add other models

* Create push-important-models.yml

* feat: add falcon-h1

* fixup

* address comment

* fix

* fix copies

* fix copies

* fix

* fix

* fix

* fix

* fix copies

* fix

* fix copies

* fix test import to at least trigget the cis

* yups

* update

* fix make fix copies

* fix inits?

* fix style

* skip annoying test

* add integration test for Falcon H1

* fix copies

* fix

---------

Co-authored-by: Arthur Zucker <arthur.zucker@gmail.com>
Co-authored-by: dhia.rhaiem <dhia.rhaiem@tii.ae>

.github/workflows/check_failed_model_tests.yml

@@ -39,55 +39,100 @@ jobs:
           name: ci_results_run_models_gpu
           path: /transformers/ci_results_run_models_gpu
 
+      - name: Check file
+        working-directory: /transformers
+        run: |
+          if [ -f ci_results_run_models_gpu/new_model_failures.json ]; then
+            echo "`ci_results_run_models_gpu/new_model_failures.json` exists, continue ..."
+            echo "process=true" >> $GITHUB_ENV
+          else
+            echo "`ci_results_run_models_gpu/new_model_failures.json` doesn't exist, abort."
+            echo "process=false" >> $GITHUB_ENV
+          fi
+
+      - uses: actions/download-artifact@v4
+        if: ${{ env.process == 'true' }}
+        with:
+          pattern: setup_values*
+          path: setup_values
+          merge-multiple: true
+
+      - name: Prepare some setup values
+        if: ${{ env.process == 'true' }}
+        run: |
+          if [ -f setup_values/prev_workflow_run_id.txt ]; then
+            echo "PREV_WORKFLOW_RUN_ID=$(cat setup_values/prev_workflow_run_id.txt)" >> $GITHUB_ENV
+          else
+            echo "PREV_WORKFLOW_RUN_ID=" >> $GITHUB_ENV
+          fi
+
+          if [ -f setup_values/other_workflow_run_id.txt ]; then
+            echo "OTHER_WORKFLOW_RUN_ID=$(cat setup_values/other_workflow_run_id.txt)" >> $GITHUB_ENV
+          else
+            echo "OTHER_WORKFLOW_RUN_ID=" >> $GITHUB_ENV
+          fi
+
       - name: Update clone
         working-directory: /transformers
+        if: ${{ env.process == 'true' }}
         run: git fetch && git checkout ${{ github.sha }}
 
       - name: Get target commit
         working-directory: /transformers/utils
+        if: ${{ env.process == 'true' }}
         run: |
-          echo "END_SHA=$(TOKEN=${{ secrets.ACCESS_REPO_INFO_TOKEN }} python3 -c 'import os; from get_previous_daily_ci import get_last_daily_ci_run_commit; commit=get_last_daily_ci_run_commit(token=os.environ["TOKEN"]); print(commit)')" >> $GITHUB_ENV
+          echo "END_SHA=$(TOKEN=${{ secrets.ACCESS_REPO_INFO_TOKEN }} python3 -c 'import os; from get_previous_daily_ci import get_last_daily_ci_run_commit; commit=get_last_daily_ci_run_commit(token=os.environ["TOKEN"], workflow_run_id=os.environ["PREV_WORKFLOW_RUN_ID"]); print(commit)')" >> $GITHUB_ENV
 
       - name: Checkout to `start_sha`
         working-directory: /transformers
+        if: ${{ env.process == 'true' }}
         run: git fetch && git checkout ${{ inputs.start_sha }}
 
       - name: Reinstall transformers in edit mode (remove the one installed during docker image build)
         working-directory: /transformers
+        if: ${{ env.process == 'true' }}
         run: python3 -m pip uninstall -y transformers && python3 -m pip install -e .
 
       - name: NVIDIA-SMI
+        if: ${{ env.process == 'true' }}
         run: |
           nvidia-smi
 
       - name: Environment
         working-directory: /transformers
+        if: ${{ env.process == 'true' }}
         run: |
           python3 utils/print_env.py
 
       - name: Show installed libraries and their versions
         working-directory: /transformers
+        if: ${{ env.process == 'true' }}
         run: pip freeze
 
       - name: Check failed tests
         working-directory: /transformers
+        if: ${{ env.process == 'true' }}
         run: python3 utils/check_bad_commit.py --start_commit ${{ inputs.start_sha }} --end_commit ${{ env.END_SHA }} --file ci_results_run_models_gpu/new_model_failures.json --output_file new_model_failures_with_bad_commit.json
 
       - name: Show results
         working-directory: /transformers
+        if: ${{ env.process == 'true' }}
         run: |
           ls -l new_model_failures_with_bad_commit.json
           cat new_model_failures_with_bad_commit.json
 
       - name: Checkout back
         working-directory: /transformers
+        if: ${{ env.process == 'true' }}
         run: |
           git checkout ${{ inputs.start_sha }}
 
       - name: Process report
         shell: bash
         working-directory: /transformers
+        if: ${{ env.process == 'true' }}
         env:
+          ACCESS_REPO_INFO_TOKEN: ${{ secrets.ACCESS_REPO_INFO_TOKEN }}
           TRANSFORMERS_CI_RESULTS_UPLOAD_TOKEN: ${{ secrets.TRANSFORMERS_CI_RESULTS_UPLOAD_TOKEN }}
         run: |
           python3 utils/process_bad_commit_report.py
@@ -95,7 +140,9 @@ jobs:
       - name: Process report
         shell: bash
         working-directory: /transformers
+        if: ${{ env.process == 'true' }}
         env:
+          ACCESS_REPO_INFO_TOKEN: ${{ secrets.ACCESS_REPO_INFO_TOKEN }}
           TRANSFORMERS_CI_RESULTS_UPLOAD_TOKEN: ${{ secrets.TRANSFORMERS_CI_RESULTS_UPLOAD_TOKEN }}
         run: |
           {
@@ -105,7 +152,7 @@ jobs:
           } >> "$GITHUB_ENV"
 
       - name: Send processed report
-        if: ${{ !endsWith(env.REPORT_TEXT, '{}') }}
+        if: ${{ env.process == 'true' && !endsWith(env.REPORT_TEXT, '{}') }}
         uses: slackapi/slack-github-action@6c661ce58804a1a20f6dc5fbee7f0381b469e001
         with:
           # Slack channel id, channel name, or user id to post message.

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

@@ -8,8 +8,43 @@ on:
   push:
     branches:
       - run_scheduled_ci*
+  workflow_dispatch:
+    inputs:
+      prev_workflow_run_id:
+        description: 'previous workflow run id to compare'
+        type: string
+        required: false
+        default: ""
+      other_workflow_run_id:
+        description: 'other workflow run id to compare'
+        type: string
+        required: false
+        default: ""
+
+
+# Used for `push` to easily modiffy the target workflow runs to compare against
+env:
+    prev_workflow_run_id: ""
+    other_workflow_run_id: ""
+
 
 jobs:
+  setup:
+    name: Setup
+    runs-on: ubuntu-22.04
+    steps:
+      - name: Setup
+        run: |
+          mkdir "setup_values"
+          echo "${{ inputs.prev_workflow_run_id || env.prev_workflow_run_id }}" > "setup_values/prev_workflow_run_id.txt"
+          echo "${{ inputs.other_workflow_run_id || env.other_workflow_run_id }}" > "setup_values/other_workflow_run_id.txt"
+
+      - name: Upload artifacts
+        uses: actions/upload-artifact@v4
+        with:
+          name: setup_values
+          path: setup_values
+
   model-ci:
     name: Model CI
     uses: ./.github/workflows/self-scheduled.yml

.github/workflows/slack-report.yml

@@ -39,6 +39,21 @@ jobs:
 
       - uses: actions/checkout@v4
       - uses: actions/download-artifact@v4
+
+      - name: Prepare some setup values
+        run: |
+          if [ -f setup_values/prev_workflow_run_id.txt ]; then
+            echo "PREV_WORKFLOW_RUN_ID=$(cat setup_values/prev_workflow_run_id.txt)" >> $GITHUB_ENV
+          else
+            echo "PREV_WORKFLOW_RUN_ID=" >> $GITHUB_ENV
+          fi
+
+          if [ -f setup_values/other_workflow_run_id.txt ]; then
+            echo "OTHER_WORKFLOW_RUN_ID=$(cat setup_values/other_workflow_run_id.txt)" >> $GITHUB_ENV
+          else
+            echo "OTHER_WORKFLOW_RUN_ID=" >> $GITHUB_ENV
+          fi
+
       - name: Send message to Slack
         if: ${{ inputs.job != 'run_quantization_torch_gpu' }}
         env:
@@ -50,7 +65,6 @@ jobs:
           ACCESS_REPO_INFO_TOKEN: ${{ secrets.ACCESS_REPO_INFO_TOKEN }}
           CI_EVENT: ${{ inputs.ci_event }}
           CI_SHA: ${{ github.sha }}
-          CI_WORKFLOW_REF: ${{ github.workflow_ref }}
           CI_TEST_JOB: ${{ inputs.job }}
           SETUP_STATUS: ${{ inputs.setup_status }}
         # We pass `needs.setup.outputs.matrix` as the argument. A processing in `notification_service.py` to change
@@ -58,7 +72,6 @@ jobs:
         # For a job that doesn't depend on (i.e. `needs`) `setup`, the value for `inputs.folder_slices` would be an
         # empty string, and the called script still get one argument (which is the emtpy string).
         run: |
-          sudo apt-get install -y curl
           pip install huggingface_hub
           pip install slack_sdk
           pip show slack_sdk
@@ -86,7 +99,6 @@ jobs:
         # We pass `needs.setup.outputs.quantization_matrix` as the argument. A processing in `notification_service_quantization.py` to change
         # `quantization/bnb` to `quantization_bnb` is required, as the artifact names use `_` instead of `/`.
         run: |
-          sudo apt-get install -y curl
           pip install huggingface_hub
           pip install slack_sdk
           pip show slack_sdk

docs/source/en/_toctree.yml

@@ -455,6 +455,8 @@
         title: Falcon
       - local: model_doc/falcon3
         title: Falcon3
+      - local: model_doc/falcon_h1
+        title: FalconH1
       - local: model_doc/falcon_mamba
         title: FalconMamba
       - local: model_doc/flan-t5

docs/source/en/model_doc/falcon_h1.md

@@ -0,0 +1,65 @@
+<!--Copyright 2025 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.
+⚠️ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
+rendered properly in your Markdown viewer.
+-->
+
+# FalconH1
+
+## Overview
+
+The FalconH1 model was developed by the TII Pretraining team. A comprehensive research paper covering the architecture, pretraining dynamics, experimental results, and conclusions is forthcoming. You can read more about this series in [this website](https://github.com/tiiuae/Falcon-H1).
+
+## Contributors
+
+This model was contributed by [DhiyaEddine](https://huggingface.co/DhiyaEddine), [ybelkada](https://huggingface.co/ybelkada), [JingweiZuo](https://huggingface.co/JingweiZuo), [IlyasChahed](https://huggingface.co/IChahed), and [MaksimVelikanov](https://huggingface.co/yellowvm).
+The original code can be found [here](https://github.com/tiiuae/Falcon-H1).
+
+
+## FalconH1Config
+
+| Model     | Depth | Dim  | Attn Heads | KV | Mamba Heads | d_head       | d_state | Ctx Len        |
+|-----------|--------|------|------------|----|--------------|--------------|------|-----------------|
+| H1 0.5B   | 36     | 1024 | 8          | 2  | 24           | 64 / 64      | 128  | 4K, 16K-SFT     |
+| H1 1.5B   | 24     | 2048 | 8          | 2  | 48           | 128 / 64     | 256  | 128K            |
+| H1 1.5B-d | 66     | 1280 | 6          | 2  | 24           | 128 / 64     | 256  | 128K            |
+| H1 3B     | 32     | 2560 | 10         | 2  | 32           | 128 / 128    | 256  | 128K            |
+| H1 7B     | 44     | 3072 | 12         | 2  | 24           | 128 / 128    | 256  | 256K            |
+| H1 34B    | 72     | 5120 | 20         | 4  | 32           | 128 / 128    | 256  | 256K            |
+
+
+
+[[autodoc]] FalconH1Config
+
+<!---
+## Usage Tips
+Tips: 
+- The architecture is based on Mamba-2 models.
+## FalconH1Model
+[[autodoc]] FalconH1Model
+    - forward
+-->
+
+## FalconH1ForCausalLM
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+model = AutoModelForCausalLM.from_pretrained("tiiuae/Falcon-H1-7B-Instruct")
+tokenizer = AutoTokenizer.from_pretrained("tiiuae/Falcon-H1-7B-Instruct")
+
+message = ["Mamba is a snake with following properties  "]
+inputs = tokenizer(message, return_tensors='pt', return_token_type_ids=False)
+response = model.generate(**inputs, max_new_tokens=64)
+print(tokenizer.batch_decode(response, skip_special_tokens=True)[0])
+```
+
+[[autodoc]] FalconH1ForCausalLM
+    - forward
+
+This HF implementation is contributed by [younesbelkada](https://github.com/younesbelkada) and [DhiaEddineRhaiem](https://github.com/dhiaEddineRhaiem). 

examples/3D_parallel.py

@@ -1,3 +1,16 @@
+# Copyright 2024 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.
 """:
 This script is used to test training a model using Tensor Parallelism and Data Parallelism.
 

examples/flax/question-answering/run_qa.py

@@ -60,7 +60,7 @@ from transformers.utils import check_min_version, send_example_telemetry
 logger = logging.getLogger(__name__)
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 Array = Any
 Dataset = datasets.arrow_dataset.Dataset

examples/flax/speech-recognition/run_flax_speech_recognition_seq2seq.py

@@ -59,7 +59,7 @@ from transformers.utils.versions import require_version
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risk.
-check_min_version("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=2.14.0", "To fix: pip install -r examples/flax/speech-recognition/requirements.txt")
 

examples/flax/text-classification/run_flax_glue.py

@@ -55,7 +55,7 @@ from transformers.utils import check_min_version, send_example_telemetry
 
 logger = logging.getLogger(__name__)
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 Array = Any
 Dataset = datasets.arrow_dataset.Dataset

examples/flax/token-classification/run_flax_ner.py

@@ -56,7 +56,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/token-classification/requirements.txt")
 

examples/pytorch/3d_parallel_checks.py

@@ -1,3 +1,16 @@
+# Copyright 2024 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.
 """:
 This script is used to test training a model using Tensor Parallelism and Data Parallelism.
 

examples/pytorch/audio-classification/run_audio_classification.py

@@ -44,7 +44,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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

@@ -53,7 +53,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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

@@ -56,7 +56,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=2.14.0", "To fix: pip install -r examples/pytorch/image-classification/requirements.txt")
 

examples/pytorch/image-classification/run_image_classification_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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 logger = get_logger(__name__)
 

examples/pytorch/image-pretraining/run_mae.py

@@ -42,7 +42,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/image-pretraining/requirements.txt")
 

examples/pytorch/image-pretraining/run_mim.py

@@ -47,7 +47,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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

@@ -52,7 +52,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/image-pretraining/requirements.txt")
 

examples/pytorch/instance-segmentation/run_instance_segmentation.py

@@ -46,7 +46,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=2.0.0", "To fix: pip install -r examples/pytorch/instance-segmentation/requirements.txt")
 

examples/pytorch/instance-segmentation/run_instance_segmentation_no_trainer.py

@@ -52,7 +52,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=2.0.0", "To fix: pip install -r examples/pytorch/instance-segmentation/requirements.txt")
 

examples/pytorch/language-modeling/run_clm.py

@@ -54,7 +54,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=2.14.0", "To fix: pip install -r examples/pytorch/language-modeling/requirements.txt")
 

examples/pytorch/language-modeling/run_clm_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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 logger = get_logger(__name__)
 

examples/pytorch/language-modeling/run_fim.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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=2.14.0", "To fix: pip install -r examples/pytorch/language-modeling/requirements.txt")
 

examples/pytorch/language-modeling/run_fim_no_trainer.py

@@ -59,7 +59,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=2.14.0", "To fix: pip install -r examples/pytorch/language-modeling/requirements.txt")
 

examples/pytorch/language-modeling/run_mlm_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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 logger = get_logger(__name__)
 require_version("datasets>=2.14.0", "To fix: pip install -r examples/pytorch/language-modeling/requirements.txt")

examples/pytorch/language-modeling/run_plm.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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=2.14.0", "To fix: pip install -r examples/pytorch/language-modeling/requirements.txt")
 

examples/pytorch/multiple-choice/run_swag.py

@@ -45,7 +45,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 logger = logging.getLogger(__name__)
 

examples/pytorch/multiple-choice/run_swag_no_trainer.py

@@ -53,7 +53,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 logger = get_logger(__name__)
 # You should update this to your particular problem to have better documentation of `model_type`

examples/pytorch/object-detection/run_object_detection.py

@@ -48,7 +48,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=2.0.0", "To fix: pip install -r examples/pytorch/object-detection/requirements.txt")
 

examples/pytorch/object-detection/run_object_detection_no_trainer.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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 logging.basicConfig(level=logging.INFO)
 logger = get_logger(__name__)

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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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

@@ -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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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

@@ -54,7 +54,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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

@@ -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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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

@@ -45,7 +45,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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

@@ -50,7 +50,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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

@@ -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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 logger = get_logger(__name__)
 

examples/pytorch/speech-recognition/run_speech_recognition_ctc.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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=1.18.0", "To fix: pip install -r examples/pytorch/speech-recognition/requirements.txt")
 

examples/pytorch/speech-recognition/run_speech_recognition_ctc_adapter.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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=1.18.0", "To fix: pip install -r examples/pytorch/speech-recognition/requirements.txt")
 

examples/pytorch/speech-recognition/run_speech_recognition_seq2seq.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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=1.18.0", "To fix: pip install -r examples/pytorch/speech-recognition/requirements.txt")
 

examples/pytorch/summarization/run_summarization.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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/summarization/requirements.txt")
 

examples/pytorch/summarization/run_summarization_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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 logger = get_logger(__name__)
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/summarization/requirements.txt")

examples/pytorch/text-classification/run_classification.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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/text-classification/requirements.txt")
 

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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 logger = get_logger(__name__)
 

examples/pytorch/text-classification/run_xnli.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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/text-classification/requirements.txt")
 

examples/pytorch/token-classification/run_ner.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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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

@@ -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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/translation/requirements.txt")
 

examples/pytorch/translation/run_translation_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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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

@@ -50,7 +50,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/image-classification/requirements.txt")
 

examples/tensorflow/multiple-choice/run_swag.py

@@ -49,7 +49,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 logger = logging.getLogger(__name__)
 

examples/tensorflow/question-answering/run_qa.py

@@ -61,7 +61,7 @@ except (ModuleNotFoundError, ImportError):
 
 
 # Will error if the minimal version of Transformers is not installed. Remove at your own risks.
-check_min_version("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 logger = logging.getLogger(__name__)
 

examples/tensorflow/summarization/run_summarization.py

@@ -52,7 +52,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/summarization/requirements.txt")
 

examples/tensorflow/text-classification/run_glue.py

@@ -46,7 +46,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 task_to_keys = {
     "cola": ("sentence", None),

examples/tensorflow/translation/run_translation.py

@@ -55,7 +55,7 @@ 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("4.52.0.dev0")
+check_min_version("4.53.0.dev0")
 
 require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/summarization/requirements.txt")
 

setup.py

@@ -451,7 +451,7 @@ install_requires = [
 
 setup(
     name="transformers",
-    version="4.52.0.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="4.53.0.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)
     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__ = "4.52.0.dev0"
+__version__ = "4.53.0.dev0"
 
 from pathlib import Path
 from typing import TYPE_CHECKING

src/transformers/data/datasets/language_modeling.py

@@ -89,8 +89,13 @@ class TextDataset(Dataset):
                 tokenized_text = tokenizer.convert_tokens_to_ids(tokenizer.tokenize(text))
 
                 for i in range(0, len(tokenized_text) - block_size + 1, block_size):  # Truncate in block of block_size
+                    token_block = tokenized_text[i: i + block_size]
                     self.examples.append(
-                        tokenizer.build_inputs_with_special_tokens(tokenized_text[i : i + block_size])
+                        tokenizer.encode(
+                            tokenizer.decode(token_block),
+                            add_special_tokens=True,
+                            truncation=True
+                        )
                     )
                 # Note that we are losing the last truncated example here for the sake of simplicity (no padding)
                 # If your dataset is small, first you should look for a bigger one :-) and second you
@@ -321,7 +326,7 @@ class LineByLineWithSOPTextDataset(Dataset):
                         raise ValueError(f"Length of sequence b is {len(tokens_b)} which must be no less than 1")
 
                     # add special tokens
-                    input_ids = tokenizer.build_inputs_with_special_tokens(tokens_a, tokens_b)
+                    input_ids = tokenizer(tokenizer.decode(tokens_a), tokenizer.decode(tokens_b))['input_ids']
                     # add token type ids, 0 for sentence a, 1 for sentence b
                     token_type_ids = tokenizer.create_token_type_ids_from_sequences(tokens_a, tokens_b)
 
@@ -506,8 +511,7 @@ class TextDatasetForNextSentencePrediction(Dataset):
                         raise ValueError(f"Length of sequence b is {len(tokens_b)} which must be no less than 1")
 
                     # add special tokens
-                    input_ids = self.tokenizer.build_inputs_with_special_tokens(tokens_a, tokens_b)
-                    # add token type ids, 0 for sentence a, 1 for sentence b
+                    input_ids = self.tokenizer(self.tokenizer.decode(tokens_a), self.tokenizer.decode(tokens_b))['input_ids']                    # add token type ids, 0 for sentence a, 1 for sentence b
                     token_type_ids = self.tokenizer.create_token_type_ids_from_sequences(tokens_a, tokens_b)
 
                     example = {

src/transformers/generation/utils.py

@@ -1985,7 +1985,9 @@ class GenerationMixin:
         instantiated, writes it to `model_kwargs`, under the name expected by the model.
         """
 
-        cache_name = "past_key_values" if "mamba" not in self.__class__.__name__.lower() else "cache_params"
+        is_hybrid_cache = any(class_name in self.__class__.__name__.lower() for class_name in ["mamba", "falconh1"])
+        cache_name = "past_key_values" if not is_hybrid_cache else "cache_params"
+
         requires_cross_attention_cache = (
             self.config.is_encoder_decoder or model_kwargs.get("encoder_outputs") is not None
         )

src/transformers/modeling_utils.py

@@ -4177,13 +4177,14 @@ class PreTrainedModel(nn.Module, ModuleUtilsMixin, PushToHubMixin, PeftAdapterMi
 
         # We need to correctly dispatch the model on the current process device. The easiest way for this is to use a simple
         # `device_map` pointing to the correct device
-        if device_mesh is None:
-            tp_plan, device_map, device_mesh = initialize_tensor_parallelism(tp_plan, tp_size=None)
-        else:
-            # TODO: make device_mesh support multiple dimensions
-            if device_mesh.ndim == 1:
-                raise ValueError("device_mesh must be 1 dimensional and will be used for TP")
-            device_map = torch.device(device_mesh.device_type, int(os.environ["LOCAL_RANK"]))
+        if tp_plan is not None:
+            if device_mesh is None and tp_plan is not None:
+                tp_plan, device_map, device_mesh = initialize_tensor_parallelism(tp_plan, tp_size=None)
+            else:
+                # TODO: make device_mesh support multiple dimensions
+                if device_mesh.ndim == 1:
+                    raise ValueError("device_mesh must be 1 dimensional and will be used for TP")
+                device_map = torch.device(device_mesh.device_type, int(os.environ["LOCAL_RANK"]))
 
         if use_auth_token is not None:
             warnings.warn(

src/transformers/models/__init__.py

@@ -103,6 +103,7 @@ if TYPE_CHECKING:
     from .ernie import *
     from .esm import *
     from .falcon import *
+    from .falcon_h1 import *
     from .falcon_mamba import *
     from .fastspeech2_conformer import *
     from .flaubert import *

src/transformers/models/albert/tokenization_albert_fast.py

@@ -134,31 +134,6 @@ class AlbertTokenizerFast(PreTrainedTokenizerFast):
     def can_save_slow_tokenizer(self) -> bool:
         return os.path.isfile(self.vocab_file) if self.vocab_file else False
 
-    def build_inputs_with_special_tokens(
-        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
-    ) -> List[int]:
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. An ALBERT sequence has the following format:
-
-        - single sequence: `[CLS] X [SEP]`
-        - pair of sequences: `[CLS] A [SEP] B [SEP]`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: list of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        sep = [self.sep_token_id]
-        cls = [self.cls_token_id]
-        if token_ids_1 is None:
-            return cls + token_ids_0 + sep
-        return cls + token_ids_0 + sep + token_ids_1 + sep
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/auto/configuration_auto.py

@@ -118,6 +118,7 @@ CONFIG_MAPPING_NAMES = OrderedDict(
         ("ernie_m", "ErnieMConfig"),
         ("esm", "EsmConfig"),
         ("falcon", "FalconConfig"),
+        ("falcon_h1", "FalconH1Config"),
         ("falcon_mamba", "FalconMambaConfig"),
         ("fastspeech2_conformer", "FastSpeech2ConformerConfig"),
         ("flaubert", "FlaubertConfig"),
@@ -481,6 +482,7 @@ MODEL_NAMES_MAPPING = OrderedDict(
         ("esm", "ESM"),
         ("falcon", "Falcon"),
         ("falcon3", "Falcon3"),
+        ("falcon_h1", "FalconH1"),
         ("falcon_mamba", "FalconMamba"),
         ("fastspeech2_conformer", "FastSpeech2Conformer"),
         ("flan-t5", "FLAN-T5"),

src/transformers/models/auto/modeling_auto.py

@@ -115,6 +115,7 @@ MODEL_MAPPING_NAMES = OrderedDict(
         ("ernie_m", "ErnieMModel"),
         ("esm", "EsmModel"),
         ("falcon", "FalconModel"),
+        ("falcon_h1", "FalconH1Model"),
         ("falcon_mamba", "FalconMambaModel"),
         ("fastspeech2_conformer", "FastSpeech2ConformerModel"),
         ("flaubert", "FlaubertModel"),
@@ -558,6 +559,7 @@ MODEL_FOR_CAUSAL_LM_MAPPING_NAMES = OrderedDict(
         ("emu3", "Emu3ForCausalLM"),
         ("ernie", "ErnieForCausalLM"),
         ("falcon", "FalconForCausalLM"),
+        ("falcon_h1", "FalconH1ForCausalLM"),
         ("falcon_mamba", "FalconMambaForCausalLM"),
         ("fuyu", "FuyuForCausalLM"),
         ("gemma", "GemmaForCausalLM"),

src/transformers/models/bart/tokenization_bart_fast.py

@@ -237,13 +237,6 @@ class BartTokenizerFast(PreTrainedTokenizerFast):
         files = self._tokenizer.model.save(save_directory, name=filename_prefix)
         return tuple(files)
 
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        output = [self.bos_token_id] + token_ids_0 + [self.eos_token_id]
-        if token_ids_1 is None:
-            return output
-
-        return output + [self.eos_token_id] + token_ids_1 + [self.eos_token_id]
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/barthez/tokenization_barthez_fast.py

@@ -126,32 +126,6 @@ class BarthezTokenizerFast(PreTrainedTokenizerFast):
     def can_save_slow_tokenizer(self) -> bool:
         return os.path.isfile(self.vocab_file) if self.vocab_file else False
 
-    def build_inputs_with_special_tokens(
-        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
-    ) -> List[int]:
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. A BARThez sequence has the following format:
-
-        - single sequence: `<s> X </s>`
-        - pair of sequences: `<s> A </s></s> B </s>`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-
-        if token_ids_1 is None:
-            return [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
-        cls = [self.cls_token_id]
-        sep = [self.sep_token_id]
-        return cls + token_ids_0 + sep + sep + token_ids_1 + sep
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/bert/tokenization_bert_fast.py

@@ -114,30 +114,6 @@ class BertTokenizerFast(PreTrainedTokenizerFast):
 
         self.do_lower_case = do_lower_case
 
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. A BERT sequence has the following format:
-
-        - single sequence: `[CLS] X [SEP]`
-        - pair of sequences: `[CLS] A [SEP] B [SEP]`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        output = [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
-
-        if token_ids_1 is not None:
-            output += token_ids_1 + [self.sep_token_id]
-
-        return output
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/big_bird/tokenization_big_bird_fast.py

@@ -123,31 +123,6 @@ class BigBirdTokenizerFast(PreTrainedTokenizerFast):
     def can_save_slow_tokenizer(self) -> bool:
         return os.path.isfile(self.vocab_file) if self.vocab_file else False
 
-    def build_inputs_with_special_tokens(
-        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
-    ) -> List[int]:
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. An BigBird sequence has the following format:
-
-        - single sequence: `[CLS] X [SEP]`
-        - pair of sequences: `[CLS] A [SEP] B [SEP]`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: list of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        sep = [self.sep_token_id]
-        cls = [self.cls_token_id]
-        if token_ids_1 is None:
-            return cls + token_ids_0 + sep
-        return cls + token_ids_0 + sep + token_ids_1 + sep
-
     def get_special_tokens_mask(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None, already_has_special_tokens: bool = False
     ) -> List[int]:

src/transformers/models/blenderbot/tokenization_blenderbot_fast.py

@@ -264,21 +264,5 @@ class BlenderbotTokenizerFast(PreTrainedTokenizerFast):
             return len(cls + token_ids_0 + sep) * [0]
         return len(cls + token_ids_0 + sep + sep + token_ids_1 + sep) * [0]
 
-    def build_inputs_with_special_tokens(self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None):
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. A Blenderbot sequence has the following format:
-        - single sequence: ` X </s>`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added
-            token_ids_1 (`List[int]`, *optional*):
-                Will be ignored
-        Returns:
-            `List[int]`: list of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        return token_ids_0 + [self.eos_token_id]
-
 
 __all__ = ["BlenderbotTokenizerFast"]

src/transformers/models/blenderbot_small/tokenization_blenderbot_small_fast.py

@@ -69,13 +69,6 @@ class BlenderbotSmallTokenizerFast(PreTrainedTokenizerFast):
         )
         self.add_prefix_space = add_prefix_space
 
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        output = [self.bos_token_id] + token_ids_0 + [self.eos_token_id]
-        if token_ids_1 is None:
-            return output
-
-        return output + [self.eos_token_id] + token_ids_1 + [self.eos_token_id]
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/camembert/tokenization_camembert_fast.py

@@ -129,32 +129,6 @@ class CamembertTokenizerFast(PreTrainedTokenizerFast):
     def can_save_slow_tokenizer(self) -> bool:
         return os.path.isfile(self.vocab_file) if self.vocab_file else False
 
-    def build_inputs_with_special_tokens(
-        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
-    ) -> List[int]:
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. An CamemBERT sequence has the following format:
-
-        - single sequence: `<s> X </s>`
-        - pair of sequences: `<s> A </s></s> B </s>`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-
-        if token_ids_1 is None:
-            return [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
-        cls = [self.cls_token_id]
-        sep = [self.sep_token_id]
-        return cls + token_ids_0 + sep + sep + token_ids_1 + sep
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/clip/tokenization_clip_fast.py

@@ -106,33 +106,6 @@ class CLIPTokenizerFast(PreTrainedTokenizerFast):
 
         self.backend_tokenizer.decode = new_decode_method
 
-    def build_inputs_with_special_tokens(
-        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
-    ) -> List[int]:
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. A CLIP sequence has the following format:
-
-        - single sequence: `<|startoftext|> X <|endoftext|>`
-
-        Pairs of sequences are not the expected use case, but they will be handled without a separator.
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        bos_token = [self.bos_token_id]
-        eos_token = [self.eos_token_id]
-
-        if token_ids_1 is None:
-            return bos_token + token_ids_0 + eos_token
-        return bos_token + token_ids_0 + eos_token + eos_token + token_ids_1 + eos_token
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/code_llama/tokenization_code_llama_fast.py

@@ -346,33 +346,5 @@ class CodeLlamaTokenizerFast(PreTrainedTokenizerFast):
 
         return (out_vocab_file,)
 
-    def build_inputs_with_special_tokens(
-        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
-    ) -> List[int]:
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. The special tokens depend on calling set_lang.
-
-        An NLLB sequence has the following format, where `X` represents the sequence:
-
-        - `input_ids` (for encoder) `X [eos, src_lang_code]`
-        - `decoder_input_ids`: (for decoder) `X [eos, tgt_lang_code]`
-
-        BOS is never used. Pairs of sequences are not the expected use case, but they will be handled without a
-        separator.
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: list of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        if token_ids_1 is None:
-            return self.bos_token_id + token_ids_0 + self.eos_token_id
-        return self.bos_token_id + token_ids_0 + token_ids_1 + self.eos_token_id
-
 
 __all__ = ["CodeLlamaTokenizerFast"]

src/transformers/models/cohere/tokenization_cohere_fast.py

@@ -496,17 +496,5 @@ class CohereTokenizerFast(PreTrainedTokenizerFast):
             **kwargs,
         )
 
-    # TODO ArthurZ let's rely on the template processor instead, refactor all fast tokenizers
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        bos_token_id = [self.bos_token_id] if self.add_bos_token else []
-        eos_token_id = [self.eos_token_id] if self.add_eos_token else []
-
-        output = bos_token_id + token_ids_0 + eos_token_id
-
-        if token_ids_1 is not None:
-            output = output + bos_token_id + token_ids_1 + eos_token_id
-
-        return output
-
 
 __all__ = ["CohereTokenizerFast"]

src/transformers/models/convbert/tokenization_convbert_fast.py

@@ -115,30 +115,6 @@ class ConvBertTokenizerFast(PreTrainedTokenizerFast):
 
         self.do_lower_case = do_lower_case
 
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. A ConvBERT sequence has the following format:
-
-        - single sequence: `[CLS] X [SEP]`
-        - pair of sequences: `[CLS] A [SEP] B [SEP]`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        output = [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
-
-        if token_ids_1 is not None:
-            output += token_ids_1 + [self.sep_token_id]
-
-        return output
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/cpm/tokenization_cpm_fast.py

@@ -148,32 +148,6 @@ class CpmTokenizerFast(PreTrainedTokenizerFast):
     def can_save_slow_tokenizer(self) -> bool:
         return os.path.isfile(self.vocab_file) if self.vocab_file else False
 
-    # Copied from transformers.models.xlnet.tokenization_xlnet_fast.XLNetTokenizerFast.build_inputs_with_special_tokens
-    def build_inputs_with_special_tokens(
-        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
-    ) -> List[int]:
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. An XLNet sequence has the following format:
-
-        - single sequence: `X <sep> <cls>`
-        - pair of sequences: `A <sep> B <sep> <cls>`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        sep = [self.sep_token_id]
-        cls = [self.cls_token_id]
-        if token_ids_1 is None:
-            return token_ids_0 + sep + cls
-        return token_ids_0 + sep + token_ids_1 + sep + cls
-
     # Copied from transformers.models.xlnet.tokenization_xlnet_fast.XLNetTokenizerFast.create_token_type_ids_from_sequences
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None

src/transformers/models/deberta/tokenization_deberta_fast.py

@@ -154,31 +154,6 @@ class DebertaTokenizerFast(PreTrainedTokenizerFast):
         value = AddedToken(value, lstrip=True, rstrip=False) if isinstance(value, str) else value
         self._mask_token = value
 
-    def build_inputs_with_special_tokens(
-        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
-    ) -> List[int]:
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. A DeBERTa sequence has the following format:
-
-        - single sequence: [CLS] X [SEP]
-        - pair of sequences: [CLS] A [SEP] B [SEP]
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        if token_ids_1 is None:
-            return [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
-        cls = [self.cls_token_id]
-        sep = [self.sep_token_id]
-        return cls + token_ids_0 + sep + token_ids_1 + sep
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/deberta_v2/tokenization_deberta_v2_fast.py

@@ -123,30 +123,6 @@ class DebertaV2TokenizerFast(PreTrainedTokenizerFast):
     def can_save_slow_tokenizer(self) -> bool:
         return os.path.isfile(self.vocab_file) if self.vocab_file else False
 
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. A DeBERTa sequence has the following format:
-
-        - single sequence: [CLS] X [SEP]
-        - pair of sequences: [CLS] A [SEP] B [SEP]
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-
-        if token_ids_1 is None:
-            return [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
-        cls = [self.cls_token_id]
-        sep = [self.sep_token_id]
-        return cls + token_ids_0 + sep + token_ids_1 + sep
-
     def get_special_tokens_mask(self, token_ids_0, token_ids_1=None, already_has_special_tokens=False):
         """
         Retrieves sequence ids from a token list that has no special tokens added. This method is called when adding

src/transformers/models/deprecated/realm/tokenization_realm_fast.py

@@ -191,30 +191,6 @@ class RealmTokenizerFast(PreTrainedTokenizerFast):
 
         return BatchEncoding(output_data, tensor_type=return_tensors)
 
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. A REALM sequence has the following format:
-
-        - single sequence: `[CLS] X [SEP]`
-        - pair of sequences: `[CLS] A [SEP] B [SEP]`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        output = [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
-
-        if token_ids_1 is not None:
-            output += token_ids_1 + [self.sep_token_id]
-
-        return output
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/deprecated/retribert/tokenization_retribert_fast.py

@@ -118,30 +118,6 @@ class RetriBertTokenizerFast(PreTrainedTokenizerFast):
 
         self.do_lower_case = do_lower_case
 
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. A BERT sequence has the following format:
-
-        - single sequence: `[CLS] X [SEP]`
-        - pair of sequences: `[CLS] A [SEP] B [SEP]`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        output = [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
-
-        if token_ids_1 is not None:
-            output += token_ids_1 + [self.sep_token_id]
-
-        return output
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/distilbert/tokenization_distilbert_fast.py

@@ -115,31 +115,6 @@ class DistilBertTokenizerFast(PreTrainedTokenizerFast):
 
         self.do_lower_case = do_lower_case
 
-    # Copied from transformers.models.bert.tokenization_bert_fast.BertTokenizerFast.build_inputs_with_special_tokens
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. A BERT sequence has the following format:
-
-        - single sequence: `[CLS] X [SEP]`
-        - pair of sequences: `[CLS] A [SEP] B [SEP]`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        output = [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
-
-        if token_ids_1 is not None:
-            output += token_ids_1 + [self.sep_token_id]
-
-        return output
-
     # Copied from transformers.models.bert.tokenization_bert_fast.BertTokenizerFast.create_token_type_ids_from_sequences
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None

src/transformers/models/electra/tokenization_electra_fast.py

@@ -111,30 +111,6 @@ class ElectraTokenizerFast(PreTrainedTokenizerFast):
 
         self.do_lower_case = do_lower_case
 
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. A ELECTRA sequence has the following format:
-
-        - single sequence: `[CLS] X [SEP]`
-        - pair of sequences: `[CLS] A [SEP] B [SEP]`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        output = [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
-
-        if token_ids_1 is not None:
-            output += token_ids_1 + [self.sep_token_id]
-
-        return output
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/falcon_h1/__init__.py

@@ -0,0 +1,27 @@
+# Copyright 2025 TII 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.
+from typing import TYPE_CHECKING
+
+from ...utils import _LazyModule
+from ...utils.import_utils import define_import_structure
+
+
+if TYPE_CHECKING:
+    from .configuration_falcon_h1 import *
+    from .modeling_falcon_h1 import *
+else:
+    import sys
+
+    _file = globals()["__file__"]
+    sys.modules[__name__] = _LazyModule(__name__, _file, define_import_structure(_file), module_spec=__spec__)

src/transformers/models/falcon_h1/configuration_falcon_h1.py

@@ -0,0 +1,283 @@
+# coding=utf-8
+# Copyright 2025 TII 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.
+"""FalconH1 model configuration"""
+
+from ...configuration_utils import PretrainedConfig
+from ...utils import logging
+
+
+logger = logging.get_logger(__name__)
+
+
+class FalconH1Config(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`FalconH1Model`]. It is used to instantiate a
+    FalconH1Model model according to the specified arguments, defining the model architecture. Instantiating a configuration
+    with defaults taken from [ibm-fms/FalconH1-9.8b-2.2T-hf](https://huggingface.co/ibm-fms/FalconH1-9.8b-2.2T-hf).
+    The FalconH1Model is a hybrid [mamba2](https://github.com/state-spaces/mamba) architecture with SwiGLU.
+    The checkpoints are  jointly trained by IBM, Princeton, and UIUC.
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+    Args:
+        vocab_size (`int`, *optional*, defaults to 128000):
+            Vocabulary size of the FalconH1 model. Defines the number of different tokens that can be represented by the
+            `inputs_ids` passed when calling [`FalconH1Model`]
+        tie_word_embeddings (`bool`, *optional*, defaults to `False`):
+            Whether the model's input and output word embeddings should be tied. Note that this is only relevant if the
+            model has a output word embedding layer.
+        hidden_size (`int`, *optional*, defaults to 4096):
+            Dimension of the hidden representations.
+        intermediate_size (`int`, *optional*, defaults to 14336):
+            Dimension of the MLP representations.
+        num_hidden_layers (`int`, *optional*, defaults to 32):
+            Number of hidden layers in the Transformer encoder.
+        num_attention_heads (`int`, *optional*, defaults to 32):
+            Number of attention heads for each attention layer in the Transformer encoder.
+        num_key_value_heads (`int`, *optional*, defaults to 8):
+            This is the number of key_value heads that should be used to implement Grouped Query Attention. If
+            `num_key_value_heads=num_attention_heads`, the model will use Multi Head Attention (MHA), if
+            `num_key_value_heads=1` the model will use Multi Query Attention (MQA) otherwise GQA is used. When
+            converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed
+            by meanpooling all the original heads within that group. For more details checkout [this
+            paper](https://arxiv.org/pdf/2305.13245.pdf). If it is not specified, will default to `8`.
+        hidden_act (`str` or `function`, *optional*, defaults to `"silu"`):
+            The non-linear activation function (function or string) in the decoder.
+        initializer_range (`float`, *optional*, defaults to 0.02):
+            The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
+        rms_norm_eps (`float`, *optional*, defaults to 1e-05):
+            The epsilon used by the rms normalization layers.
+        use_cache (`bool`, *optional*, defaults to `True`):
+            Whether or not the model should return the last key/values attentions (not used by all models). Only
+            relevant if `config.is_decoder=True`.
+        num_logits_to_keep (`int` or `None`, *optional*, defaults to 1):
+            Number of prompt logits to calculate during generation. If `None`, all logits will be calculated. If an
+            integer value, only last `num_logits_to_keep` logits will be calculated. Default is 1 because only the
+            logits of the last prompt token are needed for generation. For long sequences, the logits for the entire
+            sequence may use a lot of memory so, setting `num_logits_to_keep=1` will reduce memory footprint
+            significantly.
+        pad_token_id (`int`, *optional*, defaults to 0):
+            The id of the padding token.
+        bos_token_id (`int`, *optional*, defaults to 1):
+            The id of the "beginning-of-sequence" token.
+        eos_token_id (`int`, *optional*, defaults to 2):
+            The id of the "end-of-sequence" token.
+        max_position_embeddings (`int`, *optional*, defaults to 8192):
+            Max cached sequence length for the model
+        attention_dropout (`float`, *optional*, defaults to 0.0):
+            The dropout ratio for the attention probabilities.
+        mamba_d_ssm (`int`, *optional*, defaults to 1024):
+            The dimension of the SSM state space latents.
+        mamba_n_heads (`int`, *optional*, defaults to 128):
+            The number of mamba heads used in the v2 implementation.
+        mamba_d_head (`int`, *optional*, defaults to `"auto"`):
+            Head embeddding dimension size
+        mamba_n_groups (`int`, *optional*, defaults to 1):
+            The number of the mamba groups used in the v2 implementation.
+        mamba_d_state (`int`, *optional*, defaults to 256):
+            The dimension the mamba state space latents
+        mamba_d_conv (`int`, *optional*, defaults to 4):
+            The size of the mamba convolution kernel
+        mamba_expand (`int`, *optional*, defaults to 2):
+            Expanding factor (relative to hidden_size) used to determine the mamba intermediate size
+        mamba_chunk_size (`int`, *optional*, defaults to 256):
+            The chunks in which to break the sequence when doing prefill/training
+        mamba_conv_bias (`bool`, *optional*, defaults to `True`):
+            Flag indicating whether or not to use bias in the convolution layer of the mamba mixer block.
+        mamba_proj_bias (`bool`, *optional*, defaults to `False`):
+            Flag indicating whether or not to use bias in the input and output projections (["in_proj", "out_proj"]) of the mamba mixer block
+        mamba_norm_before_gate (`bool`, *optional*, defaults to `True`):
+            Whether to use RMSNorm before the gate in the Mamba block
+        mamba_rms_norm (`bool`, *optional*, defaults to `False`):
+            Whether to use RMSNorm instead of LayerNorm in the Mamba block
+        projectors_bias (`bool`, *optional*, defaults to `False`):
+            Flag indicating whether or not to use bias in the input and output projections (["in_proj", "out_proj"]) of the attention block
+        rope_theta (`float`, *optional*, defaults to 100000.0):
+            The theta value used for the RoPE embeddings.
+        rope_scaling (`float`, *optional*):
+            The scaling value used for the RoPE embeddings. If `None`, no scaling is applied.
+        lm_head_multiplier (`float`, *optional*, defaults to 1.0):
+            The multiplier for the LM head. This is used to scale the output of the LM head.
+        embedding_multiplier (`float`, *optional*, defaults to 1.0):
+            The multiplier for the embedding layer. This is used to scale the output of the embedding layer.
+        mlp_multipliers (`List[float]`, *optional*):
+            The multipliers for the MLP layers. This is used to scale the output of the MLP layers. The first value is
+            the multiplier of gate layer, the second value is the multiplier of the down_proj layer.
+        key_multiplier (`float`, *optional*):
+            The multiplier for the key layer. This is used to scale the output of the key layer.
+        attention_out_multiplier (`float`, *optional*):
+            The multiplier for the attention output layer. This is used to scale the output of the attention output
+        attention_in_multiplier (`float`, *optional*):
+            The multiplier for the attention input layer. This is used to scale the output of the attention input layer.
+        ssm_multipliers (`List[float]`, *optional*):
+            The multipliers for the SSM layers. This is used to scale the output of the SSM layers.
+        ssm_in_multiplier (`float`, *optional*):
+            The multiplier for the SSM input layer. This is used to scale the output of the SSM input layer.
+        ssm_out_multiplier (`float`, *optional*):
+            The multiplier for the SSM output layer. This is used to scale the output of the SSM output layer.
+    """
+
+    model_type = "falcon_h1"
+    keys_to_ignore_at_inference = ["past_key_values"]
+
+    def __init__(
+        self,
+        vocab_size=128000,
+        tie_word_embeddings=False,
+        hidden_size=4096,
+        intermediate_size=14336,
+        num_hidden_layers=32,
+        num_attention_heads=32,
+        num_key_value_heads=8,
+        hidden_act="silu",
+        initializer_range=0.02,
+        rms_norm_eps=1e-5,
+        use_cache=True,
+        num_logits_to_keep=1,
+        pad_token_id=0,
+        bos_token_id=1,
+        eos_token_id=2,
+        max_position_embeddings=8192,
+        attention_dropout=0.0,
+        mamba_d_ssm=1024,
+        mamba_n_heads=128,
+        mamba_d_head="auto",
+        mamba_n_groups=1,
+        mamba_d_state=256,
+        mamba_d_conv=4,
+        mamba_expand=2,
+        mamba_chunk_size=256,
+        mamba_conv_bias=True,
+        mamba_proj_bias=False,
+        mamba_norm_before_gate=True,
+        mamba_rms_norm=False,
+        projectors_bias=False,
+        rope_theta=100000.0,
+        rope_scaling=None,
+        lm_head_multiplier=1.0,
+        embedding_multiplier=1.0,
+        mlp_multipliers=None,
+        key_multiplier=None,
+        attention_out_multiplier=None,
+        attention_in_multiplier=None,
+        ssm_multipliers=None,
+        ssm_in_multiplier=None,
+        ssm_out_multiplier=None,
+        **kwargs,
+    ):
+        self.vocab_size = vocab_size
+        self.hidden_size = hidden_size
+        self.intermediate_size = intermediate_size
+        self.num_hidden_layers = num_hidden_layers
+        self.num_attention_heads = num_attention_heads
+        self.max_position_embeddings = max_position_embeddings
+        self.attention_dropout = attention_dropout
+        self.attention_bias = False
+        self.mlp_bias = False
+
+        # for backward compatibility
+        if num_key_value_heads is None:
+            num_key_value_heads = num_attention_heads
+
+        self.num_key_value_heads = num_key_value_heads
+        self.hidden_act = hidden_act
+        self.initializer_range = initializer_range
+        self.rms_norm_eps = rms_norm_eps
+
+        self.use_cache = use_cache
+        self.num_logits_to_keep = num_logits_to_keep
+
+        self.rope_theta = rope_theta
+        self.rope_scaling = None
+        self.rope_scaling = rope_scaling
+        self.projectors_bias = projectors_bias
+        mamba_intermediate = mamba_expand * hidden_size if mamba_d_ssm is None else mamba_d_ssm
+
+        if mamba_intermediate % mamba_n_heads != 0:
+            raise ValueError("mamba_n_heads must divide mamba_expand * hidden_size")
+
+        # for the mamba_v2, must satisfy the following
+        if mamba_d_head == "auto":
+            mamba_d_head = mamba_intermediate // mamba_n_heads
+
+        if mamba_d_head * mamba_n_heads != mamba_intermediate:
+            raise ValueError("The dimensions for the Mamba head state do not match the model intermediate_size")
+
+        self.mamba_d_ssm = mamba_d_ssm
+        self.mamba_n_heads = mamba_n_heads
+        self.mamba_d_head = mamba_d_head
+        self.mamba_n_groups = mamba_n_groups
+        self.mamba_d_state = mamba_d_state
+        self.mamba_d_conv = mamba_d_conv
+        self.mamba_expand = mamba_expand
+        self.mamba_chunk_size = mamba_chunk_size
+        self.mamba_conv_bias = mamba_conv_bias
+        self.mamba_proj_bias = mamba_proj_bias
+
+        self.mamba_norm_before_gate = mamba_norm_before_gate
+        self.mamba_rms_norm = mamba_rms_norm
+
+        self.lm_head_multiplier = lm_head_multiplier
+        self.embedding_multiplier = embedding_multiplier
+
+        if mlp_multipliers is not None:
+            self.mlp_multipliers = mlp_multipliers
+        else:
+            self.mlp_multipliers = [1.0, 1.0]
+
+        if attention_out_multiplier is not None:
+            self.attention_out_multiplier = attention_out_multiplier
+        else:
+            self.attention_out_multiplier = 1.0
+
+        if attention_in_multiplier is not None:
+            self.attention_in_multiplier = attention_in_multiplier
+        else:
+            self.attention_in_multiplier = 1.0
+
+        if key_multiplier is not None:
+            self.key_multiplier = key_multiplier
+        else:
+            self.key_multiplier = 1.0
+
+        if ssm_multipliers is not None:
+            self.ssm_multipliers = ssm_multipliers
+        else:
+            #
+            self.ssm_multipliers = [1.0, 1.0, 1.0, 1.0, 1.0]
+
+        if ssm_in_multiplier is not None:
+            self.ssm_in_multiplier = ssm_in_multiplier
+        else:
+            self.ssm_in_multiplier = 1.0
+
+        if ssm_out_multiplier is not None:
+            self.ssm_out_multiplier = ssm_out_multiplier
+        else:
+            self.ssm_out_multiplier = 1.0
+
+        super().__init__(
+            pad_token_id=pad_token_id,
+            bos_token_id=bos_token_id,
+            eos_token_id=eos_token_id,
+            tie_word_embeddings=tie_word_embeddings,
+            **kwargs,
+        )
+
+    @property
+    def layers_block_type(self):
+        return ["attention" for i in range(self.num_hidden_layers)]
+
+
+__all__ = ["FalconH1Config"]

src/transformers/models/falcon_h1/convert_mamba_ssm_checkpoint.py

@@ -0,0 +1,151 @@
+# coding=utf-8
+# Copyright 2025 TII 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.
+"""This script can be used to convert checkpoints provided in the `mamba_ssm` library into the format provided in HuggingFace `transformers`. It depends on the `mamba2_ssm` package to be installed."""
+
+import argparse
+
+import torch
+
+from transformers import AutoModelForCausalLM, AutoTokenizer, FalconH1Config, FalconH1ForCausalLM
+
+
+CONVERSION_MAPPING = {
+    "backbone": "model",
+    "embeddings": "embed_tokens",
+    "mixer.": "",
+    "mixer_ssm": "mamba",
+    "mixer_attn": "self_attn",
+    "mlp.": "feed_forward.",
+    "mlp_norm": "pre_ff_layernorm",
+    "ssm_proj": "mamba.in_proj",
+    "attn_out_proj": "o_proj",
+    ".norm.": ".input_layernorm.",
+    ".mamba.input_layernorm.": ".mamba.norm.",
+    ".ssm_out_proj.": ".mamba.out_proj.",
+    "norm_f": "final_layernorm",
+}
+
+
+def convert_falcon_h1_to_hf(input_model_path, output_path):
+    tokenizer = AutoTokenizer.from_pretrained(input_model_path)
+
+    model = AutoModelForCausalLM.from_pretrained(
+        input_model_path, torch_dtype=torch.bfloat16, trust_remote_code=True, low_cpu_mem_usage=True
+    )
+
+    intermediate_size = int(model.config.expansion_factor * model.config.hidden_size)
+
+    if intermediate_size % 2 != 0:
+        intermediate_size = intermediate_size + (intermediate_size % 2)
+
+    new_config = FalconH1Config(
+        vocab_size=model.config.vocab_size,
+        tie_word_embeddings=model.config.tie_word_embeddings,
+        hidden_size=model.config.hidden_size,
+        intermediate_size=intermediate_size,
+        mamba_d_state=model.config.state_size,
+        num_hidden_layers=model.config.num_hidden_layers,
+        mamba_use_mlp=model.config.use_mlp,
+        rms_norm_eps=model.config.layer_norm_epsilon,
+        pad_token_id=model.config.pad_token_id,
+        eos_token_id=model.config.eos_token_id,
+        mamba_expand=model.config.expand,
+        mamba_d_conv=model.config.conv_kernel,
+        mamba_n_groups=model.config.n_groups,
+        mamba_n_heads=model.config.num_heads,
+        mamba_norm_before_gate=model.config.norm_before_gate,
+        mamba_rms_norm=model.config.rms_norm,
+        mamba_d_ssm=model.config.d_ssm,
+        attention_bias=model.config.use_bias,
+        projectors_bias=model.config.use_bias,
+        mamba_conv_bias=model.config.use_conv_bias,
+        hidden_act=model.config.hidden_act,
+        use_cache=model.config.use_cache,
+        mamba_chunk_size=model.config.chunk_size,
+        num_attention_heads=model.config.num_heads_mha,
+        num_key_value_heads=model.config.num_key_value_heads,
+        head_dim=model.config.head_dim_mha,
+        lm_head_multiplier=model.config.lm_head_multiplier,
+        embedding_multiplier=model.config.embedding_multiplier,
+        mlp_multipliers=model.config.mlp_multipliers,
+        key_multiplier=model.config.key_multiplier,
+        attention_out_multiplier=model.config.attention_out_multiplier,
+        attention_in_multiplier=model.config.attention_in_multiplier,
+        ssm_multipliers=model.config.ssm_multipliers,
+        ssm_in_multiplier=model.config.ssm_in_multiplier,
+        ssm_out_multiplier=model.config.ssm_out_multiplier,
+        rope_theta=model.config.rope_theta,
+    )
+
+    old_state_dict = model.state_dict()
+    new_state_dict = {}
+
+    for old_key, old_value in old_state_dict.items():
+        new_key = old_key
+        for conversion_key, conversion_value in CONVERSION_MAPPING.items():
+            if conversion_key in old_key:
+                new_key = new_key.replace(conversion_key, conversion_value)
+
+        if "mamba.input_layernorm" in new_key:
+            new_key = new_key.replace("mamba.input_layernorm", "mamba.norm")
+
+        # Special processing for attention layers
+        if "self_attn.attn_proj" in new_key:
+            num_heads = new_config.num_attention_heads
+            num_kv_heads = new_config.num_key_value_heads
+            head_dim = new_config.head_dim
+            q_proj, k_proj, v_proj = old_value.split(
+                [
+                    num_heads * head_dim,
+                    num_kv_heads * head_dim,
+                    num_kv_heads * head_dim,
+                ],
+                dim=0,
+            )
+            new_state_dict[new_key.replace("attn_proj", "q_proj")] = q_proj
+            new_state_dict[new_key.replace("attn_proj", "k_proj")] = k_proj
+            new_state_dict[new_key.replace("attn_proj", "v_proj")] = v_proj
+        else:
+            new_state_dict[new_key] = old_value
+
+    with torch.device("meta"):
+        new_model = FalconH1ForCausalLM(new_config)
+
+    del model
+
+    new_model.load_state_dict(new_state_dict, strict=True, assign=True)
+
+    new_model.save_pretrained(output_path)
+    tokenizer.save_pretrained(output_path)
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "-i",
+        "--mamba_ssm_checkpoint_directory",
+        type=str,
+        required=True,
+        help="Path to a directory containing the `pytorch_model.bin` mamba_ssm checkpoint file to be converted.",
+    )
+    parser.add_argument(
+        "-o", "--output_dir", type=str, required=True, help="Path to directory to save the converted output model to."
+    )
+    args = parser.parse_args()
+
+    convert_falcon_h1_to_hf(
+        args.mamba_ssm_checkpoint_directory,
+        args.output_dir,
+    )

src/transformers/models/fnet/tokenization_fnet_fast.py

@@ -117,31 +117,6 @@ class FNetTokenizerFast(PreTrainedTokenizerFast):
     def can_save_slow_tokenizer(self) -> bool:
         return os.path.isfile(self.vocab_file) if self.vocab_file else False
 
-    def build_inputs_with_special_tokens(
-        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
-    ) -> List[int]:
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. An FNet sequence has the following format:
-
-        - single sequence: `[CLS] X [SEP]`
-        - pair of sequences: `[CLS] A [SEP] B [SEP]`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: list of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        sep = [self.sep_token_id]
-        cls = [self.cls_token_id]
-        if token_ids_1 is None:
-            return cls + token_ids_0 + sep
-        return cls + token_ids_0 + sep + token_ids_1 + sep
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/funnel/tokenization_funnel_fast.py

@@ -140,31 +140,6 @@ class FunnelTokenizerFast(PreTrainedTokenizerFast):
 
         self.do_lower_case = do_lower_case
 
-    # Copied from transformers.models.bert.tokenization_bert_fast.BertTokenizerFast.build_inputs_with_special_tokens with BERT->Funnel
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. A Funnel sequence has the following format:
-
-        - single sequence: `[CLS] X [SEP]`
-        - pair of sequences: `[CLS] A [SEP] B [SEP]`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        output = [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
-
-        if token_ids_1 is not None:
-            output += token_ids_1 + [self.sep_token_id]
-
-        return output
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/gemma/tokenization_gemma_fast.py

@@ -183,17 +183,5 @@ class GemmaTokenizerFast(PreTrainedTokenizerFast):
 
         return (out_vocab_file,)
 
-    # Copied from transformers.models.llama.tokenization_llama_fast.LlamaTokenizerFast.build_inputs_with_special_tokens
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        bos_token_id = [self.bos_token_id] if self.add_bos_token else []
-        eos_token_id = [self.eos_token_id] if self.add_eos_token else []
-
-        output = bos_token_id + token_ids_0 + eos_token_id
-
-        if token_ids_1 is not None:
-            output = output + bos_token_id + token_ids_1 + eos_token_id
-
-        return output
-
 
 __all__ = ["GemmaTokenizerFast"]

src/transformers/models/gpt_neox/tokenization_gpt_neox_fast.py

@@ -204,18 +204,6 @@ class GPTNeoXTokenizerFast(PreTrainedTokenizerFast):
             + eos_token_id
         )
 
-    # Copied from transformers.models.llama.tokenization_llama_fast.LlamaTokenizerFast.build_inputs_with_special_tokens
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        bos_token_id = [self.bos_token_id] if self.add_bos_token else []
-        eos_token_id = [self.eos_token_id] if self.add_eos_token else []
-
-        output = bos_token_id + token_ids_0 + eos_token_id
-
-        if token_ids_1 is not None:
-            output = output + bos_token_id + token_ids_1 + eos_token_id
-
-        return output
-
     def save_vocabulary(self, save_directory: str, filename_prefix: Optional[str] = None) -> Tuple[str]:
         files = self._tokenizer.model.save(save_directory, name=filename_prefix)
         return tuple(files)

src/transformers/models/herbert/tokenization_herbert_fast.py

@@ -71,33 +71,6 @@ class HerbertTokenizerFast(PreTrainedTokenizerFast):
             **kwargs,
         )
 
-    def build_inputs_with_special_tokens(
-        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
-    ) -> List[int]:
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. An HerBERT, like BERT sequence has the following format:
-
-        - single sequence: `<s> X </s>`
-        - pair of sequences: `<s> A </s> B </s>`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-
-        cls = [self.cls_token_id]
-        sep = [self.sep_token_id]
-        if token_ids_1 is None:
-            return cls + token_ids_0 + sep
-
-        return cls + token_ids_0 + sep + token_ids_1 + sep
-
     def get_special_tokens_mask(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None, already_has_special_tokens: bool = False
     ) -> List[int]:

src/transformers/models/layoutlm/tokenization_layoutlm_fast.py

@@ -115,30 +115,6 @@ class LayoutLMTokenizerFast(PreTrainedTokenizerFast):
 
         self.do_lower_case = do_lower_case
 
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. A LayoutLM sequence has the following format:
-
-        - single sequence: `[CLS] X [SEP]`
-        - pair of sequences: `[CLS] A [SEP] B [SEP]`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        output = [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
-
-        if token_ids_1 is not None:
-            output += token_ids_1 + [self.sep_token_id]
-
-        return output
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/layoutlmv2/tokenization_layoutlmv2_fast.py

@@ -757,30 +757,6 @@ class LayoutLMv2TokenizerFast(PreTrainedTokenizerFast):
 
         return encoded_inputs
 
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. A BERT sequence has the following format:
-
-        - single sequence: `[CLS] X [SEP]`
-        - pair of sequences: `[CLS] A [SEP] B [SEP]`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-        output = [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
-
-        if token_ids_1:
-            output += token_ids_1 + [self.sep_token_id]
-
-        return output
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/layoutlmv3/tokenization_layoutlmv3_fast.py

@@ -816,13 +816,6 @@ class LayoutLMv3TokenizerFast(PreTrainedTokenizerFast):
         files = self._tokenizer.model.save(save_directory, name=filename_prefix)
         return tuple(files)
 
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        output = [self.bos_token_id] + token_ids_0 + [self.eos_token_id]
-        if token_ids_1 is None:
-            return output
-
-        return output + [self.eos_token_id] + token_ids_1 + [self.eos_token_id]
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/layoutxlm/tokenization_layoutxlm_fast.py

@@ -745,32 +745,6 @@ class LayoutXLMTokenizerFast(PreTrainedTokenizerFast):
 
         return encoded_inputs
 
-    def build_inputs_with_special_tokens(
-        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
-    ) -> List[int]:
-        """
-        Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
-        adding special tokens. An XLM-RoBERTa sequence has the following format:
-
-        - single sequence: `<s> X </s>`
-        - pair of sequences: `<s> A </s></s> B </s>`
-
-        Args:
-            token_ids_0 (`List[int]`):
-                List of IDs to which the special tokens will be added.
-            token_ids_1 (`List[int]`, *optional*):
-                Optional second list of IDs for sequence pairs.
-
-        Returns:
-            `List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
-        """
-
-        if token_ids_1 is None:
-            return [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
-        cls = [self.cls_token_id]
-        sep = [self.sep_token_id]
-        return cls + token_ids_0 + sep + sep + token_ids_1 + sep
-
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
     ) -> List[int]:

src/transformers/models/led/tokenization_led_fast.py

@@ -241,14 +241,6 @@ class LEDTokenizerFast(PreTrainedTokenizerFast):
         files = self._tokenizer.model.save(save_directory, name=filename_prefix)
         return tuple(files)
 
-    # Copied from transformers.models.bart.tokenization_bart_fast.BartTokenizerFast.build_inputs_with_special_tokens
-    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
-        output = [self.bos_token_id] + token_ids_0 + [self.eos_token_id]
-        if token_ids_1 is None:
-            return output
-
-        return output + [self.eos_token_id] + token_ids_1 + [self.eos_token_id]
-
     # Copied from transformers.models.bart.tokenization_bart_fast.BartTokenizerFast.create_token_type_ids_from_sequences with BART->LED
     def create_token_type_ids_from_sequences(
         self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None