformer to Rule Universal Image Segmentation](https://arxiv.org/abs/2211.06220) by Jitesh Jain, Jiachen Li, MangTik Chiu, Ali Hassani, Nikita Orlov, Humphrey Shi.
-1. **[OpenLlama](https://huggingface.co/docs/transformers/model_doc/open-llama)** (from [s-JoL](https://huggingface.co/s-JoL)) released in [Open-Llama](https://github.com/s-JoL/Open-Llama). 
+1. **[OpenLlama](https://huggingface.co/docs/transformers/main/model_doc/open-llama)** (from [s-JoL](https://huggingface.co/s-JoL)) released in [Open-Llama](https://github.com/s-JoL/Open-Llama). 
 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.
@@ -395,9 +395,9 @@ conda install -c huggingface transformers
 1. **[RoBERTa-PreLayerNorm](https://huggingface.co/docs/transformers/model_doc/roberta-prelayernorm)** (from Facebook) released with the paper [fairseq: A Fast, Extensible Toolkit for Sequence Modeling](https://arxiv.org/abs/1904.01038) by Myle Ott, Sergey Edunov, Alexei Baevski, Angela Fan, Sam Gross, Nathan Ng, David Grangier, Michael Auli.
 1. **[RoCBert](https://huggingface.co/docs/transformers/model_doc/roc_bert)** (from WeChatAI) released with the paper [RoCBert: Robust Chinese Bert with Multimodal Contrastive Pretraining](https://aclanthology.org/2022.acl-long.65.pdf) by HuiSu, WeiweiShi, XiaoyuShen, XiaoZhou, TuoJi, JiaruiFang, JieZhou.
 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. **[RWKV](https://huggingface.co/docs/transformers/model_doc/rwkv)** (from Bo Peng) released with the paper [this repo](https://github.com/BlinkDL/RWKV-LM) by Bo Peng.
+1. **[RWKV](https://huggingface.co/docs/transformers/main/model_doc/rwkv)** (from Bo Peng) released with the paper [this repo](https://github.com/BlinkDL/RWKV-LM) by Bo Peng.
 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. **[Segment Anything](https://huggingface.co/docs/transformers/model_doc/sam)** (from Meta AI) released with the paper [Segment Anything](https://arxiv.org/pdf/2304.02643v1.pdf) by Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alex Berg, Wan-Yen Lo, Piotr Dollar, Ross Girshick.
+1. **[Segment Anything](https://huggingface.co/docs/transformers/main/model_doc/sam)** (from Meta AI) released with the paper [Segment Anything](https://arxiv.org/pdf/2304.02643v1.pdf) by Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alex Berg, Wan-Yen Lo, Piotr Dollar, Ross Girshick.
 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. **[SpeechT5](https://huggingface.co/docs/transformers/model_doc/speecht5)** (from Microsoft Research) released with the paper [SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing](https://arxiv.org/abs/2110.07205) by Junyi Ao, Rui Wang, Long Zhou, Chengyi Wang, Shuo Ren, Yu Wu, Shujie Liu, Tom Ko, Qing Li, Yu Zhang, Zhihua Wei, Yao Qian, Jinyu Li, Furu Wei.
@@ -405,7 +405,6 @@ conda install -c huggingface transformers
 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. **[SwiftFormer](https://huggingface.co/docs/transformers/main/model_doc/swiftformer)** (from MBZUAI) released with the paper [SwiftFormer: Efficient Additive Attention for Transformer-based Real-time Mobile Vision Applications](https://arxiv.org/abs/2303.15446) by Abdelrahman Shaker, Muhammad Maaz, Hanoona Rasheed, Salman Khan, Ming-Hsuan Yang, Fahad Shahbaz Khan.
 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/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. **[Swin2SR](https://huggingface.co/docs/transformers/model_doc/swin2sr)** (from University of Würzburg) released with the paper [Swin2SR: SwinV2 Transformer for Compressed Image Super-Resolution and Restoration](https://arxiv.org/abs/2209.11345) by Marcos V. Conde, Ui-Jin Choi, Maxime Burchi, Radu Timofte.

conftest.py

@@ -20,10 +20,6 @@ import sys
 import warnings
 from os.path import abspath, dirname, join
 
-import _pytest
-
-from transformers.testing_utils import HfDoctestModule, HfDocTestParser
-
 
 # allow having multiple repository checkouts and not needing to remember to rerun
 # 'pip install -e .[dev]' when switching between checkouts and running tests.
@@ -42,7 +38,9 @@ def pytest_configure(config):
     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_pipeline_test: mark test to run only when pipelines are tested")
+    config.addinivalue_line(
+        "markers", "is_pipeline_test: mark test to run only when pipelines are tested"
+    )
     config.addinivalue_line("markers", "is_staging_test: mark test to run only in the staging environment")
     config.addinivalue_line("markers", "accelerate_tests: mark test that require accelerate")
     config.addinivalue_line("markers", "tool_tests: mark the tool tests that are run on their specific schedule")
@@ -69,7 +67,7 @@ def pytest_sessionfinish(session, exitstatus):
 
 
 # Doctest custom flag to ignore output.
-IGNORE_RESULT = doctest.register_optionflag("IGNORE_RESULT")
+IGNORE_RESULT = doctest.register_optionflag('IGNORE_RESULT')
 
 OutputChecker = doctest.OutputChecker
 
@@ -82,5 +80,3 @@ class CustomOutputChecker(OutputChecker):
 
 
 doctest.OutputChecker = CustomOutputChecker
-_pytest.doctest.DoctestModule = HfDoctestModule
-doctest.DocTestParser = HfDocTestParser

docker/transformers-all-latest-gpu/Dockerfile

@@ -9,7 +9,7 @@ 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='2.0.1'
+ARG PYTORCH='2.0.0'
 # (not always a valid torch version)
 ARG INTEL_TORCH_EXT='1.11.0'
 # Example: `cu102`, `cu113`, etc.
@@ -32,9 +32,15 @@ RUN echo torch=$VERSION
 # TODO: We might need to specify proper versions that work with a specific torch version (especially for past CI).
 RUN [ "$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==2.12 protobuf==3.20.3 tensorflow_text tensorflow_probability
+RUN python3 -m pip install --no-cache-dir -U tensorflow==2.11
+RUN python3 -m pip install --no-cache-dir -U tensorflow_probability
 RUN python3 -m pip uninstall -y flax jax
 
+# To include the change in this commit https://github.com/onnx/tensorflow-onnx/commit/ddca3a5eb2d912f20fe7e0568dd1a3013aee9fa3
+# Otherwise, we get tf2onnx==1.8 (caused by `flatbuffers` version),  and some tests fail with `ValueError: from_keras requires input_signature`.
+# TODO: remove this line once the conflict is resolved in these libraries.
+RUN python3 -m pip install --no-cache-dir git+https://github.com/onnx/tensorflow-onnx.git@ddca3a5eb2d912f20fe7e0568dd1a3013aee9fa3
+
 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

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

@@ -4,7 +4,7 @@ LABEL maintainer="Hugging Face"
 
 ARG DEBIAN_FRONTEND=noninteractive
 
-ARG PYTORCH='2.0.1'
+ARG PYTORCH='2.0.0'
 # Example: `cu102`, `cu113`, etc.
 ARG CUDA='cu117'
 

docker/transformers-pytorch-gpu/Dockerfile

@@ -12,7 +12,7 @@ RUN git clone https://github.com/huggingface/transformers && cd transformers &&
 RUN python3 -m pip install --no-cache-dir -e ./transformers[dev-torch,testing,video]
 
 # If set to nothing, will install the latest version
-ARG PYTORCH='2.0.1'
+ARG PYTORCH='2.0.0'
 ARG TORCH_VISION=''
 ARG TORCH_AUDIO=''
 # Example: `cu102`, `cu113`, etc.

docker/transformers-tensorflow-gpu/Dockerfile

@@ -12,7 +12,7 @@ RUN git clone https://github.com/huggingface/transformers && cd transformers &&
 RUN python3 -m pip install --no-cache-dir -e ./transformers[dev-tensorflow,testing]
 
 # If set to nothing, will install the latest version
-ARG TENSORFLOW='2.12'
+ARG TENSORFLOW='2.11'
 
 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

docs/source/en/_toctree.yml

@@ -502,8 +502,6 @@
         title: ResNet
       - local: model_doc/segformer
         title: SegFormer
-      - local: model_doc/swiftformer
-        title: SwiftFormer
       - local: model_doc/swin
         title: Swin Transformer
       - local: model_doc/swinv2

docs/source/en/custom_tools.mdx

@@ -503,7 +503,7 @@ print("\n".join([f"- {a}" for a in agent.toolbox.keys()]))
 
 Note how `image_upscaler` is now part of the agents' toolbox.
 
-Let's now try out the new tools! We will re-use the image we generated in [Transformers Agents Quickstart](./transformers_agents#single-execution-run).
+Let's now try out the new tools! We will re-use the image we generated in (Transformers Agents Quickstart)[./transformers_agents#single-execution-run].
 
 ```py
 from diffusers.utils import load_image
@@ -726,7 +726,7 @@ We pass that instance to the `Tool.from_gradio` method:
 ```python
 from transformers import Tool
 
-tool = Tool.from_gradio(gradio_tool)
+tool = Tool.from_gradio(gradio_tools)
 ```
 
 Now we can manage it exactly as we would a usual custom tool. We leverage it to improve our prompt

docs/source/en/index.mdx

@@ -206,7 +206,6 @@ The documentation is organized into five sections:
 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. **[SwiftFormer](model_doc/swiftformer)** (from MBZUAI) released with the paper [SwiftFormer: Efficient Additive Attention for Transformer-based Real-time Mobile Vision Applications](https://arxiv.org/abs/2303.15446) by Abdelrahman Shaker, Muhammad Maaz, Hanoona Rasheed, Salman Khan, Ming-Hsuan Yang, Fahad Shahbaz Khan.
 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. **[Swin2SR](model_doc/swin2sr)** (from University of Würzburg) released with the paper [Swin2SR: SwinV2 Transformer for Compressed Image Super-Resolution and Restoration](https://arxiv.org/abs/2209.11345) by Marcos V. Conde, Ui-Jin Choi, Maxime Burchi, Radu Timofte.
@@ -409,7 +408,6 @@ Flax), PyTorch, and/or TensorFlow.
 |           SpeechT5            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           Splinter            |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
 |          SqueezeBERT          |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
-|          SwiftFormer          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |       Swin Transformer        |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
 |      Swin Transformer V2      |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            Swin2SR            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |

docs/source/en/main_classes/agent.mdx

@@ -19,7 +19,7 @@ can vary as the APIs or underlying models are prone to change.
 
 </Tip>
 
-To learn more about agents and tools make sure to read the [introductory guide](../transformers_agents). This page
+To learn more about agents and tools make sure to read the [introductory guide](../agents_and_tools). This page
 contains the API docs for the underlying classes.
 
 ## Agents

docs/source/en/model_doc/swiftformer.mdx

@@ -1,45 +0,0 @@
-<!--Copyright 2023 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.
--->
-
-# SwiftFormer
-
-## Overview
-
-The SwiftFormer model was proposed in [SwiftFormer: Efficient Additive Attention for Transformer-based Real-time Mobile Vision Applications](https://arxiv.org/abs/2303.15446) by Abdelrahman Shaker, Muhammad Maaz, Hanoona Rasheed, Salman Khan, Ming-Hsuan Yang, Fahad Shahbaz Khan.
-
-The SwiftFormer paper introduces a novel efficient additive attention mechanism that effectively replaces the quadratic matrix multiplication operations in the self-attention computation with linear element-wise multiplications. A series of models called 'SwiftFormer' is built based on this, which achieves state-of-the-art performance in terms of both accuracy and mobile inference speed. Even their small variant achieves 78.5% top-1 ImageNet1K accuracy with only 0.8 ms latency on iPhone 14, which is more accurate and 2× faster compared to MobileViT-v2.
-
-The abstract from the paper is the following:
-
-*Self-attention has become a defacto choice for capturing global context in various vision applications. However, its quadratic computational complexity with respect to image resolution limits its use in real-time applications, especially for deployment on resource-constrained mobile devices. Although hybrid approaches have been proposed to combine the advantages of convolutions and self-attention for a better speed-accuracy trade-off, the expensive matrix multiplication operations in self-attention remain a bottleneck. In this work, we introduce a novel efficient additive attention mechanism that effectively replaces the quadratic matrix multiplication operations with linear element-wise multiplications. Our design shows that the key-value interaction can be replaced with a linear layer without sacrificing any accuracy. Unlike previous state-of-the-art methods, our efficient formulation of self-attention enables its usage at all stages of the network. Using our proposed efficient additive attention, we build a series of models called "SwiftFormer" which achieves state-of-the-art performance in terms of both accuracy and mobile inference speed. Our small variant achieves 78.5% top-1 ImageNet-1K accuracy with only 0.8 ms latency on iPhone 14, which is more accurate and 2x faster compared to MobileViT-v2.*
-
-Tips:
-    - One can use the [`ViTImageProcessor`] API to prepare images for the model.
-
-
-This model was contributed by [shehan97](https://huggingface.co/shehan97).
-The original code can be found [here](https://github.com/Amshaker/SwiftFormer).
-
-
-## SwiftFormerConfig
-
-[[autodoc]] SwiftFormerConfig
-
-## SwiftFormerModel
-
-[[autodoc]] SwiftFormerModel
-    - forward
-
-## SwiftFormerForImageClassification
-
-[[autodoc]] SwiftFormerForImageClassification
-    - forward

docs/source/en/serialization.mdx

@@ -112,7 +112,6 @@ Ready-made configurations include the following architectures:
 - RoFormer
 - SegFormer
 - SqueezeBERT
-- SwiftFormer
 - Swin Transformer
 - T5
 - Table Transformer

docs/source/en/tasks/asr.mdx

@@ -282,7 +282,7 @@ At this point, only three steps remain:
 ...     args=training_args,
 ...     train_dataset=encoded_minds["train"],
 ...     eval_dataset=encoded_minds["test"],
-...     tokenizer=processor,
+...     tokenizer=processor.feature_extractor,
 ...     data_collator=data_collator,
 ...     compute_metrics=compute_metrics,
 ... )

docs/source/en/tasks/image_classification.mdx

@@ -30,7 +30,7 @@ The task illustrated in this tutorial is supported by the following model archit
 
 <!--This tip is automatically generated by `make fix-copies`, do not fill manually!-->
 
-[BEiT](../model_doc/beit), [BiT](../model_doc/bit), [ConvNeXT](../model_doc/convnext), [ConvNeXTV2](../model_doc/convnextv2), [CvT](../model_doc/cvt), [Data2VecVision](../model_doc/data2vec-vision), [DeiT](../model_doc/deit), [DiNAT](../model_doc/dinat), [EfficientFormer](../model_doc/efficientformer), [EfficientNet](../model_doc/efficientnet), [FocalNet](../model_doc/focalnet), [ImageGPT](../model_doc/imagegpt), [LeViT](../model_doc/levit), [MobileNetV1](../model_doc/mobilenet_v1), [MobileNetV2](../model_doc/mobilenet_v2), [MobileViT](../model_doc/mobilevit), [NAT](../model_doc/nat), [Perceiver](../model_doc/perceiver), [PoolFormer](../model_doc/poolformer), [RegNet](../model_doc/regnet), [ResNet](../model_doc/resnet), [SegFormer](../model_doc/segformer), [SwiftFormer](../model_doc/swiftformer), [Swin Transformer](../model_doc/swin), [Swin Transformer V2](../model_doc/swinv2), [VAN](../model_doc/van), [ViT](../model_doc/vit), [ViT Hybrid](../model_doc/vit_hybrid), [ViTMSN](../model_doc/vit_msn)
+[BEiT](../model_doc/beit), [BiT](../model_doc/bit), [ConvNeXT](../model_doc/convnext), [ConvNeXTV2](../model_doc/convnextv2), [CvT](../model_doc/cvt), [Data2VecVision](../model_doc/data2vec-vision), [DeiT](../model_doc/deit), [DiNAT](../model_doc/dinat), [EfficientFormer](../model_doc/efficientformer), [EfficientNet](../model_doc/efficientnet), [FocalNet](../model_doc/focalnet), [ImageGPT](../model_doc/imagegpt), [LeViT](../model_doc/levit), [MobileNetV1](../model_doc/mobilenet_v1), [MobileNetV2](../model_doc/mobilenet_v2), [MobileViT](../model_doc/mobilevit), [NAT](../model_doc/nat), [Perceiver](../model_doc/perceiver), [PoolFormer](../model_doc/poolformer), [RegNet](../model_doc/regnet), [ResNet](../model_doc/resnet), [SegFormer](../model_doc/segformer), [Swin Transformer](../model_doc/swin), [Swin Transformer V2](../model_doc/swinv2), [VAN](../model_doc/van), [ViT](../model_doc/vit), [ViT Hybrid](../model_doc/vit_hybrid), [ViTMSN](../model_doc/vit_msn)
 <!--End of the generated tip-->
 
 </Tip>

docs/source/en/tasks/text-to-speech.mdx

@@ -469,7 +469,7 @@ Instantiate the `Trainer` object  and pass the model, dataset, and data collator
 ...     train_dataset=dataset["train"],
 ...     eval_dataset=dataset["test"],
 ...     data_collator=data_collator,
-...     tokenizer=processor,
+...     tokenizer=processor.tokenizer,
 ... )
 ```
 

docs/source/en/testing.mdx

@@ -212,12 +212,20 @@ Example:
     ```"""
 
 ```
+3 steps are required to debug the docstring examples: 
+1. In order to properly run the test, **an extra line has to be added** at the end of the docstring. This can be automatically done on any file using: 
+```bash 
+python utils/prepare_for_doc_test.py <path_to_file_or_dir>
+```
 
-Just run the following line to automatically test every docstring example in the desired file: 
+2. Then, you can use the following line to automatically test every docstring example in the desired file: 
 ```bash 
 pytest --doctest-modules <path_to_file_or_dir>
 ```
-If the file has a markdown extention, you should add the `--doctest-glob="*.mdx"` argument.
+3. Once you are done debugging, you need to remove the extra line added in step **1.** by running the following: 
+```bash 
+python utils/prepare_for_doc_test.py <path_to_file_or_dir> --remove_new_line
+```
 
 ### Run only modified tests
 

docs/source/en/tools_inference_endpoints.mdx

@@ -1,132 +0,0 @@
-<!--Copyright 2023 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.
--->
-
-# Running tools on inference endpoints
-
-<Tip>
-
-This document is about running tools on inference endpoints so that agents may use these tools remotely. 
-If you do not know what tools and agents are in the context of transformers, we recommend you read the
-[Transformers Agents](transformers_agents) page first.
-
-</Tip>
-
-Agents are designed to use tools in order to respond to a natural language query. They are setup so as to load tools
-locally and use them directly in the runtime they're at.
-
-However, some of these tools can be heavy; tools that handle images, long text, or audio signals may need a 
-significant amount of memory in order to perform inference. Tools that generate images through a diffusion
-process, may require significant compute in order to perform the multiple steps they need; but end users
-may not benefit from the powerful setups required to use them.
-
-This is why we have support for **remote** tools: these have an API that can be called from the runtime, offloading
-the processing to the remote API. In this guide we'll explore how to set up an inference endpoint for a given tool
-to leverage it with the agents.
-
-Inference endpoints are one solution to handle remote tools; but they're not the only one. We integrate with
-[`gradio_tools`](custom_tools#leveraging-gradiotools) that also offers remote tools, and we'll continue adding 
-guides to other alternatives for remote tools.
-
-## Inference Endpoints
-
-
-[Inference Endpoints](https://huggingface.co/inference-endpoints) is a paid Hugging Face solution to easily deploy
-Transformers and Diffusers models on a fully-managed infrastructure. It has default deployment options for
-transformers and diffusers, but given that we're using a specific type of object here, tools, we'll set up a custom
-handler to get it to work.
-
-<Tip warning={true}>
-
-Inference Endpoints are a paid hosting service by Hugging Face, which needs to have an organization setup with 
-billing enabled.
-
-</Tip>
-
-Tools are Spaces by default in Transformers. When calling `push_to_hub` on a tool, you're effectively pushing
-the code to a Space on the Hugging Face Hub under a namespace that you own. There are many tools living on the
-[`huggingface-tools` namespace](https://huggingface.co/huggingface-tools); having them be Spaces by default means
-that users can play around with the tool directly in the browser.
-
-However, Inference Endpoints only work with **model** repositories. We'll therefore have to create a model
-repository to act as a proxy for the Space. That model repository will contain the `handler.py` file to serve
-our tool through an inference endpoint.
-
-For demonstration purposes, we'll consider that you already have a tool handy that you'd like to use remotely. If
-you'd like to setup your custom tool, we recommend reading the [Custom Tool](custom_tools#leveraging-gradiotools) 
-guide.
-
-We'll try and deploy the `huggingface-tools/text-to-video` tool to an inference endpoint. We have it available as 
-a gradio Space [here](https://huggingface.co/huggingface-tools/text-to-video).
-
-### Setting up the repository
-
-We'll start by creating a model repository that will serve as a serving point for this tool.
-It can be public or private; for the sake of this tutorial we'll keep this one public, but having it set to
-private doesn't interfere with the inference endpoint setup.
-
-The repository is created and is available [here](https://huggingface.co/huggingface-tools/text-to-video).
-In it, you'll see there is a custom handler file, called 
-[`handler.py`](https://huggingface.co/huggingface-tools/text-to-video/blob/main/handler.py), as well as a traditional
-requirements file called 
-[`requirements.txt`](https://huggingface.co/huggingface-tools/text-to-video/blob/main/requirements.txt).
-
-#### Handler file
-
-The handler file exposes an `EndpointHandler`, which serves as the link between the requests you'll be doing to the
-remote tool and the tool itself. It should:
-
-- Instantiate the tool in its initialization method
-- Have a `__call__` method which will take the serialized input and return the computed result.
-
-For text-to-text tools, the handler file is very simple; it looks like the following:
-
-```python
-from transformers.tools import load_tool
-
-
-class EndpointHandler:
-    def __init__(self, path=""):
-        self.tool = load_tool("huggingface-tools/text-to-video")
-        self.tool.setup()
-
-    def __call__(self, data):
-        inputs = data.pop("inputs", data)
-        return self.tool(**inputs)
-```
-
-However, it is different if handling different data types as it will need to serialize this data type. 
-This guide will be completed to include different serialization for text, image, audio and video.
-
-#### Requirement file
-
-The requirement file needs to specify all requirements necessary to run the tool. The basic dependencies are the 
-following:
-
-```text
-transformers>=4.29.0
-accelerate
-```
-
-but you may need to include any and all other dependencies needed by your tool
-
-### Spinning up an endpoint
-
-Once we're done creating the repository, we can go ahead and create our first endpoint. Head over to
-[the Inference Endpoints UI](https://ui.endpoints.huggingface.co/endpoints) and create your first endpoint.
-
-If the repository is setup correctly, it should spin up directly without issue.
-
-In case you encounter a "Failed" deployment, we recommend checking out 
-[this guide](https://huggingface.co/docs/inference-endpoints/guides/logs) on checking out the logs of an inference
-endpoint.
-
-TODO add images

docs/source/en/transformers_agents.mdx

@@ -65,17 +65,7 @@ We provide support for openAI models as well as opensource alternatives from Big
 models perform better (but require you to have an openAI API key, so cannot be used for free); Hugging Face is
 providing free access to endpoints for BigCode and OpenAssistant models.
 
-To start with, please install the `agents` extras in order to install all default dependencies.
-```bash
-pip install transformers[agents]
-```
-
-To use openAI models, you instantiate an [`OpenAiAgent`] after installing the `openai` dependency:
-
-```bash
-pip install openai
-```
-
+To use openAI models, you instantiate an [`OpenAiAgent`]:
 
 ```py
 from transformers import OpenAiAgent
@@ -266,16 +256,16 @@ with the code generated by the agent.
 We identify a set of tools that can empower such agents. Here is an updated list of the tools we have integrated 
 in `transformers`:
 
-- **Document question answering**: given a document (such as a PDF) in image format, answer a question on this document ([Donut](./model_doc/donut))
-- **Text question answering**: given a long text and a question, answer the question in the text ([Flan-T5](./model_doc/flan-t5))
-- **Unconditional image captioning**: Caption the image! ([BLIP](./model_doc/blip))
-- **Image question answering**: given an image, answer a question on this image ([VILT](./model_doc/vilt))
-- **Image segmentation**: given an image and a prompt, output the segmentation mask of that prompt ([CLIPSeg](./model_doc/clipseg))
-- **Speech to text**: given an audio recording of a person talking, transcribe the speech into text ([Whisper](./model_doc/whisper))
-- **Text to speech**: convert text to speech ([SpeechT5](./model_doc/speecht5))
-- **Zero-shot text classification**: given a text and a list of labels, identify to which label the text corresponds the most ([BART](./model_doc/bart))
-- **Text summarization**: summarize a long text in one or a few sentences ([BART](./model_doc/bart))
-- **Translation**: translate the text into a given language ([NLLB](./model_doc/nllb))
+- **Document question answering**: given a document (such as a PDF) in image format, answer a question on this document ([Donut](../model_doc/donut))
+- **Text question answering**: given a long text and a question, answer the question in the text ([Flan-T5](../model_doc/flan-t5))
+- **Unconditional image captioning**: Caption the image! ([BLIP](../model_doc/blip))
+- **Image question answering**: given an image, answer a question on this image ([VILT](../model_doc/vilt))
+- **Image segmentation**: given an image and a prompt, output the segmentation mask of that prompt ([CLIPSeg](../model_doc/clipseg))
+- **Speech to text**: given an audio recording of a person talking, transcribe the speech into text ([Whisper](../model_doc/whisper))
+- **Text to speech**: convert text to speech ([SpeechT5](../model_doc/speecht5))
+- **Zero-shot text classification**: given a text and a list of labels, identify to which label the text corresponds the most ([BART](../model_doc/bart))
+- **Text summarization**: summarize a long text in one or a few sentences ([BART](../model_doc/bart))
+- **Translation**: translate the text into a given language ([NLLB](../model_doc/nllb))
 
 These tools have an integration in transformers, and can be used manually as well, for example:
 
@@ -293,7 +283,7 @@ the ability to quickly create and share custom tools.
 
 By pushing the code of a tool to a Hugging Face Space or a model repository, you're then able to leverage the tool 
 directly with the agent. We've added a few 
-**transformers-agnostic** tools to the [`huggingface-tools` organization](https://huggingface.co/huggingface-tools):
+**transformers-agnostic** tools to the `huggingface-tools` organization:
 
 - **Text downloader**: to download a text from a web URL
 - **Text to image**: generate an image according to a prompt, leveraging stable diffusion
@@ -304,7 +294,7 @@ The text-to-image tool we have been using since the beginning is a remote tool t
 [*huggingface-tools/text-to-image*](https://huggingface.co/spaces/huggingface-tools/text-to-image)! We will
 continue releasing such tools on this and other organizations, to further supercharge this implementation.
 
-The agents have by default access to tools that reside on [`huggingface-tools`](https://huggingface.co/huggingface-tools).
+The agents have by default access to tools that reside on `huggingface-tools`.
 We explain how to you can write and share your tools as well as leverage any custom tool that resides on the Hub in [following guide](custom_tools).
 
 ### Code generation

examples/flax/question-answering/run_qa.py

@@ -61,7 +61,7 @@ from transformers.utils import check_min_version, get_full_repo_name, send_examp
 logger = logging.getLogger(__name__)
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 Array = Any
 Dataset = datasets.arrow_dataset.Dataset

examples/flax/text-classification/run_flax_glue.py

@@ -54,7 +54,7 @@ from transformers.utils import check_min_version, get_full_repo_name, send_examp
 
 logger = logging.getLogger(__name__)
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 Array = Any
 Dataset = datasets.arrow_dataset.Dataset

examples/flax/token-classification/run_flax_ner.py

@@ -55,7 +55,7 @@ from transformers.utils.versions import require_version
 
 logger = logging.getLogger(__name__)
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/token-classification/requirements.txt")
 

examples/pytorch/audio-classification/run_audio_classification.py

@@ -45,7 +45,7 @@ from transformers.utils.versions import require_version
 logger = logging.getLogger(__name__)
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.14.0", "To fix: pip install -r examples/pytorch/audio-classification/requirements.txt")
 

examples/pytorch/contrastive-image-text/run_clip.py

@@ -54,7 +54,7 @@ from transformers.utils.versions import require_version
 logger = logging.getLogger(__name__)
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/contrastive-image-text/requirements.txt")
 

examples/pytorch/image-classification/run_image_classification.py

@@ -55,7 +55,7 @@ from transformers.utils.versions import require_version
 logger = logging.getLogger(__name__)
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/image-classification/requirements.txt")
 

examples/pytorch/image-classification/run_image_classification_no_trainer.py

@@ -47,7 +47,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 logger = get_logger(__name__)
 

examples/pytorch/image-pretraining/run_mae.py

@@ -43,7 +43,7 @@ from transformers.utils.versions import require_version
 logger = logging.getLogger(__name__)
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/image-pretraining/requirements.txt")
 

examples/pytorch/image-pretraining/run_mim.py

@@ -48,7 +48,7 @@ Any model supported by the AutoModelForMaskedImageModeling API can be used.
 logger = logging.getLogger(__name__)
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/image-pretraining/requirements.txt")
 

examples/pytorch/image-pretraining/run_mim_no_trainer.py

@@ -53,7 +53,7 @@ Any model supported by the AutoModelForMaskedImageModeling API can be used.
 logger = logging.getLogger(__name__)
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/image-pretraining/requirements.txt")
 

examples/pytorch/language-modeling/run_clm.py

@@ -55,7 +55,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/language-modeling/requirements.txt")
 

examples/pytorch/language-modeling/run_clm_no_trainer.py

@@ -57,7 +57,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 logger = get_logger(__name__)
 

examples/pytorch/language-modeling/run_mlm.py

@@ -53,7 +53,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/language-modeling/requirements.txt")
 

examples/pytorch/language-modeling/run_mlm_no_trainer.py

@@ -57,7 +57,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 logger = get_logger(__name__)
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/language-modeling/requirements.txt")

examples/pytorch/language-modeling/run_plm.py

@@ -47,7 +47,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/language-modeling/requirements.txt")
 

examples/pytorch/multiple-choice/run_swag.py

@@ -47,7 +47,7 @@ from transformers.utils import PaddingStrategy, check_min_version, send_example_
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 logger = logging.getLogger(__name__)
 

examples/pytorch/multiple-choice/run_swag_no_trainer.py

@@ -56,7 +56,7 @@ from transformers.utils import PaddingStrategy, check_min_version, get_full_repo
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 logger = get_logger(__name__)
 # You should update this to your particular problem to have better documentation of `model_type`

examples/pytorch/question-answering/run_qa.py

@@ -49,7 +49,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/question-answering/requirements.txt")
 

examples/pytorch/question-answering/run_qa_beam_search.py

@@ -48,7 +48,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/question-answering/requirements.txt")
 

examples/pytorch/question-answering/run_qa_beam_search_no_trainer.py

@@ -56,7 +56,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/question-answering/requirements.txt")
 

examples/pytorch/question-answering/run_qa_no_trainer.py

@@ -57,7 +57,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/question-answering/requirements.txt")
 

examples/pytorch/question-answering/run_seq2seq_qa.py

@@ -46,7 +46,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/question-answering/requirements.txt")
 

examples/pytorch/semantic-segmentation/run_semantic_segmentation.py

@@ -51,7 +51,7 @@ from transformers.utils.versions import require_version
 logger = logging.getLogger(__name__)
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=2.0.0", "To fix: pip install -r examples/pytorch/semantic-segmentation/requirements.txt")
 

examples/pytorch/semantic-segmentation/run_semantic_segmentation_no_trainer.py

@@ -50,7 +50,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 logger = get_logger(__name__)
 
@@ -682,9 +682,7 @@ def main():
             if args.push_to_hub:
                 repo.push_to_hub(commit_message="End of training", auto_lfs_prune=True)
 
-            all_results = {
-                f"eval_{k}": v.tolist() if isinstance(v, np.ndarray) else v for k, v in eval_metrics.items()
-            }
+            all_results = {f"eval_{k}": v for k, v in eval_metrics.items()}
             with open(os.path.join(args.output_dir, "all_results.json"), "w") as f:
                 json.dump(all_results, f)
 

examples/pytorch/speech-recognition/run_speech_recognition_ctc.py

@@ -51,7 +51,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.18.0", "To fix: pip install -r examples/pytorch/speech-recognition/requirements.txt")
 
@@ -705,7 +705,7 @@ def main():
         compute_metrics=compute_metrics,
         train_dataset=vectorized_datasets["train"] if training_args.do_train else None,
         eval_dataset=vectorized_datasets["eval"] if training_args.do_eval else None,
-        tokenizer=processor,
+        tokenizer=feature_extractor,
     )
 
     # 8. Finally, we can start training

examples/pytorch/speech-recognition/run_speech_recognition_seq2seq.py

@@ -48,7 +48,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.18.0", "To fix: pip install -r examples/pytorch/speech-recognition/requirements.txt")
 

examples/pytorch/summarization/run_summarization.py

@@ -52,7 +52,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/summarization/requirements.txt")
 

examples/pytorch/summarization/run_summarization_no_trainer.py

@@ -56,7 +56,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 logger = get_logger(__name__)
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/summarization/requirements.txt")

examples/pytorch/test_accelerate_examples.py

@@ -21,7 +21,6 @@ import os
 import shutil
 import sys
 import tempfile
-import unittest
 from unittest import mock
 
 import torch
@@ -177,7 +176,6 @@ class ExamplesTestsNoTrainer(TestCasePlus):
         self.assertTrue(os.path.exists(os.path.join(tmp_dir, "epoch_0")))
         self.assertTrue(os.path.exists(os.path.join(tmp_dir, "ner_no_trainer")))
 
-    @unittest.skip(reason="Fix me @zack")
     @mock.patch.dict(os.environ, {"WANDB_MODE": "offline"})
     def test_run_squad_no_trainer(self):
         tmp_dir = self.get_auto_remove_tmp_dir()

examples/pytorch/text-classification/run_glue.py

@@ -48,7 +48,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/text-classification/requirements.txt")
 

examples/pytorch/text-classification/run_glue_no_trainer.py

@@ -48,7 +48,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 logger = get_logger(__name__)
 

examples/pytorch/text-classification/run_xnli.py

@@ -48,7 +48,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/text-classification/requirements.txt")
 

examples/pytorch/token-classification/run_ner.py

@@ -49,7 +49,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/token-classification/requirements.txt")
 

examples/pytorch/token-classification/run_ner_no_trainer.py

@@ -55,7 +55,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 logger = get_logger(__name__)
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/token-classification/requirements.txt")

examples/pytorch/translation/run_translation.py

@@ -52,7 +52,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/translation/requirements.txt")
 

examples/pytorch/translation/run_translation_no_trainer.py

@@ -57,7 +57,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 logger = get_logger(__name__)
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/translation/requirements.txt")

examples/tensorflow/contrastive-image-text/run_clip.py

@@ -51,7 +51,7 @@ from transformers.utils.versions import require_version
 logger = logging.getLogger(__name__)
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version(
     "datasets>=1.8.0", "To fix: pip install -r examples/tensorflow/contrastive-image-text/requirements.txt"

examples/tensorflow/image-classification/run_image_classification.py

@@ -54,7 +54,7 @@ from transformers.utils.versions import require_version
 logger = logging.getLogger(__name__)
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/image-classification/requirements.txt")
 

examples/tensorflow/multiple-choice/run_swag.py

@@ -50,7 +50,7 @@ from transformers.utils import PaddingStrategy, check_min_version, send_example_
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 logger = logging.getLogger(__name__)
 

examples/tensorflow/question-answering/run_qa.py

@@ -48,7 +48,7 @@ from transformers.utils import CONFIG_NAME, TF2_WEIGHTS_NAME, check_min_version,
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 logger = logging.getLogger(__name__)
 

examples/tensorflow/summarization/run_summarization.py

@@ -53,7 +53,7 @@ from transformers.utils.versions import require_version
 
 # region Checking dependencies
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/summarization/requirements.txt")
 

examples/tensorflow/text-classification/run_glue.py

@@ -47,7 +47,7 @@ from transformers.utils import check_min_version, send_example_telemetry
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 task_to_keys = {
     "cola": ("sentence", None),

examples/tensorflow/translation/run_translation.py

@@ -56,7 +56,7 @@ from transformers.utils.versions import require_version
 
 # region Dependencies and constants
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.30.0.dev0")
+check_min_version("4.29.0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/summarization/requirements.txt")
 

scripts/tatoeba/README.md

@@ -54,7 +54,7 @@ To upload all converted models,
 
 1. Install [git-lfs](https://git-lfs.github.com/).
 
-2. Login to `huggingface-cli`
+2. Login to `transformers-cli`
 
 ```bash
 huggingface-cli login

setup.cfg

@@ -1,3 +1,2 @@
 [tool:pytest]
 doctest_optionflags=NUMBER NORMALIZE_WHITESPACE ELLIPSIS
-doctest_glob=**/*.mdx

setup.py

@@ -112,7 +112,6 @@ _deps = [
     "datasets!=2.5.0",
     "decord==0.6.0",
     "deepspeed>=0.8.3",
-    "diffusers",
     "dill<0.3.5",
     "evaluate>=0.2.0",
     "fairscale>0.3",
@@ -124,7 +123,7 @@ _deps = [
     "fugashi>=1.0",
     "GitPython<3.1.19",
     "hf-doc-builder>=0.3.0",
-    "huggingface-hub>=0.14.1,<1.0",
+    "huggingface-hub>=0.11.0,<1.0",
     "importlib_metadata",
     "ipadic>=1.0.0,<2.0",
     "isort>=5.5.4",
@@ -141,13 +140,12 @@ _deps = [
     "onnxconverter-common",
     "onnxruntime-tools>=1.4.2",
     "onnxruntime>=1.4.0",
-    "opencv-python",
     "optuna",
     "optax>=0.0.8,<=0.1.4",
     "packaging>=20.0",
     "parameterized",
     "phonemizer",
-    "protobuf<=3.20.3",
+    "protobuf<=3.20.2",
     "psutil",
     "pyyaml>=5.1",
     "pydantic",
@@ -414,10 +412,6 @@ extras["torchhub"] = deps_list(
     "tqdm",
 )
 
-extras["agents"] = deps_list(
-    "diffusers", "accelerate", "datasets", "torch", "sentencepiece", "opencv-python", "Pillow"
-)
-
 # when modifying the following list, make sure to update src/transformers/dependency_versions_check.py
 install_requires = [
     deps["importlib_metadata"] + ";python_version<'3.8'",  # importlib_metadata for Python versions that don't have it
@@ -434,7 +428,7 @@ install_requires = [
 
 setup(
     name="transformers",
-    version="4.30.0.dev0",  # expected format is one of x.y.z.dev0, or x.y.z.rc1 or x.y.z (no to dashes, yes to dots)
+    version="4.29.0",  # expected format is one of x.y.z.dev0, or x.y.z.rc1 or x.y.z (no to dashes, yes to dots)
     author="The Hugging Face team (past and future) with the help of all our contributors (https://github.com/huggingface/transformers/graphs/contributors)",
     author_email="transformers@huggingface.co",
     description="State-of-the-art Machine Learning for JAX, PyTorch and TensorFlow",

src/transformers/__init__.py

@@ -18,7 +18,7 @@
 # to defer the actual importing for when the objects are requested. This way `import transformers` provides the names
 # in the namespace without actually importing anything (and especially none of the backends).
 
-__version__ = "4.30.0.dev0"
+__version__ = "4.29.0"
 
 from typing import TYPE_CHECKING
 
@@ -464,7 +464,6 @@ _import_structure = {
     ],
     "models.splinter": ["SPLINTER_PRETRAINED_CONFIG_ARCHIVE_MAP", "SplinterConfig", "SplinterTokenizer"],
     "models.squeezebert": ["SQUEEZEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP", "SqueezeBertConfig", "SqueezeBertTokenizer"],
-    "models.swiftformer": ["SWIFTFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP", "SwiftFormerConfig"],
     "models.swin": ["SWIN_PRETRAINED_CONFIG_ARCHIVE_MAP", "SwinConfig"],
     "models.swin2sr": ["SWIN2SR_PRETRAINED_CONFIG_ARCHIVE_MAP", "Swin2SRConfig"],
     "models.swinv2": ["SWINV2_PRETRAINED_CONFIG_ARCHIVE_MAP", "Swinv2Config"],
@@ -2464,14 +2463,6 @@ else:
             "SqueezeBertPreTrainedModel",
         ]
     )
-    _import_structure["models.swiftformer"].extend(
-        [
-            "SWIFTFORMER_PRETRAINED_MODEL_ARCHIVE_LIST",
-            "SwiftFormerForImageClassification",
-            "SwiftFormerModel",
-            "SwiftFormerPreTrainedModel",
-        ]
-    )
     _import_structure["models.swin"].extend(
         [
             "SWIN_PRETRAINED_MODEL_ARCHIVE_LIST",
@@ -4230,7 +4221,6 @@ if TYPE_CHECKING:
     )
     from .models.splinter import SPLINTER_PRETRAINED_CONFIG_ARCHIVE_MAP, SplinterConfig, SplinterTokenizer
     from .models.squeezebert import SQUEEZEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP, SqueezeBertConfig, SqueezeBertTokenizer
-    from .models.swiftformer import SWIFTFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP, SwiftFormerConfig
     from .models.swin import SWIN_PRETRAINED_CONFIG_ARCHIVE_MAP, SwinConfig
     from .models.swin2sr import SWIN2SR_PRETRAINED_CONFIG_ARCHIVE_MAP, Swin2SRConfig
     from .models.swinv2 import SWINV2_PRETRAINED_CONFIG_ARCHIVE_MAP, Swinv2Config
@@ -5886,12 +5876,6 @@ if TYPE_CHECKING:
             SqueezeBertModule,
             SqueezeBertPreTrainedModel,
         )
-        from .models.swiftformer import (
-            SWIFTFORMER_PRETRAINED_MODEL_ARCHIVE_LIST,
-            SwiftFormerForImageClassification,
-            SwiftFormerModel,
-            SwiftFormerPreTrainedModel,
-        )
         from .models.swin import (
             SWIN_PRETRAINED_MODEL_ARCHIVE_LIST,
             SwinBackbone,

src/transformers/data/test_generation_utils.py

@@ -0,0 +1,99 @@
+# 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.
+
+import random
+import unittest
+
+import timeout_decorator
+
+from ..testing_utils import require_torch
+from ..utils import cached_property, is_torch_available
+
+
+if is_torch_available():
+    import torch
+
+    from ..models.marian import MarianConfig, MarianMTModel
+
+
+@require_torch
+class GenerationUtilsTest(unittest.TestCase):
+    @cached_property
+    def config(self):
+        config = MarianConfig.from_pretrained("sshleifer/tiny-marian-en-de")
+        return config
+
+    @cached_property
+    def model(self):
+        return MarianMTModel(self.config)
+
+    def test_postprocess_next_token_scores(self):
+        config = self.config
+        model = self.model
+        # Initialize an input id tensor with batch size 8 and sequence length 12
+        input_ids = torch.arange(0, 96, 1).view((8, 12))
+        eos = config.eos_token_id
+        bad_words_ids_test_cases = [[[299]], [[23, 24], [54]], [[config.eos_token_id]], []]
+        masked_scores = [
+            [(0, 299), (1, 299), (2, 299), (3, 299), (4, 299), (5, 299), (6, 299), (7, 299)],
+            [(1, 24), (0, 54), (1, 54), (2, 54), (3, 54), (4, 54), (5, 54), (6, 54), (7, 54)],
+            [(0, eos), (1, eos), (2, eos), (3, eos), (4, eos), (5, eos), (6, eos), (7, eos)],
+            [],
+        ]
+
+        for test_case_index, bad_words_ids in enumerate(bad_words_ids_test_cases):
+            # Initialize a scores tensor with batch size 8 and vocabulary size 300
+            scores = torch.rand((8, 300))
+            output = model.postprocess_next_token_scores(
+                scores,
+                input_ids,
+                0,
+                bad_words_ids,
+                13,
+                15,
+                config.max_length,
+                config.eos_token_id,
+                config.repetition_penalty,
+                32,
+                5,
+            )
+            for masked_score in masked_scores[test_case_index]:
+                self.assertTrue(output[masked_score[0], masked_score[1]] == -float("inf"))
+
+    @timeout_decorator.timeout(10)
+    def test_postprocess_next_token_scores_large_bad_words_list(self):
+        config = self.config
+        model = self.model
+        # Initialize an input id tensor with batch size 8 and sequence length 12
+        input_ids = torch.arange(0, 96, 1).view((8, 12))
+
+        bad_words_ids = []
+        for _ in range(100):
+            length_bad_word = random.randint(1, 4)
+            bad_words_ids.append(random.sample(range(1, 300), length_bad_word))
+
+        scores = torch.rand((8, 300))
+        _ = model.postprocess_next_token_scores(
+            scores,
+            input_ids,
+            0,
+            bad_words_ids,
+            13,
+            15,
+            config.max_length,
+            config.eos_token_id,
+            config.repetition_penalty,
+            32,
+            5,
+        )

src/transformers/dependency_versions_table.py

@@ -13,7 +13,6 @@ deps = {
     "datasets": "datasets!=2.5.0",
     "decord": "decord==0.6.0",
     "deepspeed": "deepspeed>=0.8.3",
-    "diffusers": "diffusers",
     "dill": "dill<0.3.5",
     "evaluate": "evaluate>=0.2.0",
     "fairscale": "fairscale>0.3",
@@ -25,7 +24,7 @@ deps = {
     "fugashi": "fugashi>=1.0",
     "GitPython": "GitPython<3.1.19",
     "hf-doc-builder": "hf-doc-builder>=0.3.0",
-    "huggingface-hub": "huggingface-hub>=0.14.1,<1.0",
+    "huggingface-hub": "huggingface-hub>=0.11.0,<1.0",
     "importlib_metadata": "importlib_metadata",
     "ipadic": "ipadic>=1.0.0,<2.0",
     "isort": "isort>=5.5.4",
@@ -42,13 +41,12 @@ deps = {
     "onnxconverter-common": "onnxconverter-common",
     "onnxruntime-tools": "onnxruntime-tools>=1.4.2",
     "onnxruntime": "onnxruntime>=1.4.0",
-    "opencv-python": "opencv-python",
     "optuna": "optuna",
     "optax": "optax>=0.0.8,<=0.1.4",
     "packaging": "packaging>=20.0",
     "parameterized": "parameterized",
     "phonemizer": "phonemizer",
-    "protobuf": "protobuf<=3.20.3",
+    "protobuf": "protobuf<=3.20.2",
     "psutil": "psutil",
     "pyyaml": "pyyaml>=5.1",
     "pydantic": "pydantic",

src/transformers/dynamic_module_utils.py

@@ -318,7 +318,7 @@ def get_cached_module_file(
 
     if len(new_files) > 0:
         new_files = "\n".join([f"- {f}" for f in new_files])
-        repo_type_str = "" if repo_type is None else f"{repo_type}s/"
+        repo_type_str = "" if repo_type is None else f"{repo_type}/"
         url = f"https://huggingface.co/{repo_type_str}{pretrained_model_name_or_path}"
         logger.warning(
             f"A new version of the following files was downloaded from {url}:\n{new_files}"

src/transformers/file_utils.py

@@ -72,7 +72,6 @@ from .utils import (
     get_cached_models,
     get_file_from_repo,
     get_full_repo_name,
-    get_torch_version,
     has_file,
     http_user_agent,
     is_apex_available,
@@ -126,4 +125,5 @@ from .utils import (
     to_numpy,
     to_py_obj,
     torch_only_method,
+    torch_version,
 )

src/transformers/modeling_utils.py

@@ -207,21 +207,29 @@ def get_parameter_dtype(parameter: Union[nn.Module, GenerationMixin, "ModuleUtil
         # if no floating dtype was found return whatever the first dtype is
         return last_dtype
 
-    else:
-        # For nn.DataParallel compatibility in PyTorch > 1.5
-        def find_tensor_attributes(module: nn.Module) -> List[Tuple[str, Tensor]]:
-            tuples = [(k, v) for k, v in module.__dict__.items() if torch.is_tensor(v)]
-            return tuples
+    for t in parameter.buffers():
+        last_dtype = t.dtype
+        if t.is_floating_point():
+            return t.dtype
 
-        gen = parameter._named_members(get_members_fn=find_tensor_attributes)
-        last_tuple = None
-        for tuple in gen:
-            last_tuple = tuple
-            if tuple[1].is_floating_point():
-                return tuple[1].dtype
+    if last_dtype is not None:
+        # if no floating dtype was found return whatever the first dtype is
+        return last_dtype
 
-        # fallback to the last dtype
-        return last_tuple[1].dtype
+    # For nn.DataParallel compatibility in PyTorch > 1.5
+    def find_tensor_attributes(module: nn.Module) -> List[Tuple[str, Tensor]]:
+        tuples = [(k, v) for k, v in module.__dict__.items() if torch.is_tensor(v)]
+        return tuples
+
+    gen = parameter._named_members(get_members_fn=find_tensor_attributes)
+    last_tuple = None
+    for tuple in gen:
+        last_tuple = tuple
+        if tuple[1].is_floating_point():
+            return tuple[1].dtype
+
+    # fallback to the last dtype
+    return last_tuple[1].dtype
 
 
 def get_state_dict_float_dtype(state_dict):

src/transformers/models/__init__.py

@@ -173,7 +173,6 @@ from . import (
     speecht5,
     splinter,
     squeezebert,
-    swiftformer,
     swin,
     swin2sr,
     swinv2,

src/transformers/models/auto/configuration_auto.py

@@ -174,7 +174,6 @@ CONFIG_MAPPING_NAMES = OrderedDict(
         ("speecht5", "SpeechT5Config"),
         ("splinter", "SplinterConfig"),
         ("squeezebert", "SqueezeBertConfig"),
-        ("swiftformer", "SwiftFormerConfig"),
         ("swin", "SwinConfig"),
         ("swin2sr", "Swin2SRConfig"),
         ("swinv2", "Swinv2Config"),
@@ -355,7 +354,6 @@ CONFIG_ARCHIVE_MAP_MAPPING_NAMES = OrderedDict(
         ("speecht5", "SPEECHT5_PRETRAINED_CONFIG_ARCHIVE_MAP"),
         ("splinter", "SPLINTER_PRETRAINED_CONFIG_ARCHIVE_MAP"),
         ("squeezebert", "SQUEEZEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP"),
-        ("swiftformer", "SWIFTFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP"),
         ("swin", "SWIN_PRETRAINED_CONFIG_ARCHIVE_MAP"),
         ("swin2sr", "SWIN2SR_PRETRAINED_CONFIG_ARCHIVE_MAP"),
         ("swinv2", "SWINV2_PRETRAINED_CONFIG_ARCHIVE_MAP"),
@@ -560,7 +558,6 @@ MODEL_NAMES_MAPPING = OrderedDict(
         ("speecht5", "SpeechT5"),
         ("splinter", "Splinter"),
         ("squeezebert", "SqueezeBERT"),
-        ("swiftformer", "SwiftFormer"),
         ("swin", "Swin Transformer"),
         ("swin2sr", "Swin2SR"),
         ("swinv2", "Swin Transformer V2"),

src/transformers/models/auto/feature_extraction_auto.py

@@ -78,7 +78,6 @@ FEATURE_EXTRACTOR_MAPPING_NAMES = OrderedDict(
         ("sew-d", "Wav2Vec2FeatureExtractor"),
         ("speech_to_text", "Speech2TextFeatureExtractor"),
         ("speecht5", "SpeechT5FeatureExtractor"),
-        ("swiftformer", "ViTFeatureExtractor"),
         ("swin", "ViTFeatureExtractor"),
         ("swinv2", "ViTFeatureExtractor"),
         ("table-transformer", "DetrFeatureExtractor"),

src/transformers/models/auto/image_processing_auto.py

@@ -87,7 +87,6 @@ IMAGE_PROCESSOR_MAPPING_NAMES = OrderedDict(
         ("resnet", "ConvNextImageProcessor"),
         ("sam", "SamImageProcessor"),
         ("segformer", "SegformerImageProcessor"),
-        ("swiftformer", "ViTImageProcessor"),
         ("swin", "ViTImageProcessor"),
         ("swin2sr", "Swin2SRImageProcessor"),
         ("swinv2", "ViTImageProcessor"),

src/transformers/models/auto/modeling_auto.py

@@ -167,7 +167,6 @@ MODEL_MAPPING_NAMES = OrderedDict(
         ("speecht5", "SpeechT5Model"),
         ("splinter", "SplinterModel"),
         ("squeezebert", "SqueezeBertModel"),
-        ("swiftformer", "SwiftFormerModel"),
         ("swin", "SwinModel"),
         ("swin2sr", "Swin2SRModel"),
         ("swinv2", "Swinv2Model"),
@@ -469,7 +468,6 @@ MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES = OrderedDict(
         ("regnet", "RegNetForImageClassification"),
         ("resnet", "ResNetForImageClassification"),
         ("segformer", "SegformerForImageClassification"),
-        ("swiftformer", "SwiftFormerForImageClassification"),
         ("swin", "SwinForImageClassification"),
         ("swinv2", "Swinv2ForImageClassification"),
         ("van", "VanForImageClassification"),

src/transformers/models/conditional_detr/image_processing_conditional_detr.py

@@ -1328,7 +1328,7 @@ class ConditionalDetrImageProcessor(BaseImageProcessor):
 
     # Copied from transformers.models.deformable_detr.image_processing_deformable_detr.DeformableDetrImageProcessor.post_process_object_detection with DeformableDetr->ConditionalDetr
     def post_process_object_detection(
-        self, outputs, threshold: float = 0.5, target_sizes: Union[TensorType, List[Tuple]] = None, top_k: int = 100
+        self, outputs, threshold: float = 0.5, target_sizes: Union[TensorType, List[Tuple]] = None
     ):
         """
         Converts the raw output of [`ConditionalDetrForObjectDetection`] into final bounding boxes in (top_left_x,
@@ -1342,8 +1342,6 @@ class ConditionalDetrImageProcessor(BaseImageProcessor):
             target_sizes (`torch.Tensor` or `List[Tuple[int, int]]`, *optional*):
                 Tensor of shape `(batch_size, 2)` or list of tuples (`Tuple[int, int]`) containing the target size
                 (height, width) of each image in the batch. If left to None, predictions will not be resized.
-            top_k (`int`, *optional*, defaults to 100):
-                Keep only top k bounding boxes before filtering by thresholding.
 
         Returns:
             `List[Dict]`: A list of dictionaries, each dictionary containing the scores, labels and boxes for an image
@@ -1358,9 +1356,7 @@ class ConditionalDetrImageProcessor(BaseImageProcessor):
                 )
 
         prob = out_logits.sigmoid()
-        prob = prob.view(out_logits.shape[0], -1)
-        k_value = min(top_k, prob.size(1))
-        topk_values, topk_indexes = torch.topk(prob, k_value, dim=1)
+        topk_values, topk_indexes = torch.topk(prob.view(out_logits.shape[0], -1), 100, dim=1)
         scores = topk_values
         topk_boxes = torch.div(topk_indexes, out_logits.shape[2], rounding_mode="floor")
         labels = topk_indexes % out_logits.shape[2]

src/transformers/models/decision_transformer/modeling_decision_transformer.py

@@ -118,9 +118,8 @@ class DecisionTransformerGPT2Attention(nn.Module):
             torch.tril(torch.ones((max_positions, max_positions), dtype=torch.bool)).view(
                 1, 1, max_positions, max_positions
             ),
-            persistent=False,
         )
-        self.register_buffer("masked_bias", torch.tensor(-1e4), persistent=False)
+        self.register_buffer("masked_bias", torch.tensor(-1e4))
 
         self.embed_dim = config.hidden_size
         self.num_heads = config.num_attention_heads
@@ -748,7 +747,6 @@ class DecisionTransformerPreTrainedModel(PreTrainedModel):
     main_input_name = "states"
     supports_gradient_checkpointing = False
     _keys_to_ignore_on_load_missing = [r"position_ids"]
-    _keys_to_ignore_on_load_unexpected = [r"h\.\d+\.attn\.masked_bias", r"h\.\d+\.attn\.bias"]
 
     def _init_weights(self, module):
         """Initialize the weights"""

src/transformers/models/deformable_detr/image_processing_deformable_detr.py

@@ -1325,7 +1325,7 @@ class DeformableDetrImageProcessor(BaseImageProcessor):
         return results
 
     def post_process_object_detection(
-        self, outputs, threshold: float = 0.5, target_sizes: Union[TensorType, List[Tuple]] = None, top_k: int = 100
+        self, outputs, threshold: float = 0.5, target_sizes: Union[TensorType, List[Tuple]] = None
     ):
         """
         Converts the raw output of [`DeformableDetrForObjectDetection`] into final bounding boxes in (top_left_x,
@@ -1339,8 +1339,6 @@ class DeformableDetrImageProcessor(BaseImageProcessor):
             target_sizes (`torch.Tensor` or `List[Tuple[int, int]]`, *optional*):
                 Tensor of shape `(batch_size, 2)` or list of tuples (`Tuple[int, int]`) containing the target size
                 (height, width) of each image in the batch. If left to None, predictions will not be resized.
-            top_k (`int`, *optional*, defaults to 100):
-                Keep only top k bounding boxes before filtering by thresholding.
 
         Returns:
             `List[Dict]`: A list of dictionaries, each dictionary containing the scores, labels and boxes for an image
@@ -1355,9 +1353,7 @@ class DeformableDetrImageProcessor(BaseImageProcessor):
                 )
 
         prob = out_logits.sigmoid()
-        prob = prob.view(out_logits.shape[0], -1)
-        k_value = min(top_k, prob.size(1))
-        topk_values, topk_indexes = torch.topk(prob, k_value, dim=1)
+        topk_values, topk_indexes = torch.topk(prob.view(out_logits.shape[0], -1), 100, dim=1)
         scores = topk_values
         topk_boxes = torch.div(topk_indexes, out_logits.shape[2], rounding_mode="floor")
         labels = topk_indexes % out_logits.shape[2]

src/transformers/models/gpt2/modeling_gpt2.py

@@ -131,9 +131,8 @@ class GPT2Attention(nn.Module):
             torch.tril(torch.ones((max_positions, max_positions), dtype=torch.bool)).view(
                 1, 1, max_positions, max_positions
             ),
-            persistent=False,
         )
-        self.register_buffer("masked_bias", torch.tensor(-1e4), persistent=False)
+        self.register_buffer("masked_bias", torch.tensor(-1e4))
 
         self.embed_dim = config.hidden_size
         self.num_heads = config.num_attention_heads
@@ -955,8 +954,7 @@ class GPT2Model(GPT2PreTrainedModel):
     GPT2_START_DOCSTRING,
 )
 class GPT2LMHeadModel(GPT2PreTrainedModel):
-    _keys_to_ignore_on_load_missing = [r"lm_head.weight"]
-    _keys_to_ignore_on_load_unexpected = [r"h\.\d+\.attn\.masked_bias", r"h\.\d+\.attn\.bias"]
+    _keys_to_ignore_on_load_missing = [r"attn.masked_bias", r"attn.bias", r"lm_head.weight"]
 
     def __init__(self, config):
         super().__init__(config)

src/transformers/models/swiftformer/__init__.py

@@ -1,67 +0,0 @@
-# Copyright 2023 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.
-from typing import TYPE_CHECKING
-
-from ...utils import (
-    OptionalDependencyNotAvailable,
-    _LazyModule,
-    is_torch_available,
-)
-
-
-_import_structure = {
-    "configuration_swiftformer": [
-        "SWIFTFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP",
-        "SwiftFormerConfig",
-        "SwiftFormerOnnxConfig",
-    ]
-}
-
-try:
-    if not is_torch_available():
-        raise OptionalDependencyNotAvailable()
-except OptionalDependencyNotAvailable:
-    pass
-else:
-    _import_structure["modeling_swiftformer"] = [
-        "SWIFTFORMER_PRETRAINED_MODEL_ARCHIVE_LIST",
-        "SwiftFormerForImageClassification",
-        "SwiftFormerModel",
-        "SwiftFormerPreTrainedModel",
-    ]
-
-if TYPE_CHECKING:
-    from .configuration_swiftformer import (
-        SWIFTFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP,
-        SwiftFormerConfig,
-        SwiftFormerOnnxConfig,
-    )
-
-    try:
-        if not is_torch_available():
-            raise OptionalDependencyNotAvailable()
-    except OptionalDependencyNotAvailable:
-        pass
-    else:
-        from .modeling_swiftformer import (
-            SWIFTFORMER_PRETRAINED_MODEL_ARCHIVE_LIST,
-            SwiftFormerForImageClassification,
-            SwiftFormerModel,
-            SwiftFormerPreTrainedModel,
-        )
-
-else:
-    import sys
-
-    sys.modules[__name__] = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)

src/transformers/models/swiftformer/configuration_swiftformer.py

@@ -1,136 +0,0 @@
-# coding=utf-8
-# Copyright 2023 MBZUAI and The HuggingFace Inc. 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.
-""" SwiftFormer model configuration"""
-
-from collections import OrderedDict
-from typing import Mapping
-
-from packaging import version
-
-from ...configuration_utils import PretrainedConfig
-from ...onnx import OnnxConfig
-from ...utils import logging
-
-
-logger = logging.get_logger(__name__)
-
-SWIFTFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP = {
-    "MBZUAI/swiftformer-xs": "https://huggingface.co/MBZUAI/swiftformer-xs/resolve/main/config.json",
-}
-
-
-class SwiftFormerConfig(PretrainedConfig):
-    r"""
-    This is the configuration class to store the configuration of a [`SwiftFormerModel`]. It is used to instantiate an
-    SwiftFormer model according to the specified arguments, defining the model architecture. Instantiating a
-    configuration with the defaults will yield a similar configuration to that of the SwiftFormer
-    [MBZUAI/swiftformer-xs](https://huggingface.co/MBZUAI/swiftformer-xs) architecture.
-
-    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
-    documentation from [`PretrainedConfig`] for more information.
-
-
-    Args:
-        num_channels (`int`, *optional*, defaults to 3):
-            The number of input channels
-        depths (`List[int]`, *optional*, defaults to `[3, 3, 6, 4]`):
-            Depth of each stage
-        embed_dims (`List[int]`, *optional*, defaults to `[48, 56, 112, 220]`):
-            The embedding dimension at each stage
-        mlp_ratio (`int`, *optional*, defaults to 4):
-            Ratio of size of the hidden dimensionality of an MLP to the dimensionality of its input.
-        downsamples (`List[bool]`, *optional*, defaults to `[True, True, True, True]`)
-            Whether or not to downsample inputs between two stages.
-        hidden_act (`str`, *optional*, defaults to `"gelu"`):
-            The non-linear activation function (string). `"gelu"`, `"relu"`, `"selu"` and `"gelu_new"` are supported.
-        down_patch_size (`int`, *optional*, defaults to 3):
-            The size of patches in downsampling layers.
-        down_stride (`int`, *optional*, defaults to 2):
-            The stride of convolution kernels in downsampling layers.
-        down_pad (`int`, *optional*, defaults to 1):
-            Padding in downsampling layers.
-        drop_path_rate (`float`, *optional*, defaults to 0.):
-            Rate at which to increase dropout probability in DropPath.
-        use_layer_scale (`bool`, *optional*, defaults to `True`):
-            Whether to scale outputs from token mixers.
-        layer_scale_init_value (`float`, *optional*, defaults to 1e-5):
-            Factor by which outputs from token mixers are scaled.
-        batch_norm_eps (`float`, *optional*, defaults to 1e-5):
-            The epsilon used by the batch normalization layers.
-
-
-    Example:
-
-    ```python
-    >>> from transformers import SwiftFormerConfig, SwiftFormerModel
-
-    >>> # Initializing a SwiftFormer swiftformer-base-patch16-224 style configuration
-    >>> configuration = SwiftFormerConfig()
-
-    >>> # Initializing a model (with random weights) from the swiftformer-base-patch16-224 style configuration
-    >>> model = SwiftFormerModel(configuration)
-
-    >>> # Accessing the model configuration
-    >>> configuration = model.config
-    ```"""
-    model_type = "swiftformer"
-
-    def __init__(
-        self,
-        num_channels=3,
-        depths=[3, 3, 6, 4],
-        embed_dims=[48, 56, 112, 220],
-        mlp_ratio=4,
-        downsamples=[True, True, True, True],
-        hidden_act="gelu",
-        down_patch_size=3,
-        down_stride=2,
-        down_pad=1,
-        drop_path_rate=0.0,
-        use_layer_scale=True,
-        layer_scale_init_value=1e-5,
-        batch_norm_eps=1e-5,
-        **kwargs,
-    ):
-        super().__init__(**kwargs)
-        self.num_channels = num_channels
-        self.depths = depths
-        self.embed_dims = embed_dims
-        self.mlp_ratio = mlp_ratio
-        self.downsamples = downsamples
-        self.hidden_act = hidden_act
-        self.down_patch_size = down_patch_size
-        self.down_stride = down_stride
-        self.down_pad = down_pad
-        self.drop_path_rate = drop_path_rate
-        self.use_layer_scale = use_layer_scale
-        self.layer_scale_init_value = layer_scale_init_value
-        self.batch_norm_eps = batch_norm_eps
-
-
-class SwiftFormerOnnxConfig(OnnxConfig):
-    torch_onnx_minimum_version = version.parse("1.11")
-
-    @property
-    def inputs(self) -> Mapping[str, Mapping[int, str]]:
-        return OrderedDict(
-            [
-                ("pixel_values", {0: "batch", 1: "num_channels", 2: "height", 3: "width"}),
-            ]
-        )
-
-    @property
-    def atol_for_validation(self) -> float:
-        return 1e-4

src/transformers/models/swiftformer/convert_swiftformer_original_to_hf.py

@@ -1,176 +0,0 @@
-# coding=utf-8
-# Copyright 2023 The HuggingFace Inc. team.
-#
-# 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.
-"""Convert SwiftFormer checkpoints from the original implementation."""
-
-
-import argparse
-import json
-from pathlib import Path
-
-import requests
-import torch
-from huggingface_hub import hf_hub_download
-from PIL import Image
-
-from transformers import (
-    SwiftFormerConfig,
-    SwiftFormerForImageClassification,
-    ViTImageProcessor,
-)
-from transformers.utils import logging
-
-
-logging.set_verbosity_info()
-logger = logging.get_logger(__name__)
-
-device = torch.device("cpu")
-
-
-# We will verify our results on an image of cute cats
-def prepare_img():
-    url = "http://images.cocodataset.org/val2017/000000039769.jpg"
-    im = Image.open(requests.get(url, stream=True).raw)
-    return im
-
-
-def get_expected_output(swiftformer_name):
-    if swiftformer_name == "swiftformer_xs":
-        return torch.tensor([-2.1703e00, 2.1107e00, -2.0811e00, 8.8685e-01, 2.4360e-01])
-
-    elif swiftformer_name == "swiftformer_s":
-        return torch.tensor([3.9636e-01, 2.3478e-01, -1.6963e00, -1.7381e00, -8.6337e-01])
-
-    elif swiftformer_name == "swiftformer_l1":
-        return torch.tensor([-4.2768e-01, -4.7429e-01, -1.0897e00, -1.0248e00, 3.5523e-02])
-
-    elif swiftformer_name == "swiftformer_l3":
-        return torch.tensor([-2.5330e-01, 2.4211e-01, -6.0185e-01, -8.2789e-01, -6.0446e-02])
-
-
-def rename_key(dct, old, new):
-    val = dct.pop(old)
-    dct[new] = val
-
-
-def create_rename_keys(state_dict):
-    rename_keys = []
-    for k in state_dict.keys():
-        k_new = k
-        if ".pwconv" in k:
-            k_new = k_new.replace(".pwconv", ".point_wise_conv")
-        if ".dwconv" in k:
-            k_new = k_new.replace(".dwconv", ".depth_wise_conv")
-        if ".Proj." in k:
-            k_new = k_new.replace(".Proj.", ".proj.")
-        if "patch_embed" in k_new:
-            k_new = k_new.replace("patch_embed", "swiftformer.patch_embed.patch_embedding")
-        if "network" in k_new:
-            ls = k_new.split(".")
-            if ls[2].isdigit():
-                k_new = "swiftformer.encoder.network." + ls[1] + ".blocks." + ls[2] + "." + ".".join(ls[3:])
-            else:
-                k_new = k_new.replace("network", "swiftformer.encoder.network")
-        rename_keys.append((k, k_new))
-    return rename_keys
-
-
-@torch.no_grad()
-def convert_swiftformer_checkpoint(swiftformer_name, pytorch_dump_folder_path, original_ckpt):
-    """
-    Copy/paste/tweak model's weights to our SwiftFormer structure.
-    """
-
-    # define default SwiftFormer configuration
-    config = SwiftFormerConfig()
-
-    # dataset (ImageNet-21k only or also fine-tuned on ImageNet 2012), patch_size and image_size
-    config.num_labels = 1000
-    repo_id = "huggingface/label-files"
-    filename = "imagenet-1k-id2label.json"
-    id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
-    id2label = {int(k): v for k, v in id2label.items()}
-    config.id2label = id2label
-    config.label2id = {v: k for k, v in id2label.items()}
-
-    # size of the architecture
-    if swiftformer_name == "swiftformer_xs":
-        config.depths = [3, 3, 6, 4]
-        config.embed_dims = [48, 56, 112, 220]
-
-    elif swiftformer_name == "swiftformer_s":
-        config.depths = [3, 3, 9, 6]
-        config.embed_dims = [48, 64, 168, 224]
-
-    elif swiftformer_name == "swiftformer_l1":
-        config.depths = [4, 3, 10, 5]
-        config.embed_dims = [48, 96, 192, 384]
-
-    elif swiftformer_name == "swiftformer_l3":
-        config.depths = [4, 4, 12, 6]
-        config.embed_dims = [64, 128, 320, 512]
-
-    # load state_dict of original model, remove and rename some keys
-    if original_ckpt:
-        if original_ckpt.startswith("https"):
-            checkpoint = torch.hub.load_state_dict_from_url(original_ckpt, map_location="cpu", check_hash=True)
-        else:
-            checkpoint = torch.load(original_ckpt, map_location="cpu")
-    state_dict = checkpoint
-
-    rename_keys = create_rename_keys(state_dict)
-    for rename_key_src, rename_key_dest in rename_keys:
-        rename_key(state_dict, rename_key_src, rename_key_dest)
-
-    # load HuggingFace model
-    hf_model = SwiftFormerForImageClassification(config).eval()
-    hf_model.load_state_dict(state_dict)
-
-    # prepare test inputs
-    image = prepare_img()
-    processor = ViTImageProcessor.from_pretrained("preprocessor_config")
-    inputs = processor(images=image, return_tensors="pt")
-
-    # compare outputs from both models
-    timm_logits = get_expected_output(swiftformer_name)
-    hf_logits = hf_model(inputs["pixel_values"]).logits
-
-    assert hf_logits.shape == torch.Size([1, 1000])
-    assert torch.allclose(hf_logits[0, 0:5], timm_logits, atol=1e-3)
-
-    Path(pytorch_dump_folder_path).mkdir(exist_ok=True)
-    print(f"Saving model {swiftformer_name} to {pytorch_dump_folder_path}")
-    hf_model.save_pretrained(pytorch_dump_folder_path)
-
-
-if __name__ == "__main__":
-    parser = argparse.ArgumentParser()
-    # Required parameters
-    parser.add_argument(
-        "--swiftformer_name",
-        default="swiftformer_xs",
-        choices=["swiftformer_xs", "swiftformer_s", "swiftformer_l1", "swiftformer_l3"],
-        type=str,
-        help="Name of the SwiftFormer model you'd like to convert.",
-    )
-    parser.add_argument(
-        "--pytorch_dump_folder_path",
-        default="./converted_outputs/",
-        type=str,
-        help="Path to the output PyTorch model directory.",
-    )
-    parser.add_argument("--original_ckpt", default=None, type=str, help="Path to the original model checkpoint.")
-
-    args = parser.parse_args()
-    convert_swiftformer_checkpoint(args.swiftformer_name, args.pytorch_dump_folder_path, args.original_ckpt)

src/transformers/models/swiftformer/modeling_swiftformer.py

@@ -1,623 +0,0 @@
-# coding=utf-8
-# Copyright 2023 MBZUAI and The HuggingFace Inc. 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.
-""" PyTorch SwiftFormer model."""
-
-
-import collections.abc
-from typing import Optional, Tuple, Union
-
-import torch
-import torch.utils.checkpoint
-from torch import nn
-from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss
-
-from ...activations import ACT2CLS
-from ...modeling_outputs import (
-    BaseModelOutputWithNoAttention,
-    ImageClassifierOutputWithNoAttention,
-)
-from ...modeling_utils import PreTrainedModel
-from ...utils import (
-    add_code_sample_docstrings,
-    add_start_docstrings,
-    add_start_docstrings_to_model_forward,
-    logging,
-)
-from .configuration_swiftformer import SwiftFormerConfig
-
-
-logger = logging.get_logger(__name__)
-
-# General docstring
-_CONFIG_FOR_DOC = "SwiftFormerConfig"
-
-# Base docstring
-_CHECKPOINT_FOR_DOC = "MBZUAI/swiftformer-xs"
-_EXPECTED_OUTPUT_SHAPE = [1, 220, 7, 7]
-
-# Image classification docstring
-_IMAGE_CLASS_CHECKPOINT = "MBZUAI/swiftformer-xs"
-_IMAGE_CLASS_EXPECTED_OUTPUT = "tabby, tabby cat"
-
-
-SWIFTFORMER_PRETRAINED_MODEL_ARCHIVE_LIST = [
-    "MBZUAI/swiftformer-xs",
-    # See all SwiftFormer models at https://huggingface.co/models?filter=swiftformer
-]
-
-
-class SwiftFormerPatchEmbedding(nn.Module):
-    """
-    Patch Embedding Layer constructed of two 2D convolutional layers.
-
-    Input: tensor of shape `[batch_size, in_channels, height, width]`
-
-    Output: tensor of shape `[batch_size, out_channels, height/4, width/4]`
-    """
-
-    def __init__(self, config: SwiftFormerConfig):
-        super().__init__()
-
-        in_chs = config.num_channels
-        out_chs = config.embed_dims[0]
-        self.patch_embedding = nn.Sequential(
-            nn.Conv2d(in_chs, out_chs // 2, kernel_size=3, stride=2, padding=1),
-            nn.BatchNorm2d(out_chs // 2, eps=config.batch_norm_eps),
-            nn.ReLU(),
-            nn.Conv2d(out_chs // 2, out_chs, kernel_size=3, stride=2, padding=1),
-            nn.BatchNorm2d(out_chs, eps=config.batch_norm_eps),
-            nn.ReLU(),
-        )
-
-    def forward(self, x):
-        return self.patch_embedding(x)
-
-
-# Copied from transformers.models.beit.modeling_beit.drop_path
-def drop_path(x, drop_prob: float = 0.0, training: bool = False):
-    """
-    Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks).
-
-    Comment by Ross Wightman: This is the same as the DropConnect impl I created for EfficientNet, etc networks,
-    however, the original name is misleading as 'Drop Connect' is a different form of dropout in a separate paper...
-    See discussion: https://github.com/tensorflow/tpu/issues/494#issuecomment-532968956 ... I've opted for changing the
-    layer and argument names to 'drop path' rather than mix DropConnect as a layer name and use 'survival rate' as the
-    argument.
-    """
-    if drop_prob == 0.0 or not training:
-        return input
-    keep_prob = 1 - drop_prob
-    shape = (input.shape[0],) + (1,) * (input.ndim - 1)  # work with diff dim tensors, not just 2D ConvNets
-    random_tensor = keep_prob + torch.rand(shape, dtype=input.dtype, device=input.device)
-    random_tensor.floor_()  # binarize
-    output = input.div(keep_prob) * random_tensor
-    return output
-
-
-# Copied from transformers.models.beit.modeling_beit.BeitDropPath with Beit->Swiftformer
-class SwiftFormerDropPath(nn.Module):
-    """Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks)."""
-
-    def __init__(self, drop_prob: Optional[float] = None) -> None:
-        super().__init__()
-        self.drop_prob = drop_prob
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        return drop_path(hidden_states, self.drop_prob, self.training)
-
-    def extra_repr(self) -> str:
-        return "p={}".format(self.drop_prob)
-
-
-class SwiftFormerEmbeddings(nn.Module):
-    """
-    Embeddings layer consisting of a single 2D convolutional and batch normalization layer.
-
-    Input: tensor of shape `[batch_size, channels, height, width]`
-
-    Output: tensor of shape `[batch_size, channels, height/stride, width/stride]`
-    """
-
-    def __init__(self, config: SwiftFormerConfig, index: int):
-        super().__init__()
-
-        patch_size = config.down_patch_size
-        stride = config.down_stride
-        padding = config.down_pad
-        embed_dims = config.embed_dims
-
-        in_chans = embed_dims[index]
-        embed_dim = embed_dims[index + 1]
-
-        patch_size = patch_size if isinstance(patch_size, collections.abc.Iterable) else (patch_size, patch_size)
-        stride = stride if isinstance(stride, collections.abc.Iterable) else (stride, stride)
-        padding = padding if isinstance(padding, collections.abc.Iterable) else (padding, padding)
-
-        self.proj = nn.Conv2d(in_chans, embed_dim, kernel_size=patch_size, stride=stride, padding=padding)
-        self.norm = nn.BatchNorm2d(embed_dim, eps=config.batch_norm_eps)
-
-    def forward(self, x):
-        x = self.proj(x)
-        x = self.norm(x)
-        return x
-
-
-class SwiftFormerConvEncoder(nn.Module):
-    """
-    `SwiftFormerConvEncoder` with 3*3 and 1*1 convolutions.
-
-    Input: tensor of shape `[batch_size, channels, height, width]`
-
-    Output: tensor of shape `[batch_size, channels, height, width]`
-    """
-
-    def __init__(self, config: SwiftFormerConfig, dim: int):
-        super().__init__()
-        hidden_dim = int(config.mlp_ratio * dim)
-
-        self.depth_wise_conv = nn.Conv2d(dim, dim, kernel_size=3, padding=1, groups=dim)
-        self.norm = nn.BatchNorm2d(dim, eps=config.batch_norm_eps)
-        self.point_wise_conv1 = nn.Conv2d(dim, hidden_dim, kernel_size=1)
-        self.act = nn.GELU()
-        self.point_wise_conv2 = nn.Conv2d(hidden_dim, dim, kernel_size=1)
-        self.drop_path = nn.Identity()
-        self.layer_scale = nn.Parameter(torch.ones(dim).unsqueeze(-1).unsqueeze(-1), requires_grad=True)
-
-    def forward(self, x):
-        input = x
-        x = self.depth_wise_conv(x)
-        x = self.norm(x)
-        x = self.point_wise_conv1(x)
-        x = self.act(x)
-        x = self.point_wise_conv2(x)
-        x = input + self.drop_path(self.layer_scale * x)
-        return x
-
-
-class SwiftFormerMlp(nn.Module):
-    """
-    MLP layer with 1*1 convolutions.
-
-    Input: tensor of shape `[batch_size, channels, height, width]`
-
-    Output: tensor of shape `[batch_size, channels, height, width]`
-    """
-
-    def __init__(self, config: SwiftFormerConfig, in_features: int):
-        super().__init__()
-        hidden_features = int(in_features * config.mlp_ratio)
-        self.norm1 = nn.BatchNorm2d(in_features, eps=config.batch_norm_eps)
-        self.fc1 = nn.Conv2d(in_features, hidden_features, 1)
-        act_layer = ACT2CLS[config.hidden_act]
-        self.act = act_layer()
-        self.fc2 = nn.Conv2d(hidden_features, in_features, 1)
-        self.drop = nn.Dropout(p=0.0)
-
-    def forward(self, x):
-        x = self.norm1(x)
-        x = self.fc1(x)
-        x = self.act(x)
-        x = self.drop(x)
-        x = self.fc2(x)
-        x = self.drop(x)
-        return x
-
-
-class SwiftFormerEfficientAdditiveAttention(nn.Module):
-    """
-    Efficient Additive Attention module for SwiftFormer.
-
-    Input: tensor of shape `[batch_size, channels, height, width]`
-
-    Output: tensor of shape `[batch_size, channels, height, width]`
-    """
-
-    def __init__(self, config: SwiftFormerConfig, dim: int = 512):
-        super().__init__()
-
-        self.to_query = nn.Linear(dim, dim)
-        self.to_key = nn.Linear(dim, dim)
-
-        self.w_g = nn.Parameter(torch.randn(dim, 1))
-        self.scale_factor = dim**-0.5
-        self.proj = nn.Linear(dim, dim)
-        self.final = nn.Linear(dim, dim)
-
-    def forward(self, x):
-        query = self.to_query(x)
-        key = self.to_key(x)
-
-        query = torch.nn.functional.normalize(query, dim=-1)
-        key = torch.nn.functional.normalize(key, dim=-1)
-
-        query_weight = query @ self.w_g
-        scaled_query_weight = query_weight * self.scale_factor
-        scaled_query_weight = scaled_query_weight.softmax(dim=-1)
-
-        global_queries = torch.sum(scaled_query_weight * query, dim=1)
-        global_queries = global_queries.unsqueeze(1).repeat(1, key.shape[1], 1)
-
-        out = self.proj(global_queries * key) + query
-        out = self.final(out)
-
-        return out
-
-
-class SwiftFormerLocalRepresentation(nn.Module):
-    """
-    Local Representation module for SwiftFormer that is implemented by 3*3 depth-wise and point-wise convolutions.
-
-    Input: tensor of shape `[batch_size, channels, height, width]`
-
-    Output: tensor of shape `[batch_size, channels, height, width]`
-    """
-
-    def __init__(self, config: SwiftFormerConfig, dim: int):
-        super().__init__()
-
-        self.depth_wise_conv = nn.Conv2d(dim, dim, kernel_size=3, padding=1, groups=dim)
-        self.norm = nn.BatchNorm2d(dim, eps=config.batch_norm_eps)
-        self.point_wise_conv1 = nn.Conv2d(dim, dim, kernel_size=1)
-        self.act = nn.GELU()
-        self.point_wise_conv2 = nn.Conv2d(dim, dim, kernel_size=1)
-        self.drop_path = nn.Identity()
-        self.layer_scale = nn.Parameter(torch.ones(dim).unsqueeze(-1).unsqueeze(-1), requires_grad=True)
-
-    def forward(self, x):
-        input = x
-        x = self.depth_wise_conv(x)
-        x = self.norm(x)
-        x = self.point_wise_conv1(x)
-        x = self.act(x)
-        x = self.point_wise_conv2(x)
-        x = input + self.drop_path(self.layer_scale * x)
-        return x
-
-
-class SwiftFormerEncoderBlock(nn.Module):
-    """
-    SwiftFormer Encoder Block for SwiftFormer. It consists of (1) Local representation module, (2)
-    SwiftFormerEfficientAdditiveAttention, and (3) MLP block.
-
-    Input: tensor of shape `[batch_size, channels, height, width]`
-
-    Output: tensor of shape `[batch_size, channels,height, width]`
-    """
-
-    def __init__(self, config: SwiftFormerConfig, dim: int, drop_path: float = 0.0) -> None:
-        super().__init__()
-
-        layer_scale_init_value = config.layer_scale_init_value
-        use_layer_scale = config.use_layer_scale
-
-        self.local_representation = SwiftFormerLocalRepresentation(config, dim=dim)
-        self.attn = SwiftFormerEfficientAdditiveAttention(config, dim=dim)
-        self.linear = SwiftFormerMlp(config, in_features=dim)
-        self.drop_path = SwiftFormerDropPath(drop_path) if drop_path > 0.0 else nn.Identity()
-        self.use_layer_scale = use_layer_scale
-        if use_layer_scale:
-            self.layer_scale_1 = nn.Parameter(
-                layer_scale_init_value * torch.ones(dim).unsqueeze(-1).unsqueeze(-1), requires_grad=True
-            )
-            self.layer_scale_2 = nn.Parameter(
-                layer_scale_init_value * torch.ones(dim).unsqueeze(-1).unsqueeze(-1), requires_grad=True
-            )
-
-    def forward(self, x):
-        x = self.local_representation(x)
-        batch_size, channels, height, width = x.shape
-        if self.use_layer_scale:
-            x = x + self.drop_path(
-                self.layer_scale_1
-                * self.attn(x.permute(0, 2, 3, 1).reshape(batch_size, height * width, channels))
-                .reshape(batch_size, height, width, channels)
-                .permute(0, 3, 1, 2)
-            )
-            x = x + self.drop_path(self.layer_scale_2 * self.linear(x))
-
-        else:
-            x = x + self.drop_path(
-                self.attn(x.permute(0, 2, 3, 1).reshape(batch_size, height * width, channels))
-                .reshape(batch_size, height, width, channels)
-                .permute(0, 3, 1, 2)
-            )
-            x = x + self.drop_path(self.linear(x))
-        return x
-
-
-class SwiftFormerStage(nn.Module):
-    """
-    A Swiftformer stage consisting of a series of `SwiftFormerConvEncoder` blocks and a final
-    `SwiftFormerEncoderBlock`.
-
-    Input: tensor in shape `[batch_size, channels, height, width]`
-
-    Output: tensor in shape `[batch_size, channels, height, width]`
-    """
-
-    def __init__(self, config: SwiftFormerConfig, index: int) -> None:
-        super().__init__()
-
-        layer_depths = config.depths
-        dim = config.embed_dims[index]
-        depth = layer_depths[index]
-
-        blocks = []
-        for block_idx in range(depth):
-            block_dpr = config.drop_path_rate * (block_idx + sum(layer_depths[:index])) / (sum(layer_depths) - 1)
-
-            if depth - block_idx <= 1:
-                blocks.append(SwiftFormerEncoderBlock(config, dim=dim, drop_path=block_dpr))
-            else:
-                blocks.append(SwiftFormerConvEncoder(config, dim=dim))
-
-        self.blocks = nn.ModuleList(blocks)
-
-    def forward(self, input):
-        for block in self.blocks:
-            input = block(input)
-        return input
-
-
-class SwiftFormerEncoder(nn.Module):
-    def __init__(self, config: SwiftFormerConfig) -> None:
-        super().__init__()
-        self.config = config
-
-        embed_dims = config.embed_dims
-        downsamples = config.downsamples
-        layer_depths = config.depths
-
-        # Transformer model
-        network = []
-        for i in range(len(layer_depths)):
-            stage = SwiftFormerStage(config=config, index=i)
-            network.append(stage)
-            if i >= len(layer_depths) - 1:
-                break
-            if downsamples[i] or embed_dims[i] != embed_dims[i + 1]:
-                # downsampling between two stages
-                network.append(SwiftFormerEmbeddings(config, index=i))
-        self.network = nn.ModuleList(network)
-
-        self.gradient_checkpointing = False
-
-    def forward(
-        self,
-        hidden_states: torch.Tensor,
-        output_hidden_states: Optional[bool] = None,
-        return_dict: Optional[bool] = None,
-    ) -> Union[tuple, BaseModelOutputWithNoAttention]:
-        output_hidden_states = (
-            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
-        )
-        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
-
-        all_hidden_states = (hidden_states,) if output_hidden_states else None
-
-        for block in self.network:
-            hidden_states = block(hidden_states)
-            if output_hidden_states:
-                all_hidden_states = all_hidden_states + (hidden_states,)
-
-        if not return_dict:
-            return tuple(v for v in [hidden_states, all_hidden_states] if v is not None)
-
-        return BaseModelOutputWithNoAttention(
-            last_hidden_state=hidden_states,
-            hidden_states=all_hidden_states,
-        )
-
-
-class SwiftFormerPreTrainedModel(PreTrainedModel):
-    """
-    An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained
-    models.
-    """
-
-    config_class = SwiftFormerConfig
-    base_model_prefix = "swiftformer"
-    main_input_name = "pixel_values"
-    supports_gradient_checkpointing = True
-
-    def _init_weights(self, module: Union[nn.Linear, nn.Conv2d, nn.LayerNorm]) -> None:
-        """Initialize the weights"""
-        if isinstance(module, (nn.Conv2d, nn.Linear)):
-            nn.init.trunc_normal_(module.weight, std=0.02)
-            if module.bias is not None:
-                nn.init.constant_(module.bias, 0)
-        elif isinstance(module, (nn.LayerNorm)):
-            nn.init.constant_(module.bias, 0)
-            nn.init.constant_(module.weight, 1.0)
-
-    def _set_gradient_checkpointing(self, module: SwiftFormerEncoder, value: bool = False) -> None:
-        if isinstance(module, SwiftFormerEncoder):
-            module.gradient_checkpointing = value
-
-
-SWIFTFORMER_START_DOCSTRING = r"""
-    This model is a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass. Use it
-    as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and
-    behavior.
-
-    Parameters:
-        config ([`SwiftFormerConfig`]): Model configuration class with all the parameters of the model.
-            Initializing with a config file does not load the weights associated with the model, only the
-            configuration. Check out the [`~PreTrainedModel.from_pretrained`] method to load the model weights.
-"""
-
-SWIFTFORMER_INPUTS_DOCSTRING = r"""
-    Args:
-        pixel_values (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)`):
-            Pixel values. Pixel values can be obtained using [`AutoImageProcessor`]. See [`ViTImageProcessor.__call__`]
-            for details.
-
-        output_hidden_states (`bool`, *optional*):
-            Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
-            more detail.
-        return_dict (`bool`, *optional*):
-            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
-"""
-
-
-@add_start_docstrings(
-    "The bare SwiftFormer Model transformer outputting raw hidden-states without any specific head on top.",
-    SWIFTFORMER_START_DOCSTRING,
-)
-class SwiftFormerModel(SwiftFormerPreTrainedModel):
-    def __init__(self, config: SwiftFormerConfig):
-        super().__init__(config)
-        self.config = config
-
-        self.patch_embed = SwiftFormerPatchEmbedding(config)
-        self.encoder = SwiftFormerEncoder(config)
-
-        # Initialize weights and apply final processing
-        self.post_init()
-
-    @add_start_docstrings_to_model_forward(SWIFTFORMER_INPUTS_DOCSTRING)
-    @add_code_sample_docstrings(
-        checkpoint=_CHECKPOINT_FOR_DOC,
-        output_type=BaseModelOutputWithNoAttention,
-        config_class=_CONFIG_FOR_DOC,
-        modality="vision",
-        expected_output=_EXPECTED_OUTPUT_SHAPE,
-    )
-    def forward(
-        self,
-        pixel_values: Optional[torch.Tensor] = None,
-        output_hidden_states: Optional[bool] = None,
-        return_dict: Optional[bool] = None,
-    ) -> Union[Tuple, BaseModelOutputWithNoAttention]:
-        r""" """
-
-        output_hidden_states = (
-            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
-        )
-        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
-
-        if pixel_values is None:
-            raise ValueError("You have to specify pixel_values")
-
-        embedding_output = self.patch_embed(pixel_values)
-        encoder_outputs = self.encoder(
-            embedding_output,
-            output_hidden_states=output_hidden_states,
-            return_dict=return_dict,
-        )
-
-        if not return_dict:
-            return tuple(v for v in encoder_outputs if v is not None)
-
-        return BaseModelOutputWithNoAttention(
-            last_hidden_state=encoder_outputs.last_hidden_state,
-            hidden_states=encoder_outputs.hidden_states,
-        )
-
-
-@add_start_docstrings(
-    """
-    SwiftFormer Model transformer with an image classification head on top (e.g. for ImageNet).
-    """,
-    SWIFTFORMER_START_DOCSTRING,
-)
-class SwiftFormerForImageClassification(SwiftFormerPreTrainedModel):
-    def __init__(self, config: SwiftFormerConfig) -> None:
-        super().__init__(config)
-
-        embed_dims = config.embed_dims
-
-        self.num_labels = config.num_labels
-        self.swiftformer = SwiftFormerModel(config)
-
-        # Classifier head
-        self.norm = nn.BatchNorm2d(embed_dims[-1], eps=config.batch_norm_eps)
-        self.head = nn.Linear(embed_dims[-1], self.num_labels) if self.num_labels > 0 else nn.Identity()
-        self.dist_head = nn.Linear(embed_dims[-1], self.num_labels) if self.num_labels > 0 else nn.Identity()
-
-        # Initialize weights and apply final processing
-        self.post_init()
-
-    @add_start_docstrings_to_model_forward(SWIFTFORMER_INPUTS_DOCSTRING)
-    @add_code_sample_docstrings(
-        checkpoint=_IMAGE_CLASS_CHECKPOINT,
-        output_type=ImageClassifierOutputWithNoAttention,
-        config_class=_CONFIG_FOR_DOC,
-        expected_output=_IMAGE_CLASS_EXPECTED_OUTPUT,
-    )
-    def forward(
-        self,
-        pixel_values: Optional[torch.Tensor] = None,
-        labels: Optional[torch.Tensor] = None,
-        output_hidden_states: Optional[bool] = None,
-        return_dict: Optional[bool] = None,
-    ) -> Union[tuple, ImageClassifierOutputWithNoAttention]:
-        r"""
-        labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
-            Labels for computing the image classification/regression loss. Indices should be in `[0, ...,
-            config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
-            `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
-        """
-        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
-
-        # run base model
-        outputs = self.swiftformer(
-            pixel_values,
-            output_hidden_states=output_hidden_states,
-            return_dict=return_dict,
-        )
-
-        sequence_output = outputs.last_hidden_state if return_dict else outputs[0]
-
-        # run classification head
-        sequence_output = self.norm(sequence_output)
-        sequence_output = sequence_output.flatten(2).mean(-1)
-        cls_out = self.head(sequence_output)
-        distillation_out = self.dist_head(sequence_output)
-        logits = (cls_out + distillation_out) / 2
-
-        # calculate loss
-        loss = None
-        if labels is not None:
-            if self.config.problem_type is None:
-                if self.num_labels == 1:
-                    self.config.problem_type = "regression"
-                elif self.num_labels > 1 and (labels.dtype == torch.long or labels.dtype == torch.int):
-                    self.config.problem_type = "single_label_classification"
-                else:
-                    self.config.problem_type = "multi_label_classification"
-
-            if self.config.problem_type == "regression":
-                loss_fct = MSELoss()
-                if self.num_labels == 1:
-                    loss = loss_fct(logits.squeeze(), labels.squeeze())
-                else:
-                    loss = loss_fct(logits, labels)
-            elif self.config.problem_type == "single_label_classification":
-                loss_fct = CrossEntropyLoss()
-                loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1))
-            elif self.config.problem_type == "multi_label_classification":
-                loss_fct = BCEWithLogitsLoss()
-                loss = loss_fct(logits, labels)
-
-        if not return_dict:
-            output = (logits,) + outputs[1:]
-            return ((loss,) + output) if loss is not None else output
-
-        return ImageClassifierOutputWithNoAttention(
-            loss=loss,
-            logits=logits,
-            hidden_states=outputs.hidden_states,
-        )

src/transformers/models/whisper/modeling_flax_whisper.py

@@ -1515,9 +1515,7 @@ class FlaxWhisperForAudioClassificationModule(nn.Module):
     gradient_checkpointing: bool = False
 
     def setup(self) -> None:
-        self.encoder = FlaxWhisperEncoder(
-            config=self.config, dtype=self.dtype, gradient_checkpointing=self.gradient_checkpointing
-        )
+        self.encoder = FlaxWhisperEncoder(config=self.config, dtype=self.dtype)
         self.config.is_encoder_decoder = False
         num_layers = self.config.num_hidden_layers + 1
         if self.config.use_weighted_layer_sum:

src/transformers/onnx/config.py

@@ -232,9 +232,9 @@ class OnnxConfig(ABC):
             `bool`: Whether the installed version of PyTorch is compatible with the model.
         """
         if is_torch_available():
-            from transformers.utils import get_torch_version
+            from transformers.utils import torch_version
 
-            return get_torch_version() >= self.torch_onnx_minimum_version
+            return torch_version >= self.torch_onnx_minimum_version
         else:
             return False
 

src/transformers/onnx/convert.py

@@ -334,12 +334,12 @@ def export(
         preprocessor = tokenizer
 
     if is_torch_available():
-        from ..utils import get_torch_version
+        from ..utils import torch_version
 
         if not config.is_torch_support_available:
             logger.warning(
                 f"Unsupported PyTorch version for this model. Minimum required is {config.torch_onnx_minimum_version},"
-                f" got: {get_torch_version()}"
+                f" got: {torch_version}"
             )
 
     if is_torch_available() and issubclass(type(model), PreTrainedModel):

src/transformers/testing_utils.py

@@ -14,7 +14,6 @@
 
 import collections
 import contextlib
-import doctest
 import functools
 import inspect
 import logging
@@ -31,23 +30,11 @@ import unittest
 from collections.abc import Mapping
 from io import StringIO
 from pathlib import Path
-from typing import Iterable, Iterator, List, Optional, Union
+from typing import Iterator, List, Optional, Union
 from unittest import mock
 
 import huggingface_hub
 import requests
-from _pytest.doctest import (
-    Module,
-    _get_checker,
-    _get_continue_on_failure,
-    _get_runner,
-    _is_mocked,
-    _patch_unwrap_mock_aware,
-    get_optionflags,
-    import_path,
-)
-from _pytest.outcomes import skip
-from pytest import DoctestItem
 
 from transformers import logging as transformers_logging
 
@@ -1825,162 +1812,3 @@ def run_test_in_subprocess(test_case, target_func, inputs=None, timeout=None):
 
     if results["error"] is not None:
         test_case.fail(f'{results["error"]}')
-
-
-"""
-The following contains utils to run the documentation tests without having to overwrite any files.
-
-The `preprocess_string` function adds `# doctest: +IGNORE_RESULT` markers on the fly anywhere a `load_dataset` call is
-made as a print would otherwise fail the corresonding line.
-
-To skip cuda tests, make sure to call `SKIP_CUDA_DOCTEST=1 pytest --doctest-modules <path_to_files_to_test>
-"""
-
-
-def preprocess_string(string, skip_cuda_tests):
-    """Prepare a docstring or a `.mdx` file to be run by doctest.
-
-    The argument `string` would be the whole file content if it is a `.mdx` file. For a python file, it would be one of
-    its docstring. In each case, it may contain multiple python code examples. If `skip_cuda_tests` is `True` and a
-    cuda stuff is detective (with a heuristic), this method will return an empty string so no doctest will be run for
-    `string`.
-    """
-    codeblock_pattern = r"(```(?:python|py)\s*\n\s*>>> )((?:.*?\n)*?.*?```)"
-    codeblocks = re.split(re.compile(codeblock_pattern, flags=re.MULTILINE | re.DOTALL), string)
-    is_cuda_found = False
-    for i, codeblock in enumerate(codeblocks):
-        if "load_dataset(" in codeblock and "# doctest: +IGNORE_RESULT" not in codeblock:
-            codeblocks[i] = re.sub(r"(>>> .*load_dataset\(.*)", r"\1 # doctest: +IGNORE_RESULT", codeblock)
-        if (
-            (">>>" in codeblock or "..." in codeblock)
-            and re.search(r"cuda|to\(0\)|device=0", codeblock)
-            and skip_cuda_tests
-        ):
-            is_cuda_found = True
-            break
-    modified_string = ""
-    if not is_cuda_found:
-        modified_string = "".join(codeblocks)
-    return modified_string
-
-
-class HfDocTestParser(doctest.DocTestParser):
-    """
-    Overwrites the DocTestParser from doctest to properly parse the codeblocks that are formatted with black. This
-    means that there are no extra lines at the end of our snippets. The `# doctest: +IGNORE_RESULT` marker is also
-    added anywhere a `load_dataset` call is made as a print would otherwise fail the corresponding line.
-
-    Tests involving cuda are skipped base on a naive pattern that should be updated if it is not enough.
-    """
-
-    # This regular expression is used to find doctest examples in a
-    # string.  It defines three groups: `source` is the source code
-    # (including leading indentation and prompts); `indent` is the
-    # indentation of the first (PS1) line of the source code; and
-    # `want` is the expected output (including leading indentation).
-    # fmt: off
-    _EXAMPLE_RE = re.compile(r'''
-        # Source consists of a PS1 line followed by zero or more PS2 lines.
-        (?P<source>
-            (?:^(?P<indent> [ ]*) >>>    .*)    # PS1 line
-            (?:\n           [ ]*  \.\.\. .*)*)  # PS2 lines
-        \n?
-        # Want consists of any non-blank lines that do not start with PS1.
-        (?P<want> (?:(?![ ]*$)    # Not a blank line
-             (?![ ]*>>>)          # Not a line starting with PS1
-             # !!!!!!!!!!! HF Specific !!!!!!!!!!!
-             (?:(?!```).)*        # Match any character except '`' until a '```' is found (this is specific to HF because black removes the last line)
-             # !!!!!!!!!!! HF Specific !!!!!!!!!!!
-             (?:\n|$)  # Match a new line or end of string
-          )*)
-        ''', re.MULTILINE | re.VERBOSE
-    )
-    # fmt: on
-
-    # !!!!!!!!!!! HF Specific !!!!!!!!!!!
-    skip_cuda_tests: bool = bool(os.environ.get("SKIP_CUDA_DOCTEST", False))
-    # !!!!!!!!!!! HF Specific !!!!!!!!!!!
-
-    def parse(self, string, name="<string>"):
-        """
-        Overwrites the `parse` method to incorporate a skip for CUDA tests, and remove logs and dataset prints before
-        calling `super().parse`
-        """
-        string = preprocess_string(string, self.skip_cuda_tests)
-        return super().parse(string, name)
-
-
-class HfDoctestModule(Module):
-    """
-    Overwrites the `DoctestModule` of the pytest package to make sure the HFDocTestParser is used when discovering
-    tests.
-    """
-
-    def collect(self) -> Iterable[DoctestItem]:
-        class MockAwareDocTestFinder(doctest.DocTestFinder):
-            """A hackish doctest finder that overrides stdlib internals to fix a stdlib bug.
-
-            https://github.com/pytest-dev/pytest/issues/3456 https://bugs.python.org/issue25532
-            """
-
-            def _find_lineno(self, obj, source_lines):
-                """Doctest code does not take into account `@property`, this
-                is a hackish way to fix it. https://bugs.python.org/issue17446
-
-                Wrapped Doctests will need to be unwrapped so the correct line number is returned. This will be
-                reported upstream. #8796
-                """
-                if isinstance(obj, property):
-                    obj = getattr(obj, "fget", obj)
-
-                if hasattr(obj, "__wrapped__"):
-                    # Get the main obj in case of it being wrapped
-                    obj = inspect.unwrap(obj)
-
-                # Type ignored because this is a private function.
-                return super()._find_lineno(  # type:ignore[misc]
-                    obj,
-                    source_lines,
-                )
-
-            def _find(self, tests, obj, name, module, source_lines, globs, seen) -> None:
-                if _is_mocked(obj):
-                    return
-                with _patch_unwrap_mock_aware():
-                    # Type ignored because this is a private function.
-                    super()._find(  # type:ignore[misc]
-                        tests, obj, name, module, source_lines, globs, seen
-                    )
-
-        if self.path.name == "conftest.py":
-            module = self.config.pluginmanager._importconftest(
-                self.path,
-                self.config.getoption("importmode"),
-                rootpath=self.config.rootpath,
-            )
-        else:
-            try:
-                module = import_path(
-                    self.path,
-                    root=self.config.rootpath,
-                    mode=self.config.getoption("importmode"),
-                )
-            except ImportError:
-                if self.config.getvalue("doctest_ignore_import_errors"):
-                    skip("unable to import module %r" % self.path)
-                else:
-                    raise
-
-        # !!!!!!!!!!! HF Specific !!!!!!!!!!!
-        finder = MockAwareDocTestFinder(parser=HfDocTestParser())
-        # !!!!!!!!!!! HF Specific !!!!!!!!!!!
-        optionflags = get_optionflags(self)
-        runner = _get_runner(
-            verbose=False,
-            optionflags=optionflags,
-            checker=_get_checker(),
-            continue_on_failure=_get_continue_on_failure(self.config),
-        )
-        for test in finder.find(module, module.__name__):
-            if test.examples:  # skip empty doctests and cuda
-                yield DoctestItem.from_parent(self, name=test.name, runner=runner, dtest=test)

src/transformers/tools/__init__.py

@@ -38,6 +38,7 @@ else:
     _import_structure["image_captioning"] = ["ImageCaptioningTool"]
     _import_structure["image_question_answering"] = ["ImageQuestionAnsweringTool"]
     _import_structure["image_segmentation"] = ["ImageSegmentationTool"]
+    _import_structure["language_identifier"] = ["LanguageIdentificationTool"]
     _import_structure["speech_to_text"] = ["SpeechToTextTool"]
     _import_structure["text_classification"] = ["TextClassificationTool"]
     _import_structure["text_question_answering"] = ["TextQuestionAnsweringTool"]
@@ -59,6 +60,7 @@ if TYPE_CHECKING:
         from .image_captioning import ImageCaptioningTool
         from .image_question_answering import ImageQuestionAnsweringTool
         from .image_segmentation import ImageSegmentationTool
+        from .language_identifier import LanguageIdentificationTool
         from .speech_to_text import SpeechToTextTool
         from .text_classification import TextClassificationTool
         from .text_question_answering import TextQuestionAnsweringTool

src/transformers/tools/prompts.py

@@ -172,7 +172,7 @@ Human: It worked! Can you translate the summary in German?
 Assistant: I will use the tool `translator` to translate the text in German.
 
 ```py
-translated_summary = translator(summarized_text, src_lang="English", tgt_lang="German")
+translated_summary = translator(summarized_text, src_lang="English", tgt_lang="German)
 ```
 
 ====

src/transformers/utils/__init__.py

@@ -99,7 +99,6 @@ from .import_utils import (
     _LazyModule,
     ccl_version,
     direct_transformers_import,
-    get_torch_version,
     is_accelerate_available,
     is_apex_available,
     is_bitsandbytes_available,
@@ -171,6 +170,7 @@ from .import_utils import (
     is_vision_available,
     requires_backends,
     torch_only_method,
+    torch_version,
 )
 
 

src/transformers/utils/dummy_pt_objects.py

@@ -6440,30 +6440,6 @@ class SqueezeBertPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-SWIFTFORMER_PRETRAINED_MODEL_ARCHIVE_LIST = None
-
-
-class SwiftFormerForImageClassification(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class SwiftFormerModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class SwiftFormerPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
 SWIN_PRETRAINED_MODEL_ARCHIVE_LIST = None