Release: v4.18.0

This reverts commit b1a7dfe099.

.circleci/config.yml

@@ -81,7 +81,7 @@ jobs:
             - run: sudo apt-get -y update && sudo apt-get install -y libsndfile1-dev espeak-ng
             - run: pip install --upgrade pip
             - run: pip install .[sklearn,tf-cpu,torch,testing,sentencepiece,torch-speech,vision]
-            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.10.0+cpu.html
+            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.11.0+cpu.html
             - run: pip install tensorflow_probability
             - run: pip install https://github.com/kpu/kenlm/archive/master.zip
             - save_cache:
@@ -119,7 +119,7 @@ jobs:
             - run: sudo apt-get -y update && sudo apt-get install -y libsndfile1-dev espeak-ng
             - run: pip install --upgrade pip
             - run: pip install .[sklearn,tf-cpu,torch,testing,sentencepiece,torch-speech,vision]
-            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.10.0+cpu.html
+            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.11.0+cpu.html
             - run: pip install tensorflow_probability
             - run: pip install https://github.com/kpu/kenlm/archive/master.zip
             - save_cache:
@@ -152,7 +152,7 @@ jobs:
             - run: sudo apt-get -y update && sudo apt-get install -y libsndfile1-dev espeak-ng
             - run: pip install --upgrade pip
             - run: pip install .[sklearn,flax,torch,testing,sentencepiece,torch-speech,vision]
-            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.10.0+cpu.html
+            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.11.0+cpu.html
             - run: pip install https://github.com/kpu/kenlm/archive/master.zip
             - save_cache:
                 key: v0.4-{{ checksum "setup.py" }}
@@ -189,7 +189,7 @@ jobs:
             - run: sudo apt-get -y update && sudo apt-get install -y libsndfile1-dev espeak-ng
             - run: pip install --upgrade pip
             - run: pip install .[sklearn,flax,torch,testing,sentencepiece,torch-speech,vision]
-            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.10.0+cpu.html
+            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.11.0+cpu.html
             - run: pip install https://github.com/kpu/kenlm/archive/master.zip
             - save_cache:
                 key: v0.4-{{ checksum "setup.py" }}
@@ -220,7 +220,7 @@ jobs:
             - run: sudo apt-get -y update && sudo apt-get install -y libsndfile1-dev espeak-ng
             - run: pip install --upgrade pip
             - run: pip install .[sklearn,torch,testing,sentencepiece,torch-speech,vision,timm]
-            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.10.0+cpu.html
+            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.11.0+cpu.html
             - run: pip install https://github.com/kpu/kenlm/archive/master.zip
             - save_cache:
                   key: v0.4-torch-{{ checksum "setup.py" }}
@@ -256,7 +256,7 @@ jobs:
             - run: sudo apt-get -y update && sudo apt-get install -y libsndfile1-dev espeak-ng
             - run: pip install --upgrade pip
             - run: pip install .[sklearn,torch,testing,sentencepiece,torch-speech,vision,timm]
-            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.10.0+cpu.html
+            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.11.0+cpu.html
             - run: pip install https://github.com/kpu/kenlm/archive/master.zip
             - save_cache:
                   key: v0.4-torch-{{ checksum "setup.py" }}
@@ -420,7 +420,7 @@ jobs:
             - run: sudo apt-get -y update && sudo apt-get install -y libsndfile1-dev espeak-ng
             - run: pip install --upgrade pip
             - run: pip install .[sklearn,torch,testing,sentencepiece,torch-speech,vision,timm]
-            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.10.0+cpu.html
+            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.11.0+cpu.html
             - run: pip install https://github.com/kpu/kenlm/archive/master.zip
             - save_cache:
                   key: v0.4-torch-{{ checksum "setup.py" }}
@@ -457,7 +457,7 @@ jobs:
             - run: sudo apt-get -y update && sudo apt-get install -y libsndfile1-dev espeak-ng
             - run: pip install --upgrade pip
             - run: pip install .[sklearn,torch,testing,sentencepiece,torch-speech,vision,timm]
-            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.10.0+cpu.html
+            - run: pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.11.0+cpu.html
             - run: pip install https://github.com/kpu/kenlm/archive/master.zip
             - save_cache:
                   key: v0.4-torch-{{ checksum "setup.py" }}
@@ -559,6 +559,10 @@ jobs:
                   if [ -f test_list.txt ]; then
                     python -m pytest -s --make-reports=tests_custom_tokenizers ./tests/test_tokenization_bert_japanese.py ./tests/test_tokenization_openai.py | tee tests_output.txt
                   fi
+            - run: |
+                  if [ -f test_list.txt ]; then
+                    python -m pytest -n 1 tests/test_tokenization_clip.py --dist=loadfile -s --make-reports=tests_tokenization_clip --durations=100 | tee tests_output.txt
+                  fi
             - store_artifacts:
                   path: ~/transformers/tests_output.txt
             - store_artifacts:
@@ -779,7 +783,7 @@ jobs:
                       - v0.4-torch-{{ checksum "setup.py" }}
                       - v0.4-{{ checksum "setup.py" }}
             - run: pip install --upgrade pip
-            - run: pip install .[torch,testing,sentencepiece,onnxruntime]
+            - run: pip install .[torch,testing,sentencepiece,onnxruntime,vision]
             - save_cache:
                   key: v0.4-onnx-{{ checksum "setup.py" }}
                   paths:
@@ -812,7 +816,7 @@ jobs:
                       - v0.4-torch-{{ checksum "setup.py" }}
                       - v0.4-{{ checksum "setup.py" }}
             - run: pip install --upgrade pip
-            - run: pip install .[torch,testing,sentencepiece,onnxruntime]
+            - run: pip install .[torch,testing,sentencepiece,onnxruntime,vision]
             - save_cache:
                   key: v0.4-onnx-{{ checksum "setup.py" }}
                   paths:
@@ -848,7 +852,7 @@ jobs:
             - run: isort --check-only examples tests src utils
             - run: python utils/custom_init_isort.py --check_only
             - run: flake8 examples tests src utils
-            - run: python utils/style_doc.py src/transformers docs/source --max_len 119 --check_only
+            - run: doc-builder style src/transformers docs/source --max_len 119 --check_only --path_to_docs docs/source
 
     check_repository_consistency:
         working_directory: ~/transformers
@@ -951,7 +955,7 @@ workflow_filters: &workflow_filters
     filters:
         branches:
             only:
-                - master
+                - main
 workflows:
     version: 2
     build_and_test:
@@ -978,7 +982,7 @@ workflows:
                 filters:
                     branches:
                         only:
-                            - master
+                            - main
         jobs:
             - run_examples_torch_all
             - run_examples_flax_all
@@ -1000,7 +1004,7 @@ workflows:
 #                filters:
 #                    branches:
 #                        only:
-#                            - master
+#                            - main
 #        jobs:
 #            - cleanup-gke-jobs
 #            - run_examples_tpu

.github/ISSUE_TEMPLATE/bug-report.md

@@ -28,12 +28,13 @@ assignees: ''
 Models:
 
 - ALBERT, BERT, XLM, DeBERTa, DeBERTa-v2, ELECTRA, MobileBert, SqueezeBert: @LysandreJik
-- T5, BART, Marian, Pegasus, EncoderDecoder: @patrickvonplaten
-- Blenderbot, MBART: @patil-suraj
-- Longformer, Reformer, TransfoXL, XLNet, FNet, BigBird: @patrickvonplaten
+- T5, Pegasus, EncoderDecoder: @patrickvonplaten
+- Blenderbot, MBART, BART, Marian, Pegasus: @patil-suraj
+- Reformer, TransfoXL, XLNet, FNet: @patrickvonplaten
+- Longformer, BigBird: @ydshieh
 - FSMT: @stas00
 - Funnel: @sgugger
-- GPT-2, GPT: @patrickvonplaten, @LysandreJik
+- GPT-2, GPT: @patil-suraj, @patrickvonplaten, @LysandreJik
 - RAG, DPR: @patrickvonplaten, @lhoestq
 - TensorFlow: @Rocketknight1
 - JAX/Flax: @patil-suraj

.github/ISSUE_TEMPLATE/feature-request.md

@@ -22,4 +22,4 @@ assignees: ''
 
 <!-- Is there any way that you could help, e.g. by submitting a PR?
      Make sure to read the CONTRIBUTING.MD readme:
-     https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md -->
+     https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md -->

.github/PULL_REQUEST_TEMPLATE.md

@@ -17,13 +17,13 @@ Fixes # (issue)
 
 ## Before submitting
 - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
-- [ ] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md#start-contributing-pull-requests),
+- [ ] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests),
       Pull Request section?
 - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link
       to it if that's the case.
 - [ ] Did you make sure to update the documentation with your changes? Here are the
-      [documentation guidelines](https://github.com/huggingface/transformers/tree/master/docs), and
-      [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/master/docs#writing-source-documentation).
+      [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and
+      [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation).
 - [ ] Did you write any new necessary tests?
 
 

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

@@ -3,7 +3,7 @@ name: Add model like runner
 on:
   push:
     branches:
-      - master
+      - main
   pull_request:
     paths:
       - "src/**"
@@ -12,7 +12,8 @@ on:
     types: [opened, synchronize, reopened]
 
 jobs:
-  run_tests_templates:
+  run_tests_templates_like:
+    name: "Add new model like template tests"
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v2
@@ -30,6 +31,7 @@ jobs:
       - name: Install dependencies
         run: |
           pip install --upgrade pip!=21.3
+          pip install -U click  # Click 7 is installed in the environment by default, but we need at least version 8 for Black
           sudo apt -y update && sudo apt install -y libsndfile1-dev
           pip install .[dev]
 
@@ -41,7 +43,7 @@ jobs:
 
       - name: Run all PyTorch modeling test
         run: |
-          python -m pytest -n 2 --dist=loadfile -s --make-reports=tests_new_models tests/test_modeling_bert_new.py
+          python -m pytest -n 2 --dist=loadfile -s --make-reports=tests_new_models tests/bert_new/test_modeling_bert_new.py
 
       - name: Run style changes
         run: |
@@ -49,11 +51,11 @@ jobs:
 
       - name: Failure short reports
         if: ${{ always() }}
-        run: cat reports/tests_new_models_failures_short.txt
+        run: cat reports/tests_new_models/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
           name: run_all_tests_new_models_test_reports
-          path: reports
+          path: reports/tests_new_models

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

@@ -0,0 +1,145 @@
+name: Build docker images (scheduled)
+
+on:
+  push:
+    branches:
+      - docker-image*
+  repository_dispatch:
+  schedule:
+    - cron: "0 1 * * *"
+
+concurrency:
+  group: docker-images-builds
+  cancel-in-progress: false
+
+jobs:
+  latest-docker:
+    name: "Latest PyTorch + TensorFlow [dev]"
+    runs-on: ubuntu-latest
+    steps:
+      -
+        name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v1
+      -
+        name: Check out code
+        uses: actions/checkout@v2
+      -
+        name: Login to DockerHub
+        uses: docker/login-action@v1
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_PASSWORD }}
+      -
+        name: Build and push
+        uses: docker/build-push-action@v2
+        with:
+          context: ./docker/transformers-all-latest-gpu
+          build-args: |
+            REF=main
+          push: true
+          tags: huggingface/transformers-all-latest-gpu
+
+  latest-torch-deepspeed-docker:
+    name: "Latest PyTorch + DeepSpeed"
+    needs: latest-docker
+    runs-on: ubuntu-latest
+    steps:
+      -
+        name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v1
+      -
+        name: Check out code
+        uses: actions/checkout@v2
+      -
+        name: Login to DockerHub
+        uses: docker/login-action@v1
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_PASSWORD }}
+      -
+        name: Build and push
+        uses: docker/build-push-action@v2
+        with:
+          context: ./docker/transformers-pytorch-deepspeed-latest-gpu
+          build-args: |
+            REF=main
+          push: true
+          tags: huggingface/transformers-pytorch-deepspeed-latest-gpu
+
+  doc-builder:
+    name: "Doc builder"
+    runs-on: ubuntu-latest
+    steps:
+      -
+        name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v1
+      -
+        name: Check out code
+        uses: actions/checkout@v2
+      -
+        name: Login to DockerHub
+        uses: docker/login-action@v1
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_PASSWORD }}
+      -
+        name: Build and push
+        uses: docker/build-push-action@v2
+        with:
+          context: ./docker/transformers-doc-builder
+          push: true
+          tags: huggingface/transformers-doc-builder
+
+  latest-pytorch:
+    name: "Latest PyTorch [dev]"
+    runs-on: ubuntu-latest
+    needs: latest-torch-deepspeed-docker
+    steps:
+      -
+        name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v1
+      -
+        name: Check out code
+        uses: actions/checkout@v2
+      -
+        name: Login to DockerHub
+        uses: docker/login-action@v1
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_PASSWORD }}
+      -
+        name: Build and push
+        uses: docker/build-push-action@v2
+        with:
+          context: ./docker/transformers-pytorch-gpu
+          build-args: |
+            REF=main
+          push: true
+          tags: huggingface/transformers-pytorch-gpu
+
+  latest-tensorflow:
+    needs: latest-pytorch
+    name: "Latest TensorFlow [dev]"
+    runs-on: ubuntu-latest
+    steps:
+      -
+        name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v1
+      -
+        name: Check out code
+        uses: actions/checkout@v2
+      -
+        name: Login to DockerHub
+        uses: docker/login-action@v1
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_PASSWORD }}
+      -
+        name: Build and push
+        uses: docker/build-push-action@v2
+        with:
+          context: ./docker/transformers-tensorflow-gpu
+          build-args: |
+            REF=main
+          push: true
+          tags: huggingface/transformers-tensorflow-gpu

.github/workflows/build_dev_documentation.yml

@@ -1,120 +0,0 @@
-name: Build dev documentation
-
-on:
-  pull_request:
-
-
-jobs:
-  build_and_package:
-    runs-on: [self-hosted, doc-builder]
-    container:
-      image: huggingface/doc-builder-transformers
-      options: "-v /home/github_actions:/mnt"
-    env:
-      PR_NUMBER: ${{ github.event.number }}
-      EVENT_CONTEXT: ${{ toJSON(github.event) }}
-
-    steps:
-      - uses: actions/checkout@v2
-        with:
-          repository: 'huggingface/doc-builder'
-          path: doc-builder
-
-      - uses: actions/checkout@v2
-        with:
-          repository: 'huggingface/transformers'
-          path: transformers
-
-      - uses: actions/checkout@v2
-        with:
-          repository: 'huggingface/notebooks'
-          path: notebooks
-
-      - name: Set env
-        run: echo "WRITE=$(cat /mnt/WRITE)" >> $GITHUB_ENV
-
-      - name: Comment PR
-        uses: thollander/actions-comment-pull-request@v1
-        if: github.event.action == 'opened'
-
-        with:
-          message: 'The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/transformers/pr_${{ env.PR_NUMBER }}). All of your documentation changes will be reflected on that endpoint.'
-          GITHUB_TOKEN: ${{ env.WRITE }}
-
-      - name: Find Comment
-        if: github.event.action == 'reopened'
-        uses: peter-evans/find-comment@v1
-        id: fc
-        with:
-          issue-number: ${{ env.PR_NUMBER }}
-          comment-author: HuggingFaceDocBuilder
-
-      - name: Update comment
-        if: github.event.action == 'reopened'
-        uses: peter-evans/create-or-update-comment@v1
-        with:
-          comment-id: ${{ steps.fc.outputs.comment-id }}
-          token: ${{ env.WRITE }}
-          edit-mode: replace
-          body: |
-            The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/transformers/pr_${{ env.PR_NUMBER }}). All of your documentation changes will be reflected on that endpoint.
-
-      - name: Loading cache.
-        uses: actions/cache@v2
-        id: cache
-        with:
-          path: ~/.cache/pip
-          key: v1-test_build_doc
-          restore-keys: |
-            v1-test_build_doc-${{ hashFiles('setup.py') }}
-            v1-test_build_doc
-
-      - name: Setup environment
-        run: |
-          apt-get -y update && apt-get install -y libsndfile1-dev
-
-          pip uninstall -y doc-builder
-          pip install git+https://github.com/huggingface/doc-builder -U
-          cd transformers
-          pip install .[dev]
-          cd ..
-
-          export TORCH_VERSION=$(python -c "from torch import version; print(version.__version__.split('+')[0])")
-          pip install torch-scatter -f https://data.pyg.org/whl/torch-${TORCH_VERSION}+cpu.html
-
-          pip install torchvision
-          python -m pip install 'git+https://github.com/facebookresearch/detectron2.git'
-
-          apt install -y tesseract-ocr
-          pip install pytesseract
-          pip install pytorch-quantization --extra-index-url https://pypi.ngc.nvidia.com
-          pip install https://github.com/kpu/kenlm/archive/master.zip
-
-      - name: Setup git
-        run: |
-          git config --global user.name "Hugging Face Doc Builder"
-          git config --global user.email docs@huggingface.co
-          
-          cd doc-builder
-          git pull origin main
-          cd ..
-
-          cd notebooks
-          git pull origin master
-          cd ..
-          WRITE=`cat /mnt/WRITE`
-          
-          rm -rf doc-build-dev
-          git clone https://HuggingFaceDocBuilder:$WRITE@github.com/huggingface/doc-build-dev
-
-      - name: Make documentation
-        run: |
-          doc-builder build transformers transformers/docs/source --build_dir doc-build-dev --notebook_dir notebooks/transformers_doc --clean --version pr_$PR_NUMBER
-
-      - name: Push to repositories
-        run: |
-          cd doc-build-dev
-          ls
-          git add .
-          git commit -m "Updated with commit ${{ github.sha }} See: https://github.com/huggingface/transformers/commit/${{ github.sha }}"
-          git push origin main

.github/workflows/build_doc_test.yml

@@ -1,50 +0,0 @@
-name: Documentation test build
-
-on:
-  pull_request:
-    paths:
-      - "src/**"
-      - "docs/**"
-      - ".github/**"
-
-jobs:
-  build_and_package:
-    runs-on: ubuntu-latest
-    defaults:
-      run:
-        shell: bash -l {0}
-
-    steps:
-      - uses: actions/checkout@v2
-      
-      - name: Loading cache.
-        uses: actions/cache@v2
-        id: cache
-        with:
-          path: ~/.cache/pip
-          key: v1-test_build_doc
-          restore-keys: |
-            v1-test_build_doc-${{ hashFiles('setup.py') }}
-            v1-test_build_doc
-
-      - name: Setup environment
-        run: |
-          pip install --upgrade pip
-          sudo apt-get -y update && sudo apt-get install -y libsndfile1-dev
-
-          pip install git+https://github.com/huggingface/doc-builder
-          pip install .[dev]
-
-          export TORCH_VERSION=$(python -c "from torch import version; print(version.__version__.split('+')[0])")
-          pip install torch-scatter -f https://data.pyg.org/whl/torch-${TORCH_VERSION}+cpu.html
-
-          pip install torchvision
-          python -m pip install 'git+https://github.com/facebookresearch/detectron2.git'
-
-          sudo apt install tesseract-ocr
-          pip install pytesseract
-          pip install pytorch-quantization --extra-index-url https://pypi.ngc.nvidia.com
-
-      - name: Make documentation
-        run: |
-          doc-builder build transformers ./docs/source

.github/workflows/build_documentation.yml

@@ -3,100 +3,18 @@ name: Build documentation
 on:
   push:
     branches:
-      - master
+      - main
       - doc-builder*
       - v*-release
+      - use_templates
 
 jobs:
-  build_and_package:
-    runs-on: ubuntu-latest
-    defaults:
-      run:
-        shell: bash -l {0}
-
-    steps:
-      - uses: actions/checkout@v2
-        with:
-          repository: 'huggingface/doc-build'
-          path: doc-build
-          token: ${{ secrets.HUGGINGFACE_PUSH }}
-      
-      - uses: actions/checkout@v2
-        with:
-          repository: 'huggingface/transformers'
-          path: transformers
-      
-      - uses: actions/checkout@v2
-        with:
-          repository: 'huggingface/notebooks'
-          path: notebooks
-          token: ${{ secrets.HUGGINGFACE_PUSH }}
-      
-      - name: Loading cache.
-        uses: actions/cache@v2
-        id: cache
-        with:
-          path: ~/.cache/pip
-          key: v1-test_build_doc
-          restore-keys: |
-            v1-test_build_doc-${{ hashFiles('setup.py') }}
-            v1-test_build_doc
-
-      - name: Setup environment
-        run: |
-          sudo apt-get -y update && sudo apt-get install -y libsndfile1-dev
-
-          pip install git+https://github.com/huggingface/doc-builder
-          cd transformers
-          pip install .[dev]
-          cd ..
-
-          export TORCH_VERSION=$(python -c "from torch import version; print(version.__version__.split('+')[0])")
-          pip install torch-scatter -f https://data.pyg.org/whl/torch-${TORCH_VERSION}+cpu.html
-
-          pip install torchvision
-          python -m pip install 'git+https://github.com/facebookresearch/detectron2.git'
-
-          sudo apt install tesseract-ocr
-          pip install pytesseract
-          pip install pytorch-quantization --extra-index-url https://pypi.ngc.nvidia.com
-          pip install https://github.com/kpu/kenlm/archive/master.zip
-
-      - name: Setup git
-        run: |
-          git config --global user.name "Hugging Face Doc Builder"
-          git config --global user.email docs@huggingface.co
-          
-          cd doc-build
-          git pull origin main
-          cd ..
-
-          cd notebooks
-          git pull origin master
-          cd ..
-
-      - name: Make documentation
-        run: |
-          doc-builder build transformers transformers/docs/source --build_dir doc-build --notebook_dir notebooks/transformers_doc --clean
-
-      - name: Push to repositories
-        run: |
-          cd doc-build
-          if [[ `git status --porcelain` ]]; then
-            git add .
-            git commit -m "Updated with commit ${{ github.sha }} \n\nSee: https://github.com/huggingface/transformers/commit/${{ github.sha }}"
-            git push origin main
-          else
-            echo "No diff in the documentation."
-          fi
-          cd ..
-
-          cd notebooks
-          if [[ `git status --porcelain` ]]; then
-            git add transformers_doc
-            git commit -m "Updated Transformer doc notebooks with commit ${{ github.sha }} \n\nSee: https://github.com/huggingface/transformers/commit/${{ github.sha }}"
-            git push origin master
-          else
-            echo "No diff in the notebooks."
-          fi
-          cd ..
+   build:
+    uses: huggingface/doc-builder/.github/workflows/build_main_documentation.yml@main
+    with:
+      commit_sha: ${{ github.sha }}
+      package: transformers
+      notebook_folder: transformers_doc
+      languages: en es
+    secrets:
+      token: ${{ secrets.HUGGINGFACE_PUSH }}

.github/workflows/build_pr_documentation.yml

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

.github/workflows/delete_dev_documentation.yml

@@ -1,59 +0,0 @@
-name: Delete dev documentation
-
-on:
-  pull_request:
-    types: [ closed ]
-
-
-jobs:
-  build_and_package:
-    runs-on: [self-hosted, doc-builder]
-    container:
-      image: huggingface/doc-builder-transformers
-      options: "-v /home/github_actions:/mnt"
-    env:
-      PR_NUMBER: ${{ github.event.number }}
-
-    steps:
-      - uses: actions/checkout@v2
-      - name: Set env
-        run: echo "WRITE=$(cat /mnt/WRITE)" >> $GITHUB_ENV
-
-      - uses: actions/checkout@v2
-        with:
-          repository: 'huggingface/doc-build-dev'
-          path: doc-build-dev
-          token: ${{ env.WRITE }}
-
-      - name: Setup git
-        run: |
-          git config --global user.name "Hugging Face Doc Builder"
-          git config --global user.email docs@huggingface.co
-
-      - name: Push to repositories
-        run: |
-          cd doc-build-dev
-          ls
-          rm -rf transformers/pr_$PR_NUMBER
-          ls
-          git add .
-          git commit -m "Closed PR ${GITHUB_REF##*/}"
-          git push origin main
-
-      - name: Find Comment
-        if: ${{ always() }}
-        uses: peter-evans/find-comment@v1
-        id: fc
-        with:
-          issue-number: ${{ env.PR_NUMBER }}
-          comment-author: HuggingFaceDocBuilder
-
-      - name: Update comment
-        if: ${{ always() }}
-        uses: peter-evans/create-or-update-comment@v1
-        with:
-          comment-id: ${{ steps.fc.outputs.comment-id }}
-          token: ${{ env.WRITE }}
-          edit-mode: replace
-          body: |
-            _The documentation is not available anymore as the PR was closed or merged._

.github/workflows/delete_doc_comment.yml

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

.github/workflows/doctests.yml

@@ -15,36 +15,66 @@ env:
   RUN_SLOW: yes
   OMP_NUM_THREADS: 16
   MKL_NUM_THREADS: 16
-  PYTEST_TIMEOUT: 600
+  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
+  TF_FORCE_GPU_ALLOW_GROWTH: true
 
 jobs:
   run_doctests:
-    runs-on: [self-hosted, docker-gpu-test, single-gpu]
+    runs-on: [self-hosted, doc-tests-gpu]
     container:
-      image: pytorch/pytorch:1.9.0-cuda11.1-cudnn8-runtime
+      image: huggingface/transformers-all-latest-gpu
       options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
     steps:
-      - name: Launcher docker
-        uses: actions/checkout@v2
-
+      - uses: actions/checkout@v2
       - name: NVIDIA-SMI
         run: |
           nvidia-smi
 
-      - name: Install dependencies
+      - name: GPU visibility
         run: |
-          apt -y update && apt install -y libsndfile1-dev
-          pip install --upgrade pip
-          pip install .[testing,torch-speech]
+          utils/print_env_pt.py
+          TF_CPP_MIN_LOG_LEVEL=3 python3 -c "import tensorflow as tf; print('TF GPUs available:', bool(tf.config.list_physical_devices('GPU')))"
+          TF_CPP_MIN_LOG_LEVEL=3 python3 -c "import tensorflow as tf; print('Number of TF GPUs available:', len(tf.config.list_physical_devices('GPU')))"
 
       - name: Prepare files for doctests
         run: |
-          python utils/prepare_for_doc_test.py src docs
+          python3 utils/prepare_for_doc_test.py src docs
 
       - name: Run doctests
         run: |
-          pytest --doctest-modules $(cat utils/documentation_tests.txt) -sv --doctest-continue-on-failure --doctest-glob="*.mdx"
+          python3 -m pytest -v --make-reports doc_tests_gpu --doctest-modules $(cat utils/documentation_tests.txt) -sv --doctest-continue-on-failure --doctest-glob="*.mdx"
 
       - name: Clean files after doctests
         run: |
-          python utils/prepare_for_doc_test.py src docs --remove_new_line
+          python3 utils/prepare_for_doc_test.py src docs --remove_new_line
+
+      - name: Failure short reports
+        if: ${{ failure() }}
+        continue-on-error: true
+        run: cat reports/doc_tests_gpu/failures_short.txt
+
+      - name: Test suite reports artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: doc_tests_gpu_test_reports
+          path: reports/doc_tests_gpu
+
+
+  send_results:
+    name: Send results to webhook
+    runs-on: ubuntu-latest
+    if: always()
+    needs: [run_doctests]
+    steps:
+      - uses: actions/checkout@v2
+      - uses: actions/download-artifact@v2
+      - name: Send message to Slack
+        env:
+          CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
+          CI_SLACK_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY_DOCS }}
+          CI_SLACK_CHANNEL_ID_DAILY: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY_DOCS }}
+          CI_SLACK_CHANNEL_DUMMY_TESTS: ${{ secrets.CI_SLACK_CHANNEL_DUMMY_TESTS }}
+        run: |
+          pip install slack_sdk
+          python utils/notification_service_doc_tests.py

.github/workflows/model-templates.yml

@@ -3,7 +3,7 @@ name: Model templates runner
 on:
   push:
     branches:
-      - master
+      - main
   pull_request:
     paths:
       - "src/**"
@@ -60,16 +60,16 @@ jobs:
 
       - name: Run style changes
         run: |
-          git fetch origin master:master
+          git fetch origin main:main
           make style && make quality && make repo-consistency
 
       - name: Failure short reports
         if: ${{ always() }}
-        run: cat reports/tests_templates_failures_short.txt
+        run: cat reports/tests_templates/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
           name: run_all_tests_templates_test_reports
-          path: reports
+          path: reports/tests_templates

.github/workflows/self-push.yml

@@ -3,7 +3,7 @@ name: Self-hosted runner (push)
 on:
   push:
     branches:
-      - master
+      - main
       - ci_*
       - ci-*
     paths:
@@ -82,13 +82,17 @@ jobs:
       image: tensorflow/tensorflow:2.4.1-gpu
       options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
     steps:
+      - name: Set up Python 3.7
+        uses: actions/setup-python@v2
+        with:
+          python-version: 3.7
+
       - name: Install dependencies
         run: |
           apt -y update && apt install -y software-properties-common && apt -y update && add-apt-repository -y ppa:git-core/ppa && apt -y update && apt install -y git espeak-ng
           pip install --upgrade "jax[cuda111]" -f https://storage.googleapis.com/jax-releases/jax_releases.html
           pip install --upgrade pip
           pip install .[sklearn,testing,sentencepiece,flax,flax-speech,vision]
-          pip install https://github.com/kpu/kenlm/archive/master.zip
 
       - name: Launcher docker
         uses: actions/checkout@v2
@@ -492,4 +496,4 @@ jobs:
 
         run: |
           pip install slack_sdk
-          python utils/notification_service.py push
+          python utils/notification_service_deprecated.py push

.github/workflows/self-scheduled.yml

@@ -1,531 +1,254 @@
 name: Self-hosted runner (scheduled)
 
 on:
-  push:
-    branches:
-      - multi_ci_*
   repository_dispatch:
   schedule:
-    - cron: "0 0 * * *"
+    - cron: "0 2 * * *"
 
 env:
   HF_HOME: /mnt/cache
   TRANSFORMERS_IS_CI: yes
+  OMP_NUM_THREADS: 8
+  MKL_NUM_THREADS: 8
   RUN_SLOW: yes
-  OMP_NUM_THREADS: 16
-  MKL_NUM_THREADS: 16
-  PYTEST_TIMEOUT: 600
   SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
+  TF_FORCE_GPU_ALLOW_GROWTH: true
+  RUN_PT_TF_CROSS_TESTS: 1
 
 jobs:
-  run_all_tests_torch_gpu:
-    runs-on: [self-hosted, docker-gpu, single-gpu]
+  setup:
+    name: Setup
+    strategy:
+      matrix:
+        machines: [multi-gpu-docker, single-gpu-docker]
+    runs-on: ${{ matrix.machines }}
     container:
-      image: pytorch/pytorch:1.9.0-cuda11.1-cudnn8-runtime
+      image: huggingface/transformers-all-latest-gpu
       options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    outputs:
+      matrix: ${{ steps.set-matrix.outputs.matrix }}
     steps:
-      - name: Launcher docker
-        uses: actions/checkout@v2
+      - name: Update clone
+        working-directory: /transformers
+        run: |
+          git fetch && git checkout ${{ github.sha }}
+
+      - name: Cleanup
+        working-directory: /transformers
+        run: |
+          rm -rf tests/__pycache__
+          rm -rf reports
+
+      - id: set-matrix
+        name: Identify models to test
+        working-directory: /transformers/tests
+        run: |
+          echo "::set-output name=matrix::$(python3 -c 'import os; x = list(filter(os.path.isdir, os.listdir(os.getcwd()))); x.sort(); print(x)')"
 
       - name: NVIDIA-SMI
         run: |
           nvidia-smi
 
-      - name: Install dependencies
-        run: |
-          apt -y update && apt install -y libsndfile1-dev git espeak-ng
-          pip install --upgrade pip
-          pip install .[integrations,sklearn,testing,onnxruntime,sentencepiece,torch-speech,vision,timm]
-          pip install https://github.com/kpu/kenlm/archive/master.zip
-          python -m pip install 'git+https://github.com/facebookresearch/detectron2.git'
-
-      - name: Are GPUs recognized by our DL frameworks
+      - name: GPU visibility
+        working-directory: /transformers
         run: |
           utils/print_env_pt.py
+          TF_CPP_MIN_LOG_LEVEL=3 python3 -c "import tensorflow as tf; print('TF GPUs available:', bool(tf.config.list_physical_devices('GPU')))"
+          TF_CPP_MIN_LOG_LEVEL=3 python3 -c "import tensorflow as tf; print('Number of TF GPUs available:', len(tf.config.list_physical_devices('GPU')))"
+
+  run_tests_gpu:
+    name: Model tests
+    strategy:
+      fail-fast: false
+      matrix:
+        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
+        machines: [multi-gpu-docker, single-gpu-docker]
+    runs-on: ${{ matrix.machines }}
+    container:
+      image: huggingface/transformers-all-latest-gpu
+      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    needs: setup
+    steps:
+      - name: Echo folder ${{ matrix.folders }}
+        run: echo "${{ matrix.folders }}"
+
+      - name: Update clone
+        working-directory: /transformers
+        run: git fetch && git checkout ${{ github.sha }}
 
       - name: Run all tests on GPU
-        run: |
-          python -m pytest -n 1 -v --dist=loadfile --make-reports=tests_torch_gpu tests
+        working-directory: /transformers
+        run: python3 -m pytest -v --make-reports=${{ matrix.machines }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
 
       - name: Failure short reports
-        if: ${{ always() }}
-        run: cat reports/tests_torch_gpu_failures_short.txt
+        if: ${{ failure() }}
+        continue-on-error: true
+        run: cat /transformers/reports/${{ matrix.machines }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
 
-      - name: Test durations
+      - name: Test suite reports artifacts
         if: ${{ always() }}
-        run: cat reports/tests_torch_gpu_durations.txt
+        uses: actions/upload-artifact@v2
+        with:
+          name: ${{ matrix.machines }}_run_all_tests_gpu_${{ matrix.folders }}_test_reports
+          path: /transformers/reports/${{ matrix.machines }}_tests_gpu_${{ matrix.folders }}
+
+  run_examples_gpu:
+    name: Examples directory
+    runs-on: [self-hosted, single-gpu-docker]
+    container:
+      image: huggingface/transformers-all-latest-gpu
+      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    needs: setup
+    steps:
+      - name: Update clone
+        working-directory: /transformers
+        run: git fetch && git checkout ${{ github.sha }}
 
       - name: Run examples tests on GPU
-        if: ${{ always() }}
-        env:
-          OMP_NUM_THREADS: 16
-          MKL_NUM_THREADS: 16
-          RUN_SLOW: yes
-          HF_HOME: /mnt/cache
-          TRANSFORMERS_IS_CI: yes
+        working-directory: /transformers
         run: |
           pip install -r examples/pytorch/_tests_requirements.txt
-          python -m pytest -n 1 -v --dist=loadfile --make-reports=examples_torch_gpu examples
+          python3 -m pytest -v --make-reports=examples_gpu examples/pytorch
 
       - name: Failure short reports
-        if: ${{ always() }}
-        run: cat reports/examples_torch_gpu_failures_short.txt
-
-      - name: Test durations
-        if: ${{ always() }}
-        run: cat reports/examples_torch_gpu_durations.txt
-
-      - name: Run all pipeline tests on GPU
-        if: ${{ always() }}
-        env:
-          RUN_PIPELINE_TESTS: yes
-        run: |
-          python -m pytest -n 1 -v --dist=loadfile -m is_pipeline_test --make-reports=tests_torch_pipeline_gpu tests
-
-      - name: Failure short reports
-        if: ${{ always() }}
-        run: cat reports/tests_torch_pipeline_gpu_failures_short.txt
-
-      - name: Test durations
-        if: ${{ always() }}
-        run: cat reports/tests_torch_pipeline_gpu_durations.txt
+        if: ${{ failure() }}
+        continue-on-error: true
+        run: cat /transformers/reports/examples_gpu/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
-          name: run_all_tests_torch_gpu_test_reports
-          path: reports
+          name: run_examples_gpu
+          path: /transformers/reports/examples_gpu
 
-  run_all_tests_flax_gpu:
-    runs-on: [self-hosted, docker-gpu-test, single-gpu]
+  run_pipelines_torch_gpu:
+    name: PyTorch pipelines
+    strategy:
+      fail-fast: false
+      matrix:
+        machines: [multi-gpu-docker, single-gpu-docker]
+    runs-on: ${{ matrix.machines }}
     container:
-      image: tensorflow/tensorflow:2.4.1-gpu
+      image: huggingface/transformers-pytorch-gpu
       options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    needs: setup
     steps:
-      - name: Launcher docker
-        uses: actions/checkout@v2
+      - name: Update clone
+        working-directory: /transformers
+        run: git fetch && git checkout ${{ github.sha }}
 
-      - name: NVIDIA-SMI
-        continue-on-error: true
+      - name: Run all pipeline tests on GPU
+        working-directory: /transformers
+        env:
+          RUN_PIPELINE_TESTS: yes
         run: |
-          nvidia-smi
-
-      - name: Install dependencies
-        run: |
-          pip install --upgrade pip
-          pip install --upgrade "jax[cuda111]" -f https://storage.googleapis.com/jax-releases/jax_releases.html
-          pip install .[flax,integrations,sklearn,testing,sentencepiece,flax-speech,vision]
-          pip install https://github.com/kpu/kenlm/archive/master.zip
-
-      - name: Are GPUs recognized by our DL frameworks
-        run: |
-          python -c "from jax.lib import xla_bridge; print('GPU available:', xla_bridge.get_backend().platform)"
-          python -c "import jax; print('Number of GPUs available:', len(jax.local_devices()))"
-
-      - name: Run all tests on GPU
-        run: |
-          python -m pytest -n 1 -v --dist=loadfile --make-reports=tests_flax_gpu tests
+          python3 -m pytest -n 1 -v --dist=loadfile -m is_pipeline_test --make-reports=${{ matrix.machines }}_tests_torch_pipeline_gpu tests
 
       - name: Failure short reports
-        if: ${{ always() }}
-        run: cat reports/tests_flax_gpu_failures_short.txt
-
-      - name: Test durations
-        if: ${{ always() }}
-        run: cat reports/tests_flax_gpu_durations.txt
+        if: ${{ failure() }}
+        continue-on-error: true
+        run: cat /transformers/reports/${{ matrix.machines }}_tests_torch_pipeline_gpu/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
-          name: run_all_tests_flax_gpu_test_reports
-          path: reports
+          name: ${{ matrix.machines }}_run_tests_torch_pipeline_gpu
+          path: /transformers/reports/${{ matrix.machines }}_tests_torch_pipeline_gpu
 
-  run_all_tests_tf_gpu:
-    runs-on: [self-hosted, docker-gpu, single-gpu]
+  run_pipelines_tf_gpu:
+    name: TensorFlow pipelines
+    strategy:
+      fail-fast: false
+      matrix:
+        machines: [multi-gpu-docker, single-gpu-docker]
+    runs-on: ${{ matrix.machines }}
     container:
-      image: tensorflow/tensorflow:2.4.1-gpu
+      image: huggingface/transformers-tensorflow-gpu
       options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    needs: setup
     steps:
-      - name: Launcher docker
-        uses: actions/checkout@v2
-
-      - name: NVIDIA-SMI
+      - name: Update clone
+        working-directory: /transformers
         run: |
-          nvidia-smi
-
-      - name: Install dependencies
-        run: |
-          apt -y update && apt install -y libsndfile1-dev git espeak-ng
-          pip install --upgrade pip
-          pip install .[sklearn,testing,onnx,sentencepiece,tf-speech,vision]
-          pip install https://github.com/kpu/kenlm/archive/master.zip
-
-
-      - name: Are GPUs recognized by our DL frameworks
-        run: |
-          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('TF GPUs available:', bool(tf.config.list_physical_devices('GPU')))"
-          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('Number of TF GPUs available:', len(tf.config.list_physical_devices('GPU')))"
-
-      - name: Run all tests on GPU
-        env:
-          TF_NUM_INTEROP_THREADS: 1
-          TF_NUM_INTRAOP_THREADS: 16
-        run: |
-          python -m pytest -n 1 -v --dist=loadfile --make-reports=tests_tf_gpu tests
-
-      - name: Failure short reports
-        if: ${{ always() }}
-        run: cat reports/tests_tf_gpu_failures_short.txt
-
-      - name: Test durations
-        if: ${{ always() }}
-        run: cat reports/tests_tf_gpu_durations.txt
+          git fetch && git checkout ${{ github.sha }}
 
       - name: Run all pipeline tests on GPU
-        if: ${{ always() }}
-        env:
-          RUN_PIPELINE_TESTS: yes
-          TF_NUM_INTEROP_THREADS: 1
-          TF_NUM_INTRAOP_THREADS: 16
-        run: |
-          python -m pytest -n 1 -v --dist=loadfile -m is_pipeline_test --make-reports=tests_tf_pipeline_gpu tests
-
-      - name: Failure short reports
-        if: ${{ always() }}
-        run: cat reports/tests_tf_pipeline_gpu_failures_short.txt
-
-      - name: Test durations
-        if: ${{ always() }}
-        run: cat reports/tests_tf_pipeline_gpu_durations.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: run_all_tests_tf_gpu_test_reports
-          path: reports
-
-  run_all_examples_torch_xla_tpu:
-    runs-on: [self-hosted, docker-tpu-test, tpu-v3-8]
-    container:
-      image: gcr.io/tpu-pytorch/xla:nightly_3.8_tpuvm
-      options: --privileged -v "/lib/libtpu.so:/lib/libtpu.so" -v /mnt/cache/.cache/huggingface:/mnt/cache/ --shm-size 16G
-    steps:
-      - name: Launcher docker
-        uses: actions/checkout@v2
-
-      - name: Install dependencies
-        run: |
-          pip install --upgrade pip
-          pip install .[testing]
-
-      - name: Are TPUs recognized by our DL frameworks
-        env:
-          XRT_TPU_CONFIG: localservice;0;localhost:51011
-        run: |
-          python -c "import torch_xla.core.xla_model as xm; print(xm.xla_device())"
-
-      - name: Run example tests on TPU
-        env:
-          XRT_TPU_CONFIG: "localservice;0;localhost:51011"
-          MKL_SERVICE_FORCE_INTEL: "1"  # See: https://github.com/pytorch/pytorch/issues/37377
-
-        run: |
-          python -m pytest -n 1 -v --dist=loadfile --make-reports=tests_torch_xla_tpu examples/pytorch/test_xla_examples.py
-
-      - name: Failure short reports
-        if: ${{ always() }}
-        run: cat reports/tests_torch_xla_tpu_failures_short.txt
-
-      - name: Tests durations
-        if: ${{ always() }}
-        run: cat reports/tests_torch_xla_tpu_durations.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: run_all_examples_torch_xla_tpu
-          path: reports
-
-  run_all_tests_torch_multi_gpu:
-    runs-on: [self-hosted, docker-gpu, multi-gpu]
-    container:
-      image: pytorch/pytorch:1.9.0-cuda11.1-cudnn8-runtime
-      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    steps:
-      - name: Launcher docker
-        uses: actions/checkout@v2
-
-      - name: NVIDIA-SMI
-        continue-on-error: true
-        run: |
-          nvidia-smi
-
-      - name: Install dependencies
-        run: |
-          apt -y update && apt install -y libsndfile1-dev git espeak-ng
-          pip install --upgrade pip
-          pip install .[integrations,sklearn,testing,onnxruntime,sentencepiece,torch-speech,vision,timm]
-          pip install https://github.com/kpu/kenlm/archive/master.zip
-          python -m pip install 'git+https://github.com/facebookresearch/detectron2.git'
-
-      - name: Are GPUs recognized by our DL frameworks
-        run: |
-          utils/print_env_pt.py
-
-      - name: Run all tests on GPU
-        env:
-          MKL_SERVICE_FORCE_INTEL: 1
-        run: |
-          python -m pytest -n 1 -v --dist=loadfile --make-reports=tests_torch_multi_gpu tests
-
-      - name: Failure short reports
-        if: ${{ always() }}
-        run: cat reports/tests_torch_multi_gpu_failures_short.txt
-
-      - name: Test durations
-        if: ${{ always() }}
-        run: cat reports/tests_torch_multi_gpu_durations.txt
-
-      - name: Run all pipeline tests on GPU
-        if: ${{ always() }}
+        working-directory: /transformers
         env:
           RUN_PIPELINE_TESTS: yes
         run: |
-          python -m pytest -n 1 -v --dist=loadfile -m is_pipeline_test --make-reports=tests_torch_pipeline_multi_gpu tests
+          python3 -m pytest -n 1 -v --dist=loadfile -m is_pipeline_test --make-reports=${{ matrix.machines }}_tests_tf_pipeline_gpu tests
 
       - name: Failure short reports
         if: ${{ always() }}
-        run: cat reports/tests_torch_pipeline_multi_gpu_failures_short.txt
-
-      - name: Test durations
-        if: ${{ always() }}
-        run: cat reports/tests_torch_pipeline_multi_gpu_durations.txt
+        run: |
+          cat /transformers/reports/${{ matrix.machines }}_tests_tf_pipeline_gpu/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
-          name: run_all_tests_torch_multi_gpu_test_reports
-          path: reports
-
-  run_all_tests_tf_multi_gpu:
-    runs-on: [self-hosted, docker-gpu, multi-gpu]
-    container:
-      image: tensorflow/tensorflow:2.4.1-gpu
-      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    steps:
-      - name: Launcher docker
-        uses: actions/checkout@v2
-
-      - name: NVIDIA-SMI
-        continue-on-error: true
-        run: |
-          nvidia-smi
-
-      - name: Install dependencies
-        run: |
-          apt -y update && apt install -y libsndfile1-dev git espeak-ng
-          pip install --upgrade pip
-          pip install .[sklearn,testing,onnx,sentencepiece,tf-speech,vision]
-          pip install https://github.com/kpu/kenlm/archive/master.zip
-
-      - name: Are GPUs recognized by our DL frameworks
-        run: |
-          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('TF GPUs available:', bool(tf.config.list_physical_devices('GPU')))"
-          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('Number of TF GPUs available:', len(tf.config.list_physical_devices('GPU')))"
-
-      - name: Run all tests on GPU
-        env:
-          TF_NUM_INTEROP_THREADS: 1
-          TF_NUM_INTRAOP_THREADS: 16
-        run: |
-          python -m pytest -n 1 -v --dist=loadfile --make-reports=tests_tf_multi_gpu tests
-
-      - name: Failure short reports
-        if: ${{ always() }}
-        run: cat reports/tests_tf_multi_gpu_failures_short.txt
-
-      - name: Test durations
-        if: ${{ always() }}
-        run: cat reports/tests_tf_multi_gpu_durations.txt
-
-      - name: Run all pipeline tests on GPU
-        if: ${{ always() }}
-        env:
-          RUN_PIPELINE_TESTS: yes
-          TF_NUM_INTEROP_THREADS: 1
-          TF_NUM_INTRAOP_THREADS: 16
-        run: |
-          python -m pytest -n 1 -v --dist=loadfile -m is_pipeline_test --make-reports=tests_tf_pipeline_multi_gpu tests
-
-      - name: Failure short reports
-        if: ${{ always() }}
-        run: cat reports/tests_tf_pipeline_multi_gpu_failures_short.txt
-
-      - name: Test durations
-        if: ${{ always() }}
-        run: cat reports/tests_tf_pipeline_multi_gpu_durations.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: run_all_tests_tf_multi_gpu_test_reports
-          path: reports
-
-#  run_all_tests_flax_multi_gpu:
-#    runs-on: [self-hosted, docker-gpu, multi-gpu]
-#    container:
-#      image: tensorflow/tensorflow:2.4.1-gpu
-#      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-#    steps:
-#      - name: Launcher docker
-#        uses: actions/checkout@v2
-#
-#      - name: NVIDIA-SMI
-#        run: |
-#          nvidia-smi
-#
-#      - name: Install dependencies
-#        run: |
-#          pip install --upgrade pip
-#          pip install --upgrade "jax[cuda111]" -f https://storage.googleapis.com/jax-releases/jax_releases.html
-#          pip install .[flax,integrations,sklearn,testing,sentencepiece,flax-speech,vision]
-#
-#      - name: Are GPUs recognized by our DL frameworks
-#        run: |
-#          python -c "from jax.lib import xla_bridge; print('GPU available:', xla_bridge.get_backend().platform)"
-#          python -c "import jax; print('Number of GPUs available:', len(jax.local_devices()))"
-#
-#      - name: Run all tests on GPU
-#        run: |
-#          python -m pytest -n 1 -v --dist=loadfile --make-reports=tests_flax_gpu tests
-#
-#      - name: Failure short reports
-#        if: ${{ always() }}
-#        run: cat reports/tests_flax_gpu_failures_short.txt
-#
-#      - name: Test suite reports artifacts
-#        if: ${{ always() }}
-#        uses: actions/upload-artifact@v2
-#        with:
-#          name: run_all_tests_flax_gpu_test_reports
-#          path: reports
+          name: ${{ matrix.machines }}_run_tests_tf_pipeline_gpu
+          path: /transformers/reports/${{ matrix.machines }}_tests_tf_pipeline_gpu
 
   run_all_tests_torch_cuda_extensions_gpu:
-    runs-on: [self-hosted, docker-gpu, single-gpu]
+    name: Torch CUDA extension tests
+    strategy:
+      fail-fast: false
+      matrix:
+        machines: [multi-gpu-docker, single-gpu-docker]
+    runs-on: ${{ matrix.machines }}
+    needs: setup
     container:
-      image: nvcr.io/nvidia/pytorch:21.03-py3
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+      image: huggingface/transformers-pytorch-deepspeed-latest-gpu
+      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
     steps:
-      - name: Launcher docker
-        uses: actions/checkout@v2
+      - name: Update clone
+        working-directory: /workspace/transformers
+        run: git fetch && git checkout ${{ github.sha }}
 
-      - name: NVIDIA-SMI
+      - name: Re-compile DeepSpeed
+        working-directory: /workspace
         run: |
-          nvidia-smi
-
-      - name: Install dependencies
-        run: |
-          apt -y update && apt install -y libaio-dev
-          pip install --upgrade pip
-          pip install .[testing,deepspeed]
-
-      - name: Are GPUs recognized by our DL frameworks
-        run: |
-          utils/print_env_pt.py
+          pip install deepspeed # installs the deps correctly
+          rm -rf DeepSpeed
+          git clone https://github.com/microsoft/DeepSpeed && cd DeepSpeed && rm -rf build
+          DS_BUILD_CPU_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install -e . --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check
 
       - name: Run all tests on GPU
+        working-directory: /workspace/transformers
         run: |
-          python -m pytest -n 1 -v --dist=loadfile --make-reports=tests_torch_cuda_extensions_gpu tests/deepspeed tests/extended
+          python -m pytest -v --make-reports=${{ matrix.machines }}_tests_torch_cuda_extensions_gpu tests/deepspeed tests/extended
 
       - name: Failure short reports
-        if: ${{ always() }}
-        run: cat reports/tests_torch_cuda_extensions_gpu_failures_short.txt
-
-      - name: Test durations
-        if: ${{ always() }}
-        run: cat reports/tests_torch_cuda_extensions_gpu_durations.txt
-
-      - name: Test suite reports artifacts
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v2
-        with:
-          name: run_tests_torch_cuda_extensions_gpu_test_reports
-          path: reports
-
-  run_all_tests_torch_cuda_extensions_multi_gpu:
-    runs-on: [self-hosted, docker-gpu, multi-gpu]
-    container:
-      image: nvcr.io/nvidia/pytorch:21.03-py3
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    steps:
-      - name: Launcher docker
-        uses: actions/checkout@v2
-
-      - name: NVIDIA-SMI
+        if: ${{ failure() }}
         continue-on-error: true
-        run: |
-          nvidia-smi
-
-      - name: Install dependencies
-        run: |
-          apt -y update && apt install -y libaio-dev
-          pip install --upgrade pip
-          rm -rf ~/.cache/torch_extensions/ # shared between conflicting builds
-          pip install .[testing,deepspeed,fairscale]
-
-      - name: Are GPUs recognized by our DL frameworks
-        run: |
-          utils/print_env_pt.py
-
-      - name: Run all tests on GPU
-        run: |
-          python -m pytest -n 1 -v --dist=loadfile --make-reports=tests_torch_cuda_extensions_multi_gpu tests/deepspeed tests/extended
-
-      - name: Failure short reports
-        if: ${{ always() }}
-        run: cat reports/tests_torch_cuda_extensions_multi_gpu_failures_short.txt
-
-      - name: Test durations
-        if: ${{ always() }}
-        run: cat reports/tests_torch_cuda_extensions_multi_gpu_durations.txt
+        run: cat /workspace/transformers/reports/${{ matrix.machines }}_tests_torch_cuda_extensions_gpu/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
-          name: run_tests_torch_cuda_extensions_multi_gpu_test_reports
-          path: reports
+          name: ${{ matrix.machines }}_run_tests_torch_cuda_extensions_gpu_test_reports
+          path: /workspace/transformers/reports/${{ matrix.machines }}_tests_torch_cuda_extensions_gpu
+
 
   send_results:
     name: Send results to webhook
     runs-on: ubuntu-latest
     if: always()
-    needs: [
-        run_all_tests_torch_gpu,
-        run_all_tests_tf_gpu,
-        run_all_tests_torch_multi_gpu,
-        run_all_tests_tf_multi_gpu,
-        run_all_tests_torch_cuda_extensions_gpu,
-        run_all_tests_torch_cuda_extensions_multi_gpu
-    ]
+    needs: [setup, run_tests_gpu, run_examples_gpu, run_pipelines_tf_gpu, run_pipelines_torch_gpu, run_all_tests_torch_cuda_extensions_gpu]
     steps:
       - uses: actions/checkout@v2
-
       - uses: actions/download-artifact@v2
-
       - name: Send message to Slack
         env:
           CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
           CI_SLACK_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID }}
           CI_SLACK_CHANNEL_ID_DAILY: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY }}
-
-
+          CI_SLACK_CHANNEL_DUMMY_TESTS: ${{ secrets.CI_SLACK_CHANNEL_DUMMY_TESTS }}
         run: |
           pip install slack_sdk
-          python utils/notification_service.py scheduled
+          python utils/notification_service.py "${{ needs.setup.outputs.matrix }}"

.github/workflows/update_metdata.yml

@@ -3,7 +3,7 @@ name: Update Transformers metadata
 on:
   push:
     branches:
-      - master
+      - main
       - update_transformers_metadata
 
 jobs:

.gitignore

@@ -160,4 +160,7 @@ tags
 .pre-commit*
 
 # .lock
-*.lock
+*.lock
+
+# DS_Store (MacOS)
+.DS_Store

CONTRIBUTING.md

@@ -26,7 +26,7 @@ on the awesome projects it made possible, shout out on Twitter every time it has
 helped you, or simply star the repo to say "thank you".
 
 Whichever way you choose to contribute, please be mindful to respect our
-[code of conduct](https://github.com/huggingface/transformers/blob/master/CODE_OF_CONDUCT.md).
+[code of conduct](https://github.com/huggingface/transformers/blob/main/CODE_OF_CONDUCT.md).
 
 ## You can contribute in so many ways!
 
@@ -92,7 +92,7 @@ If you are willing to contribute the model yourself, let us know so we can best
 guide you.
 
 We have added a **detailed guide and templates** to guide you in the process of adding a new model. You can find them
-in the [`templates`](https://github.com/huggingface/transformers/tree/master/templates) folder.
+in the [`templates`](https://github.com/huggingface/transformers/tree/main/templates) folder.
 
 ### Do you want a new feature (that is not a model)?
 
@@ -114,7 +114,7 @@ If your issue is well written we're already 80% of the way there by the time you
 post it.
 
 We have added **templates** to guide you in the process of adding a new example script for training or testing the
-models in the library. You can find them in the [`templates`](https://github.com/huggingface/transformers/tree/master/templates)
+models in the library. You can find them in the [`templates`](https://github.com/huggingface/transformers/tree/main/templates)
 folder.
 
 ## Start contributing! (Pull Requests)
@@ -148,7 +148,7 @@ Follow these steps to start contributing:
    $ git checkout -b a-descriptive-name-for-my-changes
    ```
 
-   **Do not** work on the `master` branch.
+   **Do not** work on the `main` branch.
 
 4. Set up a development environment by running the following command in a virtual environment:
 
@@ -267,7 +267,7 @@ Follow these steps to start contributing:
 
    ```bash
    $ git fetch upstream
-   $ git rebase upstream/master
+   $ git rebase upstream/main
    ```
 
    Push the changes to your account using:
@@ -317,8 +317,8 @@ See more about the checks run on a pull request in our [PR guide](pr_checks)
 ### Tests
 
 An extensive test suite is included to test the library behavior and several examples. Library tests can be found in
-the [tests folder](https://github.com/huggingface/transformers/tree/master/tests) and examples tests in the
-[examples folder](https://github.com/huggingface/transformers/tree/master/examples).
+the [tests folder](https://github.com/huggingface/transformers/tree/main/tests) and examples tests in the
+[examples folder](https://github.com/huggingface/transformers/tree/main/examples).
 
 We like `pytest` and `pytest-xdist` because it's faster. From the root of the
 repository, here's how to run tests with `pytest` for the library:
@@ -365,10 +365,10 @@ $ python -m unittest discover -s examples -t examples -v
 ### Style guide
 
 For documentation strings, 🤗 Transformers follows the [google style](https://google.github.io/styleguide/pyguide.html).
-Check our [documentation writing guide](https://github.com/huggingface/transformers/tree/master/docs#writing-documentation---specification)
+Check our [documentation writing guide](https://github.com/huggingface/transformers/tree/main/docs#writing-documentation---specification)
 for more information.
 
-#### This guide was heavily inspired by the awesome [scikit-learn guide to contributing](https://github.com/scikit-learn/scikit-learn/blob/master/CONTRIBUTING.md)
+#### This guide was heavily inspired by the awesome [scikit-learn guide to contributing](https://github.com/scikit-learn/scikit-learn/blob/main/CONTRIBUTING.md)
 
 
 ### Develop on Windows
@@ -386,15 +386,15 @@ One way one can run the make command on Window is to pass by MSYS2:
 
 You can now use `make` from any terminal (Powershell, cmd.exe, etc) 🎉
 
-### Syncing forked master with upstream (HuggingFace) master
+### Syncing forked main with upstream (HuggingFace) main
 
 To avoid pinging the upstream repository which adds reference notes to each upstream PR and sends unnecessary notifications to the developers involved in these PRs,
-when syncing the master branch of a forked repository, please, follow these steps:
-1. When possible, avoid syncing with the upstream using a branch and PR on the forked repository. Instead merge directly into the forked master.
+when syncing the main branch of a forked repository, please, follow these steps:
+1. When possible, avoid syncing with the upstream using a branch and PR on the forked repository. Instead merge directly into the forked main.
 2. If a PR is absolutely necessary, use the following steps after checking out your branch:
 ```
 $ git checkout -b your-branch-for-syncing
-$ git pull --squash --no-commit upstream master
+$ git pull --squash --no-commit upstream main
 $ git commit -m '<your message without GitHub references>'
 $ git push --set-upstream origin your-branch-for-syncing
 ```

ISSUES.md

@@ -71,8 +71,8 @@ You are not required to read the following guidelines before opening an issue. H
      File "/transformers/src/transformers/__init__.py", line 34, in <module>
        from . import dependency_versions_check
      File "/transformers/src/transformers/dependency_versions_check.py", line 34, in <module>
-       from .file_utils import is_tokenizers_available
-     File "/transformers/src/transformers/file_utils.py", line 40, in <module>
+       from .utils import is_tokenizers_available
+     File "/transformers/src/transformers/utils/import_utils.py", line 40, in <module>
        from tqdm.auto import tqdm
     ModuleNotFoundError: No module named 'tqdm.auto'
     ```
@@ -124,8 +124,8 @@ You are not required to read the following guidelines before opening an issue. H
      File "/transformers/src/transformers/__init__.py", line 34, in <module>
        from . import dependency_versions_check
      File "/transformers/src/transformers/dependency_versions_check.py", line 34, in <module>
-       from .file_utils import is_tokenizers_available
-     File "/transformers/src/transformers/file_utils.py", line 40, in <module>
+       from .utils import is_tokenizers_available
+     File "/transformers/src/transformers/utils/import_utils.py", line 40, in <module>
        from tqdm.auto import tqdm
    ModuleNotFoundError: No module named 'tqdm.auto'
    ```

Makefile

@@ -1,4 +1,4 @@
-.PHONY: deps_table_update modified_only_fixup extra_quality_checks quality style fixup fix-copies test test-examples docs
+.PHONY: deps_table_update modified_only_fixup extra_style_checks quality style fixup fix-copies test test-examples
 
 # make sure to test the local checkout in scripts and not the pre-installed one (don't use quotes!)
 export PYTHONPATH = src
@@ -48,13 +48,13 @@ quality:
 	isort --check-only $(check_dirs)
 	python utils/custom_init_isort.py --check_only
 	flake8 $(check_dirs)
-	python utils/style_doc.py src/transformers docs/source --max_len 119 --check_only
+	doc-builder style src/transformers docs/source --max_len 119 --check_only --path_to_docs docs/source
 
 # Format source code automatically and check is there are any problems left that need manual fixing
 
 extra_style_checks:
 	python utils/custom_init_isort.py
-	python utils/style_doc.py src/transformers docs/source --max_len 119
+	doc-builder style src/transformers docs/source --max_len 119 --path_to_docs docs/source
 
 # this target runs checks on all files and potentially modifies some of them
 

README.md

@@ -21,9 +21,9 @@ limitations under the License.
 <p>
 <p align="center">
     <a href="https://circleci.com/gh/huggingface/transformers">
-        <img alt="Build" src="https://img.shields.io/circleci/build/github/huggingface/transformers/master">
+        <img alt="Build" src="https://img.shields.io/circleci/build/github/huggingface/transformers/main">
     </a>
-    <a href="https://github.com/huggingface/transformers/blob/master/LICENSE">
+    <a href="https://github.com/huggingface/transformers/blob/main/LICENSE">
         <img alt="GitHub" src="https://img.shields.io/github/license/huggingface/transformers.svg?color=blue">
     </a>
     <a href="https://huggingface.co/docs/transformers/index">
@@ -32,7 +32,7 @@ limitations under the License.
     <a href="https://github.com/huggingface/transformers/releases">
         <img alt="GitHub release" src="https://img.shields.io/github/release/huggingface/transformers.svg">
     </a>
-    <a href="https://github.com/huggingface/transformers/blob/master/CODE_OF_CONDUCT.md">
+    <a href="https://github.com/huggingface/transformers/blob/main/CODE_OF_CONDUCT.md">
         <img alt="Contributor Covenant" src="https://img.shields.io/badge/Contributor%20Covenant-v2.0%20adopted-ff69b4.svg">
     </a>
     <a href="https://zenodo.org/badge/latestdoi/155220641"><img src="https://zenodo.org/badge/155220641.svg" alt="DOI"></a>
@@ -41,9 +41,9 @@ limitations under the License.
 <h4 align="center">
     <p>
         <b>English</b> |
-        <a href="https://github.com/huggingface/transformers/blob/master/README_zh-hans.md">简体中文</a> |
-        <a href="https://github.com/huggingface/transformers/blob/master/README_zh-hant.md">繁體中文</a> |
-        <a href="https://github.com/huggingface/transformers/blob/master/README_ko.md">한국어</a>
+        <a href="https://github.com/huggingface/transformers/blob/main/README_zh-hans.md">简体中文</a> |
+        <a href="https://github.com/huggingface/transformers/blob/main/README_zh-hant.md">繁體中文</a> |
+        <a href="https://github.com/huggingface/transformers/blob/main/README_ko.md">한국어</a>
     <p>
 </h4>
 
@@ -185,7 +185,7 @@ The model itself is a regular [Pytorch `nn.Module`](https://pytorch.org/docs/sta
 
 - This library is not a modular toolbox of building blocks for neural nets. The code in the model files is not refactored with additional abstractions on purpose, so that researchers can quickly iterate on each of the models without diving into additional abstractions/files.
 - The training API is not intended to work on any model but is optimized to work with the models provided by the library. For generic machine learning loops, you should use another library.
-- While we strive to present as many use cases as possible, the scripts in our [examples folder](https://github.com/huggingface/transformers/tree/master/examples) are just that: examples. It is expected that they won't work out-of-the box on your specific problem and that you will be required to change a few lines of code to adapt them to your needs.
+- While we strive to present as many use cases as possible, the scripts in our [examples folder](https://github.com/huggingface/transformers/tree/main/examples) are just that: examples. It is expected that they won't work out-of-the box on your specific problem and that you will be required to change a few lines of code to adapt them to your needs.
 
 ## Installation
 
@@ -229,7 +229,7 @@ Current number of checkpoints: ![](https://img.shields.io/endpoint?url=https://h
 🤗 Transformers currently provides the following architectures (see [here](https://huggingface.co/docs/transformers/model_summary) for a high-level summary of each them):
 
 1. **[ALBERT](https://huggingface.co/docs/transformers/model_doc/albert)** (from Google Research and the Toyota Technological Institute at Chicago) released with the paper [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://arxiv.org/abs/1909.11942), by Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut.
-1. **[BART](https://huggingface.co/docs/transformers/model_doc/bart)** (from Facebook) released with the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/pdf/1910.13461.pdf) by Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Ves Stoyanov and Luke Zettlemoyer.
+1. **[BART](https://huggingface.co/docs/transformers/model_doc/bart)** (from Facebook) released with the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/abs/1910.13461) by Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Ves Stoyanov and Luke Zettlemoyer.
 1. **[BARThez](https://huggingface.co/docs/transformers/model_doc/barthez)** (from École polytechnique) released with the paper [BARThez: a Skilled Pretrained French Sequence-to-Sequence Model](https://arxiv.org/abs/2010.12321) by Moussa Kamal Eddine, Antoine J.-P. Tixier, Michalis Vazirgiannis.
 1. **[BARTpho](https://huggingface.co/docs/transformers/model_doc/bartpho)** (from VinAI Research) released with the paper [BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese](https://arxiv.org/abs/2109.09701) by Nguyen Luong Tran, Duong Minh Le and Dat Quoc Nguyen.
 1. **[BEiT](https://huggingface.co/docs/transformers/model_doc/beit)** (from Microsoft) released with the paper [BEiT: BERT Pre-Training of Image Transformers](https://arxiv.org/abs/2106.08254) by Hangbo Bao, Li Dong, Furu Wei.
@@ -244,31 +244,37 @@ Current number of checkpoints: ![](https://img.shields.io/endpoint?url=https://h
 1. **[ByT5](https://huggingface.co/docs/transformers/model_doc/byt5)** (from Google Research) released with the paper [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) by Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel.
 1. **[CamemBERT](https://huggingface.co/docs/transformers/model_doc/camembert)** (from Inria/Facebook/Sorbonne) released with the paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) by Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suárez*, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah and Benoît Sagot.
 1. **[CANINE](https://huggingface.co/docs/transformers/model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
+1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
 1. **[CLIP](https://huggingface.co/docs/transformers/model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
 1. **[CPM](https://huggingface.co/docs/transformers/model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
 1. **[CTRL](https://huggingface.co/docs/transformers/model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
+1. **[Data2Vec](https://huggingface.co/docs/transformers/model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
 1. **[DeBERTa](https://huggingface.co/docs/transformers/model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
 1. **[DeBERTa-v2](https://huggingface.co/docs/transformers/model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
+1. **[Decision Transformer](https://huggingface.co/docs/transformers/model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
+1. **[DiT](https://huggingface.co/docs/transformers/model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
 1. **[DeiT](https://huggingface.co/docs/transformers/model_doc/deit)** (from Facebook) released with the paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) by Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou.
 1. **[DETR](https://huggingface.co/docs/transformers/model_doc/detr)** (from Facebook) released with the paper [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) by Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko.
 1. **[DialoGPT](https://huggingface.co/docs/transformers/model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
-1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/master/examples/research_projects/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/master/examples/research_projects/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/master/examples/research_projects/distillation) and a German version of DistilBERT.
+1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation) and a German version of DistilBERT.
 1. **[DPR](https://huggingface.co/docs/transformers/model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval
 for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon
 Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
+1. **[DPT](https://huggingface.co/docs/transformers/master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by René Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
 1. **[EncoderDecoder](https://huggingface.co/docs/transformers/model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
 1. **[ELECTRA](https://huggingface.co/docs/transformers/model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
 1. **[FlauBERT](https://huggingface.co/docs/transformers/model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab.
 1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
 1. **[Funnel Transformer](https://huggingface.co/docs/transformers/model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
+1. **[GLPN](https://huggingface.co/docs/transformers/model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
 1. **[GPT](https://huggingface.co/docs/transformers/model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
 1. **[GPT-2](https://huggingface.co/docs/transformers/model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
 1. **[GPT-J](https://huggingface.co/docs/transformers/model_doc/gptj)** (from EleutherAI) released in the repository [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
 1. **[GPT Neo](https://huggingface.co/docs/transformers/model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
 1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (from Facebook) released with the paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) by Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
 1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (from Berkeley) released with the paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) by Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
-1. **[ImageGPT](https://huggingface.co/docs/transformers/master/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
+1. **[ImageGPT](https://huggingface.co/docs/transformers/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
 1. **[LayoutLM](https://huggingface.co/docs/transformers/model_doc/layoutlm)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
 1. **[LayoutLMv2](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) by Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
 1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
@@ -279,23 +285,27 @@ Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
 1. **[LXMERT](https://huggingface.co/docs/transformers/model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
 1. **[M2M100](https://huggingface.co/docs/transformers/model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
 1. **[MarianMT](https://huggingface.co/docs/transformers/model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jörg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
+1. **[MaskFormer](https://huggingface.co/docs/transformers/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
 1. **[MBart](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
 1. **[MBart-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
 1. **[Megatron-BERT](https://huggingface.co/docs/transformers/model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
 1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
 1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
 1. **[MT5](https://huggingface.co/docs/transformers/model_doc/mt5)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
-1. **[Nyströmformer](https://huggingface.co/docs/transformers/master/model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
+1. **[Nyströmformer](https://huggingface.co/docs/transformers/model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
 1. **[Pegasus](https://huggingface.co/docs/transformers/model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
 1. **[Perceiver IO](https://huggingface.co/docs/transformers/model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira.
 1. **[PhoBERT](https://huggingface.co/docs/transformers/model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
+1. **[PLBart](https://huggingface.co/docs/transformers/model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
+1. **[PoolFormer](https://huggingface.co/docs/transformers/model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
 1. **[ProphetNet](https://huggingface.co/docs/transformers/model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
 1. **[QDQBert](https://huggingface.co/docs/transformers/model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
-1. **[REALM](https://huggingface.co/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
+1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
-1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/pdf/2010.12821.pdf) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
-1. **[RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta)** (from Facebook), released together with the paper a [Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
-1. **[RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper a [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/pdf/2104.09864v1.pdf) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
+1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
+1. **[ResNet](https://huggingface.co/docs/transformers/model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun.
+1. **[RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta)** (from Facebook), released together with the paper [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
+1. **[RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/abs/2104.09864) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
 1. **[SegFormer](https://huggingface.co/docs/transformers/model_doc/segformer)** (from NVIDIA) released with the paper [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) by Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo.
 1. **[SEW](https://huggingface.co/docs/transformers/model_doc/sew)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
 1. **[SEW-D](https://huggingface.co/docs/transformers/model_doc/sew_d)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
@@ -303,7 +313,7 @@ Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
 1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (from Facebook), released together with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
 1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (from Tel Aviv University), released together with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
 1. **[SqueezeBert](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
-1. **[Swin Transformer](https://huggingface.co/docs/transformers/master/model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
+1. **[Swin Transformer](https://huggingface.co/docs/transformers/model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
 1. **[T5](https://huggingface.co/docs/transformers/model_doc/t5)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
 1. **[T5v1.1](https://huggingface.co/docs/transformers/model_doc/t5v1.1)** (from Google AI) released in the repository [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
 1. **[TAPAS](https://huggingface.co/docs/transformers/model_doc/tapas)** (from Google AI) released with the paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) by Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos.
@@ -312,20 +322,23 @@ Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
 1. **[UniSpeech](https://huggingface.co/docs/transformers/model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
 1. **[UniSpeechSat](https://huggingface.co/docs/transformers/model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER
 AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
-1. **[ViLT)](https://huggingface.co/docs/transformers/master/model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
+1. **[VAN](https://huggingface.co/docs/transformers/model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/abs/2202.09741) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
+1. **[ViLT](https://huggingface.co/docs/transformers/model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
 1. **[Vision Transformer (ViT)](https://huggingface.co/docs/transformers/model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby.
-1. **[ViTMAE)](https://huggingface.co/docs/transformers/master/model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
+1. **[ViTMAE](https://huggingface.co/docs/transformers/model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
 1. **[VisualBERT](https://huggingface.co/docs/transformers/model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
-1. **[WavLM](https://huggingface.co/docs/transformers/master/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[WavLM](https://huggingface.co/docs/transformers/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
 1. **[Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
-1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/master/transformers/model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
+1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/transformers/model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
+1. **[XGLM](https://huggingface.co/docs/transformers/model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li.
 1. **[XLM](https://huggingface.co/docs/transformers/model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
 1. **[XLM-ProphetNet](https://huggingface.co/docs/transformers/model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
 1. **[XLM-RoBERTa](https://huggingface.co/docs/transformers/model_doc/xlm-roberta)** (from Facebook AI), released together with the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov.
+1. **[XLM-RoBERTa-XL](https://huggingface.co/docs/transformers/model_doc/xlm-roberta-xl)** (from Facebook AI), released together with the paper [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) by Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau.
 1. **[XLNet](https://huggingface.co/docs/transformers/model_doc/xlnet)** (from Google/CMU) released with the paper [​XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
 1. **[XLSR-Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
-1. **[XLS-R](https://huggingface.co/docs/master/transformers/model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
-1. **[YOSO](https://huggingface.co/docs/transformers/master/model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) Once: Linear Cost Self-Attention Via Bernoulli Sampling](https://arxiv.org/abs/2111.09714) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
+1. **[XLS-R](https://huggingface.co/docs/transformers/model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
+1. **[YOSO](https://huggingface.co/docs/transformers/model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) Once: Linear Cost Self-Attention Via Bernoulli Sampling](https://arxiv.org/abs/2111.09714) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
 1. Want to contribute a new model? We have added a **detailed guide and templates** to guide you in the process of adding a new model. You can find them in the [`templates`](./templates) folder of the repository. Be sure to check the [contributing guidelines](./CONTRIBUTING.md) and contact the maintainers or open an issue to collect feedbacks before starting your PR.
 
 To check if each model has an implementation in Flax, PyTorch or TensorFlow, or has an associated tokenizer backed by the 🤗 Tokenizers library, refer to [this table](https://huggingface.co/docs/transformers/index#supported-frameworks).
@@ -339,9 +352,9 @@ These implementations have been tested on several datasets (see the example scri
 |-|-|
 | [Documentation](https://huggingface.co/docs/transformers/) | Full API documentation and tutorials |
 | [Task summary](https://huggingface.co/docs/transformers/task_summary) | Tasks supported by 🤗 Transformers |
-| [Preprocessing tutorial](https://huggingface.co/docstransformers/preprocessing) | Using the `Tokenizer` class to prepare data for the models |
+| [Preprocessing tutorial](https://huggingface.co/docs/transformers/preprocessing) | Using the `Tokenizer` class to prepare data for the models |
 | [Training and fine-tuning](https://huggingface.co/docs/transformers/training) | Using the models provided by 🤗 Transformers in a PyTorch/TensorFlow training loop and the `Trainer` API |
-| [Quick tour: Fine-tuning/usage scripts](https://github.com/huggingface/transformers/tree/master/examples) | Example scripts for fine-tuning models on a wide range of tasks |
+| [Quick tour: Fine-tuning/usage scripts](https://github.com/huggingface/transformers/tree/main/examples) | Example scripts for fine-tuning models on a wide range of tasks |
 | [Model sharing and uploading](https://huggingface.co/docs/transformers/model_sharing) | Upload and share your fine-tuned models with the community |
 | [Migration](https://huggingface.co/docs/transformers/migration) | Migrate to 🤗 Transformers from `pytorch-transformers` or `pytorch-pretrained-bert` |
 

README_ko.md

@@ -21,9 +21,9 @@ limitations under the License.
 <p>
 <p align="center">
     <a href="https://circleci.com/gh/huggingface/transformers">
-        <img alt="Build" src="https://img.shields.io/circleci/build/github/huggingface/transformers/master">
+        <img alt="Build" src="https://img.shields.io/circleci/build/github/huggingface/transformers/main">
     </a>
-    <a href="https://github.com/huggingface/transformers/blob/master/LICENSE">
+    <a href="https://github.com/huggingface/transformers/blob/main/LICENSE">
         <img alt="GitHub" src="https://img.shields.io/github/license/huggingface/transformers.svg?color=blue">
     </a>
     <a href="https://huggingface.co/docs/transformers/index">
@@ -32,7 +32,7 @@ limitations under the License.
     <a href="https://github.com/huggingface/transformers/releases">
         <img alt="GitHub release" src="https://img.shields.io/github/release/huggingface/transformers.svg">
     </a>
-    <a href="https://github.com/huggingface/transformers/blob/master/CODE_OF_CONDUCT.md">
+    <a href="https://github.com/huggingface/transformers/blob/main/CODE_OF_CONDUCT.md">
         <img alt="Contributor Covenant" src="https://img.shields.io/badge/Contributor%20Covenant-v2.0%20adopted-ff69b4.svg">
     </a>
     <a href="https://zenodo.org/badge/latestdoi/155220641"><img src="https://zenodo.org/badge/155220641.svg" alt="DOI"></a>
@@ -41,8 +41,8 @@ limitations under the License.
 <h4 align="center">
     <p>
         <a href="https://github.com/huggingface/transformers/">English</a> |
-        <a href="https://github.com/huggingface/transformers/blob/master/README_zh-hans.md">简体中文</a> |
-        <a href="https://github.com/huggingface/transformers/blob/master/README_zh-hant.md">繁體中文</a> |
+        <a href="https://github.com/huggingface/transformers/blob/main/README_zh-hans.md">简体中文</a> |
+        <a href="https://github.com/huggingface/transformers/blob/main/README_zh-hant.md">繁體中文</a> |
         <b>한국어</b>
     <p>
 </h4>
@@ -166,7 +166,7 @@ limitations under the License.
 
 - 이 라이브러리는 신경망 블록을 만들기 위한 모듈이 아닙니다. 연구자들이 여러 파일을 살펴보지 않고 바로 각 모델을 사용할 수 있도록, 모델 파일 코드의 추상화 수준을 적정하게 유지했습니다.
 - 학습 API는 모든 모델에 적용할 수 있도록 만들어지진 않았지만, 라이브러리가 제공하는 모델들에 적용할 수 있도록 최적화되었습니다. 일반적인 머신 러닝을 위해선, 다른 라이브러리를 사용하세요.
-- 가능한 많은 사용 예시를 보여드리고 싶어서, [예시 폴더](https://github.com/huggingface/transformers/tree/master/examples)의 스크립트를 준비했습니다. 이 스크립트들을 수정 없이 특정한 문제에 바로 적용하지 못할 수 있습니다. 필요에 맞게 일부 코드를 수정해야 할 수 있습니다.
+- 가능한 많은 사용 예시를 보여드리고 싶어서, [예시 폴더](https://github.com/huggingface/transformers/tree/main/examples)의 스크립트를 준비했습니다. 이 스크립트들을 수정 없이 특정한 문제에 바로 적용하지 못할 수 있습니다. 필요에 맞게 일부 코드를 수정해야 할 수 있습니다.
 
 ## 설치
 
@@ -227,27 +227,33 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[CANINE](https://huggingface.co/docs/transformers/model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
 1. **[CLIP](https://huggingface.co/docs/transformers/model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
+1. **[ConvNeXT](https://huggingface.co/docs/transformers/main/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
 1. **[CPM](https://huggingface.co/docs/transformers/model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
 1. **[CTRL](https://huggingface.co/docs/transformers/model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
+1. **[Data2Vec](https://huggingface.co/docs/transformers/main/model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
 1. **[DeBERTa](https://huggingface.co/docs/transformers/model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
 1. **[DeBERTa-v2](https://huggingface.co/docs/transformers/model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
+1. **[Decision Transformer](https://huggingface.co/docs/transformers/model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
 1. **[DeiT](https://huggingface.co/docs/transformers/model_doc/deit)** (from Facebook) released with the paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) by Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou.
 1. **[DETR](https://huggingface.co/docs/transformers/model_doc/detr)** (from Facebook) released with the paper [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) by Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko.
 1. **[DialoGPT](https://huggingface.co/docs/transformers/model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
-1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/master/examples/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/master/examples/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/master/examples/distillation) and a German version of DistilBERT.
+1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/distillation) and a German version of DistilBERT.
+1. **[DiT](https://huggingface.co/docs/transformers/model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
 1. **[DPR](https://huggingface.co/docs/transformers/model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
+1. **[DPT](https://huggingface.co/docs/transformers/master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by René Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
 1. **[ELECTRA](https://huggingface.co/docs/transformers/model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
 1. **[EncoderDecoder](https://huggingface.co/docs/transformers/model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
 1. **[FlauBERT](https://huggingface.co/docs/transformers/model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab.
 1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
 1. **[Funnel Transformer](https://huggingface.co/docs/transformers/model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
+1. **[GLPN](https://huggingface.co/docs/transformers/main/model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
 1. **[GPT](https://huggingface.co/docs/transformers/model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
 1. **[GPT Neo](https://huggingface.co/docs/transformers/model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
 1. **[GPT-2](https://huggingface.co/docs/transformers/model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
 1. **[GPT-J](https://huggingface.co/docs/transformers/model_doc/gptj)** (from EleutherAI) released in the repository [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
 1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (from Facebook) released with the paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) by Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
 1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (from Berkeley) released with the paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) by Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
-1. **[ImageGPT](https://huggingface.co/docs/transformers/master/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
+1. **[ImageGPT](https://huggingface.co/docs/transformers/main/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
 1. **[LayoutLM](https://huggingface.co/docs/transformers/model_doc/layoutlm)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
 1. **[LayoutLMv2](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) by Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
 1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
@@ -257,6 +263,7 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[LXMERT](https://huggingface.co/docs/transformers/model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
 1. **[M2M100](https://huggingface.co/docs/transformers/model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
 1. **[MarianMT](https://huggingface.co/docs/transformers/model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jörg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
+1. **[MaskFormer](https://huggingface.co/docs/transformers/main/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
 1. **[MBart](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
 1. **[MBart-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
 1. **[Megatron-BERT](https://huggingface.co/docs/transformers/model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
@@ -264,15 +271,18 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
 1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
 1. **[MT5](https://huggingface.co/docs/transformers/model_doc/mt5)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
-1. **[Nyströmformer](https://huggingface.co/docs/transformers/master/model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
+1. **[Nyströmformer](https://huggingface.co/docs/transformers/main/model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
 1. **[Pegasus](https://huggingface.co/docs/transformers/model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
 1. **[Perceiver IO](https://huggingface.co/docs/transformers/model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira.
 1. **[PhoBERT](https://huggingface.co/docs/transformers/model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
+1. **[PLBart](https://huggingface.co/docs/transformers/main/model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
+1. **[PoolFormer](https://huggingface.co/docs/transformers/main/model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
 1. **[ProphetNet](https://huggingface.co/docs/transformers/model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
 1. **[QDQBert](https://huggingface.co/docs/transformers/model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
-1. **[REALM](https://huggingface.co/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
+1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/pdf/2010.12821.pdf) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
+1. **[ResNet](https://huggingface.co/docs/transformers/main/model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun. 
 1. **[RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta)** (from Facebook), released together with the paper a [Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
 1. **[RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper a [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/pdf/2104.09864v1.pdf) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
 1. **[SegFormer](https://huggingface.co/docs/transformers/model_doc/segformer)** (from NVIDIA) released with the paper [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) by Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo.
@@ -282,7 +292,7 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (from Facebook), released together with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
 1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (from Tel Aviv University), released together with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
 1. **[SqueezeBert](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
-1. **[Swin Transformer](https://huggingface.co/docs/transformers/master/model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
+1. **[Swin Transformer](https://huggingface.co/docs/transformers/main/model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
 1. **[T5](https://huggingface.co/docs/transformers/model_doc/t5)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
 1. **[T5v1.1](https://huggingface.co/docs/transformers/model_doc/t5v1.1)** (from Google AI) released in the repository [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
 1. **[TAPAS](https://huggingface.co/docs/transformers/model_doc/tapas)** (from Google AI) released with the paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) by Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos.
@@ -290,20 +300,23 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[TrOCR](https://huggingface.co/docs/transformers/model_doc/trocr)** (from Microsoft), released together with the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
 1. **[UniSpeech](https://huggingface.co/docs/transformers/model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
 1. **[UniSpeechSat](https://huggingface.co/docs/transformers/model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
-1. **[ViLT)](https://huggingface.co/docs/transformers/master/model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
+1. **[VAN](https://huggingface.co/docs/transformers/main/model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/pdf/2202.09741.pdf) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
+1. **[ViLT](https://huggingface.co/docs/transformers/main/model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
 1. **[Vision Transformer (ViT)](https://huggingface.co/docs/transformers/model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby.
 1. **[VisualBERT](https://huggingface.co/docs/transformers/model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
-1. **[ViTMAE)](https://huggingface.co/docs/transformers/master/model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
+1. **[ViTMAE](https://huggingface.co/docs/transformers/main/model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
 1. **[Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
-1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/master/transformers/model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
-1. **[WavLM](https://huggingface.co/docs/transformers/master/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/transformers/model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
+1. **[WavLM](https://huggingface.co/docs/transformers/main/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[XGLM](https://huggingface.co/docs/transformers/model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li. 
 1. **[XLM](https://huggingface.co/docs/transformers/model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
 1. **[XLM-ProphetNet](https://huggingface.co/docs/transformers/model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
 1. **[XLM-RoBERTa](https://huggingface.co/docs/transformers/model_doc/xlm-roberta)** (from Facebook AI), released together with the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov.
+1. **[XLM-RoBERTa-XL](https://huggingface.co/docs/transformers/model_doc/xlm-roberta-xl)** (from Facebook AI) released with the paper [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) by Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau.
 1. **[XLNet](https://huggingface.co/docs/transformers/model_doc/xlnet)** (from Google/CMU) released with the paper [​XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
-1. **[XLS-R](https://huggingface.co/docs/master/transformers/model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
+1. **[XLS-R](https://huggingface.co/docs/transformers/model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
 1. **[XLSR-Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
-1. **[YOSO](https://huggingface.co/docs/transformers/master/model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
+1. **[YOSO](https://huggingface.co/docs/transformers/main/model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
 1. 새로운 모델을 올리고 싶나요? 우리가 **상세한 가이드와 템플릿** 으로 새로운 모델을 올리도록 도와드릴게요. 가이드와 템플릿은 이 저장소의 [`templates`](./templates) 폴더에서 확인하실 수 있습니다. [컨트리뷰션 가이드라인](./CONTRIBUTING.md)을 꼭 확인해주시고, PR을 올리기 전에 메인테이너에게 연락하거나 이슈를 오픈해 피드백을 받으시길 바랍니다. 
 
 각 모델이 Flax, PyTorch, TensorFlow으로 구현되었는지 또는 🤗 Tokenizers 라이브러리가 지원하는 토크나이저를 사용하는지 확인하려면, [이 표](https://huggingface.co/docs/transformers/index#supported-frameworks)를 확인하세요.
@@ -318,7 +331,7 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 | [과제 요약](https://huggingface.co/docs/transformers/task_summary) | 🤗 Transformers가 지원하는 과제들 |
 | [전처리 튜토리얼](https://huggingface.co/docs/transformers/preprocessing) | `Tokenizer` 클래스를 이용해 모델을 위한 데이터 준비하기 |
 | [학습과 fine-tuning](https://huggingface.co/docs/transformers/training) | 🤗 Transformers가 제공하는 모델 PyTorch/TensorFlow 학습 과정과 `Trainer` API에서 사용하기 |
-| [퀵 투어: Fine-tuning/사용 스크립트](https://github.com/huggingface/transformers/tree/master/examples) | 다양한 과제에서 모델 fine-tuning하는 예시 스크립트 |
+| [퀵 투어: Fine-tuning/사용 스크립트](https://github.com/huggingface/transformers/tree/main/examples) | 다양한 과제에서 모델 fine-tuning하는 예시 스크립트 |
 | [모델 공유 및 업로드](https://huggingface.co/docs/transformers/model_sharing) | 커뮤니티에 fine-tune된 모델을 업로드 및 공유하기 |
 | [마이그레이션](https://huggingface.co/docs/transformers/migration) | `pytorch-transformers`나 `pytorch-pretrained-bert`에서 🤗 Transformers로 이동하기|
 

README_zh-hans.md

@@ -46,9 +46,9 @@ checkpoint: 检查点
 <p>
 <p align="center">
     <a href="https://circleci.com/gh/huggingface/transformers">
-        <img alt="Build" src="https://img.shields.io/circleci/build/github/huggingface/transformers/master">
+        <img alt="Build" src="https://img.shields.io/circleci/build/github/huggingface/transformers/main">
     </a>
-    <a href="https://github.com/huggingface/transformers/blob/master/LICENSE">
+    <a href="https://github.com/huggingface/transformers/blob/main/LICENSE">
         <img alt="GitHub" src="https://img.shields.io/github/license/huggingface/transformers.svg?color=blue">
     </a>
     <a href="https://huggingface.co/docs/transformers/index">
@@ -57,7 +57,7 @@ checkpoint: 检查点
     <a href="https://github.com/huggingface/transformers/releases">
         <img alt="GitHub release" src="https://img.shields.io/github/release/huggingface/transformers.svg">
     </a>
-    <a href="https://github.com/huggingface/transformers/blob/master/CODE_OF_CONDUCT.md">
+    <a href="https://github.com/huggingface/transformers/blob/main/CODE_OF_CONDUCT.md">
         <img alt="Contributor Covenant" src="https://img.shields.io/badge/Contributor%20Covenant-v2.0%20adopted-ff69b4.svg">
     </a>
     <a href="https://zenodo.org/badge/latestdoi/155220641"><img src="https://zenodo.org/badge/155220641.svg" alt="DOI"></a>
@@ -67,8 +67,8 @@ checkpoint: 检查点
     <p>
         <a href="https://github.com/huggingface/transformers/">English</a> |
         <b>简体中文</b> |
-        <a href="https://github.com/huggingface/transformers/blob/master/README_zh-hant.md">繁體中文</a> |
-        <a href="https://github.com/huggingface/transformers/blob/master/README_ko.md">한국어</a>
+        <a href="https://github.com/huggingface/transformers/blob/main/README_zh-hant.md">繁體中文</a> |
+        <a href="https://github.com/huggingface/transformers/blob/main/README_ko.md">한국어</a>
     <p>
 </h4>
 
@@ -191,7 +191,7 @@ checkpoint: 检查点
 
 - 本库并不是模块化的神经网络工具箱。模型文件中的代码特意呈若璞玉，未经额外抽象封装，以便研究人员快速迭代魔改而不致溺于抽象和文件跳转之中。
 - `Trainer` API 并非兼容任何模型，只为本库之模型优化。若是在寻找适用于通用机器学习的训练循环实现，请另觅他库。
-- 尽管我们已尽力而为，[examples 目录](https://github.com/huggingface/transformers/tree/master/examples)中的脚本也仅为用例而已。对于你的特定问题，它们并不一定开箱即用，可能需要改几行代码以适之。
+- 尽管我们已尽力而为，[examples 目录](https://github.com/huggingface/transformers/tree/main/examples)中的脚本也仅为用例而已。对于你的特定问题，它们并不一定开箱即用，可能需要改几行代码以适之。
 
 ## 安装
 
@@ -251,27 +251,33 @@ conda install -c huggingface transformers
 1. **[CANINE](https://huggingface.co/docs/transformers/model_doc/canine)** (来自 Google Research) 伴随论文 [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) 由 Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting 发布。
 1. **[CLIP](https://huggingface.co/docs/transformers/model_doc/clip)** (来自 OpenAI) 伴随论文 [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) 由 Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever 发布。
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (来自 YituTech) 伴随论文 [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) 由 Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan 发布。
+1. **[ConvNeXT](https://huggingface.co/docs/transformers/main/model_doc/convnext)** (来自 Facebook AI) 伴随论文 [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) 由 Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie 发布。
 1. **[CPM](https://huggingface.co/docs/transformers/model_doc/cpm)** (来自 Tsinghua University) 伴随论文 [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) 由 Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun 发布。
 1. **[CTRL](https://huggingface.co/docs/transformers/model_doc/ctrl)** (来自 Salesforce) 伴随论文 [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) 由 Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher 发布。
+1. **[Data2Vec](https://huggingface.co/docs/transformers/main/model_doc/data2vec)** (来自 Facebook) 伴随论文 [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) 由 Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli 发布。
 1. **[DeBERTa](https://huggingface.co/docs/transformers/model_doc/deberta)** (来自 Microsoft) 伴随论文 [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) 由 Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen 发布。
 1. **[DeBERTa-v2](https://huggingface.co/docs/transformers/model_doc/deberta-v2)** (来自 Microsoft) 伴随论文 [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) 由 Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen 发布。
+1. **[Decision Transformer](https://huggingface.co/docs/transformers/model_doc/decision_transformer)** (来自 Berkeley/Facebook/Google) 伴随论文 [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) 由 Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch 发布。
 1. **[DeiT](https://huggingface.co/docs/transformers/model_doc/deit)** (来自 Facebook) 伴随论文 [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) 由 Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou 发布。
 1. **[DETR](https://huggingface.co/docs/transformers/model_doc/detr)** (来自 Facebook) 伴随论文 [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) 由 Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko 发布。
 1. **[DialoGPT](https://huggingface.co/docs/transformers/model_doc/dialogpt)** (来自 Microsoft Research) 伴随论文 [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) 由 Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan 发布。
-1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (来自 HuggingFace), 伴随论文 [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) 由 Victor Sanh, Lysandre Debut and Thomas Wolf 发布。 同样的方法也应用于压缩 GPT-2 到 [DistilGPT2](https://github.com/huggingface/transformers/tree/master/examples/distillation), RoBERTa 到 [DistilRoBERTa](https://github.com/huggingface/transformers/tree/master/examples/distillation), Multilingual BERT 到 [DistilmBERT](https://github.com/huggingface/transformers/tree/master/examples/distillation) 和德语版 DistilBERT。
+1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (来自 HuggingFace), 伴随论文 [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) 由 Victor Sanh, Lysandre Debut and Thomas Wolf 发布。 同样的方法也应用于压缩 GPT-2 到 [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/distillation), RoBERTa 到 [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/distillation), Multilingual BERT 到 [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/distillation) 和德语版 DistilBERT。
+1. **[DiT](https://huggingface.co/docs/transformers/model_doc/dit)** (来自 Microsoft Research) 伴随论文 [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) 由 Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei 发布。
 1. **[DPR](https://huggingface.co/docs/transformers/model_doc/dpr)** (来自 Facebook) 伴随论文 [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) 由 Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih 发布。
+1. **[DPT](https://huggingface.co/docs/transformers/master/model_doc/dpt)** (来自 Intel Labs) 伴随论文 [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) 由 René Ranftl, Alexey Bochkovskiy, Vladlen Koltun 发布。
 1. **[ELECTRA](https://huggingface.co/docs/transformers/model_doc/electra)** (来自 Google Research/Stanford University) 伴随论文 [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) 由 Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning 发布。
 1. **[EncoderDecoder](https://huggingface.co/docs/transformers/model_doc/encoder-decoder)** (来自 Google Research) 伴随论文 [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) 由 Sascha Rothe, Shashi Narayan, Aliaksei Severyn 发布。
 1. **[FlauBERT](https://huggingface.co/docs/transformers/model_doc/flaubert)** (来自 CNRS) 伴随论文 [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) 由 Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab 发布。
 1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (来自 Google Research) 伴随论文 [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) 由 James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon 发布。
 1. **[Funnel Transformer](https://huggingface.co/docs/transformers/model_doc/funnel)** (来自 CMU/Google Brain) 伴随论文 [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) 由 Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le 发布。
+1. **[GLPN](https://huggingface.co/docs/transformers/main/model_doc/glpn)** (来自 KAIST) 伴随论文 [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) 由 Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim 发布。
 1. **[GPT](https://huggingface.co/docs/transformers/model_doc/openai-gpt)** (来自 OpenAI) 伴随论文 [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) 由 Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever 发布。
 1. **[GPT Neo](https://huggingface.co/docs/transformers/model_doc/gpt_neo)** (来自 EleutherAI) 随仓库 [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) 发布。作者为 Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy 发布。
 1. **[GPT-2](https://huggingface.co/docs/transformers/model_doc/gpt2)** (来自 OpenAI) 伴随论文 [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) 由 Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever** 发布。
 1. **[GPT-J](https://huggingface.co/docs/transformers/model_doc/gptj)** (来自 EleutherAI) 伴随论文 [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) 由 Ben Wang and Aran Komatsuzaki 发布。
 1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (来自 Facebook) 伴随论文 [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) 由 Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed 发布。
 1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (来自 Berkeley) 伴随论文 [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) 由 Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer 发布。
-1. **[ImageGPT](https://huggingface.co/docs/transformers/master/model_doc/imagegpt)** (来自 OpenAI) 伴随论文 [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) 由 Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever 发布。
+1. **[ImageGPT](https://huggingface.co/docs/transformers/main/model_doc/imagegpt)** (来自 OpenAI) 伴随论文 [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) 由 Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever 发布。
 1. **[LayoutLM](https://huggingface.co/docs/transformers/model_doc/layoutlm)** (来自 Microsoft Research Asia) 伴随论文 [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) 由 Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou 发布。
 1. **[LayoutLMv2](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (来自 Microsoft Research Asia) 伴随论文 [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) 由 Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou 发布。
 1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (来自 Microsoft Research Asia) 伴随论文 [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) 由 Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei 发布。
@@ -281,6 +287,7 @@ conda install -c huggingface transformers
 1. **[LXMERT](https://huggingface.co/docs/transformers/model_doc/lxmert)** (来自 UNC Chapel Hill) 伴随论文 [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) 由 Hao Tan and Mohit Bansal 发布。
 1. **[M2M100](https://huggingface.co/docs/transformers/model_doc/m2m_100)** (来自 Facebook) 伴随论文 [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) 由 Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin 发布。
 1. **[MarianMT](https://huggingface.co/docs/transformers/model_doc/marian)** 用 [OPUS](http://opus.nlpl.eu/) 数据训练的机器翻译模型由 Jörg Tiedemann 发布。[Marian Framework](https://marian-nmt.github.io/) 由微软翻译团队开发。
+1. **[MaskFormer](https://huggingface.co/docs/transformers/main/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov 
 1. **[MBart](https://huggingface.co/docs/transformers/model_doc/mbart)** (来自 Facebook) 伴随论文 [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) 由 Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer 发布。
 1. **[MBart-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (来自 Facebook) 伴随论文 [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) 由 Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan 发布。
 1. **[Megatron-BERT](https://huggingface.co/docs/transformers/model_doc/megatron-bert)** (来自 NVIDIA) 伴随论文 [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) 由 Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro 发布。
@@ -288,15 +295,18 @@ conda install -c huggingface transformers
 1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (来自 Studio Ousia) 伴随论文 [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) 由 Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka 发布。
 1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (来自 Microsoft Research) 伴随论文 [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) 由 Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu 发布。
 1. **[MT5](https://huggingface.co/docs/transformers/model_doc/mt5)** (来自 Google AI) 伴随论文 [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) 由 Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel 发布。
-1. **[Nyströmformer](https://huggingface.co/docs/transformers/master/model_doc/nystromformer)** (来自 the University of Wisconsin - Madison) 伴随论文 [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) 由 Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh 发布。
+1. **[Nyströmformer](https://huggingface.co/docs/transformers/main/model_doc/nystromformer)** (来自 the University of Wisconsin - Madison) 伴随论文 [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) 由 Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh 发布。
 1. **[Pegasus](https://huggingface.co/docs/transformers/model_doc/pegasus)** (来自 Google) 伴随论文 [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) 由 Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu 发布。
 1. **[Perceiver IO](https://huggingface.co/docs/transformers/model_doc/perceiver)** (来自 Deepmind) 伴随论文 [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) 由 Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira 发布。
 1. **[PhoBERT](https://huggingface.co/docs/transformers/model_doc/phobert)** (来自 VinAI Research) 伴随论文 [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) 由 Dat Quoc Nguyen and Anh Tuan Nguyen 发布。
+1. **[PLBart](https://huggingface.co/docs/transformers/main/model_doc/plbart)** (来自 UCLA NLP) 伴随论文 [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) 由 Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang 发布。
+1. **[PoolFormer](https://huggingface.co/docs/transformers/main/model_doc/poolformer)** (来自 Sea AI Labs) 伴随论文 [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) 由 Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng 发布。
 1. **[ProphetNet](https://huggingface.co/docs/transformers/model_doc/prophetnet)** (来自 Microsoft Research) 伴随论文 [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) 由 Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou 发布。
 1. **[QDQBert](https://huggingface.co/docs/transformers/model_doc/qdqbert)** (来自 NVIDIA) 伴随论文 [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) 由 Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius 发布。
-1. **[REALM](https://huggingface.co/transformers/model_doc/realm.html)** (来自 Google Research) 伴随论文 [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) 由 Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang 发布。
+1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (来自 Google Research) 伴随论文 [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) 由 Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang 发布。
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (来自 Google Research) 伴随论文 [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) 由 Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya 发布。
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (来自 Google Research) 伴随论文 [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/pdf/2010.12821.pdf) 由 Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder 发布。
+1. **[ResNet](https://huggingface.co/docs/transformers/main/model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun. 
 1. **[RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta)** (来自 Facebook), 伴随论文 [Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) 由 Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov 发布。
 1. **[RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer)** (来自 ZhuiyiTechnology), 伴随论文 [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/pdf/2104.09864v1.pdf) 由 Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu 发布。
 1. **[SegFormer](https://huggingface.co/docs/transformers/model_doc/segformer)** (来自 NVIDIA) 伴随论文 [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) 由 Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo 发布。
@@ -306,7 +316,7 @@ conda install -c huggingface transformers
 1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (来自 Facebook) 伴随论文 [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) 由 Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau 发布。
 1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (来自 Tel Aviv University) 伴随论文 [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) 由 Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy 发布。
 1. **[SqueezeBert](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (来自 Berkeley) 伴随论文 [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) 由 Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer 发布。
-1. **[Swin Transformer](https://huggingface.co/docs/transformers/master/model_doc/swin)** (来自 Microsoft) 伴随论文 [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) 由 Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo 发布。
+1. **[Swin Transformer](https://huggingface.co/docs/transformers/main/model_doc/swin)** (来自 Microsoft) 伴随论文 [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) 由 Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo 发布。
 1. **[T5](https://huggingface.co/docs/transformers/model_doc/t5)** (来自 Google AI) 伴随论文 [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) 由 Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu 发布。
 1. **[T5v1.1](https://huggingface.co/docs/transformers/model_doc/t5v1.1)** (来自 Google AI) 伴随论文 [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) 由 Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu 发布。
 1. **[TAPAS](https://huggingface.co/docs/transformers/model_doc/tapas)** (来自 Google AI) 伴随论文 [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) 由 Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos 发布。
@@ -314,20 +324,23 @@ conda install -c huggingface transformers
 1. **[TrOCR](https://huggingface.co/docs/transformers/model_doc/trocr)** (来自 Microsoft) 伴随论文 [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) 由 Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei 发布。
 1. **[UniSpeech](https://huggingface.co/docs/transformers/model_doc/unispeech)** (来自 Microsoft Research) 伴随论文 [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) 由 Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang 发布。
 1. **[UniSpeechSat](https://huggingface.co/docs/transformers/model_doc/unispeech-sat)** (来自 Microsoft Research) 伴随论文 [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) 由 Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu 发布。
-1. **[ViLT)](https://huggingface.co/docs/transformers/master/model_doc/vilt)** (来自 NAVER AI Lab/Kakao Enterprise/Kakao Brain) 伴随论文 [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) 由 Wonjae Kim, Bokyung Son, Ildoo Kim 发布。
+1. **[VAN](https://huggingface.co/docs/transformers/main/model_doc/van)** (来自 Tsinghua University and Nankai University) 伴随论文 [Visual Attention Network](https://arxiv.org/pdf/2202.09741.pdf) 由 Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu 发布。
+1. **[ViLT](https://huggingface.co/docs/transformers/main/model_doc/vilt)** (来自 NAVER AI Lab/Kakao Enterprise/Kakao Brain) 伴随论文 [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) 由 Wonjae Kim, Bokyung Son, Ildoo Kim 发布。
 1. **[Vision Transformer (ViT)](https://huggingface.co/docs/transformers/model_doc/vit)** (来自 Google AI) 伴随论文 [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) 由 Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby 发布。
 1. **[VisualBERT](https://huggingface.co/docs/transformers/model_doc/visual_bert)** (来自 UCLA NLP) 伴随论文 [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) 由 Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang 发布。
-1. **[ViTMAE)](https://huggingface.co/docs/transformers/master/model_doc/vit_mae)** (来自 Meta AI) 伴随论文 [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) 由 Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick 发布。
+1. **[ViTMAE](https://huggingface.co/docs/transformers/main/model_doc/vit_mae)** (来自 Meta AI) 伴随论文 [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) 由 Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick 发布。
 1. **[Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/wav2vec2)** (来自 Facebook AI) 伴随论文 [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) 由 Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli 发布。
-1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/master/transformers/model_doc/wav2vec2_phoneme)** (来自 Facebook AI) 伴随论文 [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) 由 Qiantong Xu, Alexei Baevski, Michael Auli 发布。
-1. **[WavLM](https://huggingface.co/docs/transformers/master/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/transformers/model_doc/wav2vec2_phoneme)** (来自 Facebook AI) 伴随论文 [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) 由 Qiantong Xu, Alexei Baevski, Michael Auli 发布。
+1. **[WavLM](https://huggingface.co/docs/transformers/main/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[XGLM](https://huggingface.co/docs/transformers/model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li. 
 1. **[XLM](https://huggingface.co/docs/transformers/model_doc/xlm)** (来自 Facebook) 伴随论文 [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) 由 Guillaume Lample and Alexis Conneau 发布。
 1. **[XLM-ProphetNet](https://huggingface.co/docs/transformers/model_doc/xlm-prophetnet)** (来自 Microsoft Research) 伴随论文 [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) 由 Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou 发布。
 1. **[XLM-RoBERTa](https://huggingface.co/docs/transformers/model_doc/xlm-roberta)** (来自 Facebook AI), 伴随论文 [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) 由 Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov 发布。
+1. **[XLM-RoBERTa-XL](https://huggingface.co/docs/transformers/model_doc/xlm-roberta-xl)** (来自 Facebook AI) 伴随论文 [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) 由 Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau 发布。
 1. **[XLNet](https://huggingface.co/docs/transformers/model_doc/xlnet)** (来自 Google/CMU) 伴随论文 [XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) 由 Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le 发布。
-1. **[XLS-R](https://huggingface.co/docs/master/transformers/model_doc/xls_r)** (来自 Facebook AI) 伴随论文 [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) 由 Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli 发布。
+1. **[XLS-R](https://huggingface.co/docs/transformers/model_doc/xls_r)** (来自 Facebook AI) 伴随论文 [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) 由 Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli 发布。
 1. **[XLSR-Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/xlsr_wav2vec2)** (来自 Facebook AI) 伴随论文 [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) 由 Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli 发布。
-1. **[YOSO](https://huggingface.co/docs/transformers/master/model_doc/yoso)** (来自 the University of Wisconsin - Madison) 伴随论文 [You Only Sample (Almost) 由 Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh 发布。
+1. **[YOSO](https://huggingface.co/docs/transformers/main/model_doc/yoso)** (来自 the University of Wisconsin - Madison) 伴随论文 [You Only Sample (Almost) 由 Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh 发布。
 1. 想要贡献新的模型？我们这里有一份**详细指引和模板**来引导你添加新的模型。你可以在 [`templates`](./templates) 目录中找到他们。记得查看 [贡献指南](./CONTRIBUTING.md) 并在开始写 PR 前联系维护人员或开一个新的 issue 来获得反馈。
 
 要检查某个模型是否已有 Flax、PyTorch 或 TensorFlow 的实现，或其是否在 🤗 Tokenizers 库中有对应词符化器（tokenizer），敬请参阅[此表](https://huggingface.co/docs/transformers/index#supported-frameworks)。
@@ -343,7 +356,7 @@ conda install -c huggingface transformers
 | [任务总结](https://huggingface.co/docs/transformers/task_summary) | 🤗 Transformers 支持的任务 |
 | [预处理教程](https://huggingface.co/docs/transformers/preprocessing) | 使用 `Tokenizer` 来为模型准备数据 |
 | [训练和微调](https://huggingface.co/docstransformers/training) | 在 PyTorch/TensorFlow 的训练循环或 `Trainer` API 中使用 🤗 Transformers 提供的模型 |
-| [快速上手：微调和用例脚本](https://github.com/huggingface/transformers/tree/master/examples) | 为各种任务提供的用例脚本 |
+| [快速上手：微调和用例脚本](https://github.com/huggingface/transformers/tree/main/examples) | 为各种任务提供的用例脚本 |
 | [模型分享和上传](https://huggingface.co/docs/transformers/model_sharing) | 和社区上传和分享你微调的模型 |
 | [迁移](https://huggingface.co/docs/transformers/migration) | 从 `pytorch-transformers` 或 `pytorch-pretrained-bert` 迁移到 🤗 Transformers |
 

README_zh-hant.md

@@ -58,9 +58,9 @@ user: 使用者
 <p>
 <p align="center">
     <a href="https://circleci.com/gh/huggingface/transformers">
-        <img alt="Build" src="https://img.shields.io/circleci/build/github/huggingface/transformers/master">
+        <img alt="Build" src="https://img.shields.io/circleci/build/github/huggingface/transformers/main">
     </a>
-    <a href="https://github.com/huggingface/transformers/blob/master/LICENSE">
+    <a href="https://github.com/huggingface/transformers/blob/main/LICENSE">
         <img alt="GitHub" src="https://img.shields.io/github/license/huggingface/transformers.svg?color=blue">
     </a>
     <a href="https://huggingface.co/docs/transformers/index">
@@ -69,7 +69,7 @@ user: 使用者
     <a href="https://github.com/huggingface/transformers/releases">
         <img alt="GitHub release" src="https://img.shields.io/github/release/huggingface/transformers.svg">
     </a>
-    <a href="https://github.com/huggingface/transformers/blob/master/CODE_OF_CONDUCT.md">
+    <a href="https://github.com/huggingface/transformers/blob/main/CODE_OF_CONDUCT.md">
         <img alt="Contributor Covenant" src="https://img.shields.io/badge/Contributor%20Covenant-v2.0%20adopted-ff69b4.svg">
     </a>
     <a href="https://zenodo.org/badge/latestdoi/155220641"><img src="https://zenodo.org/badge/155220641.svg" alt="DOI"></a>
@@ -78,9 +78,9 @@ user: 使用者
 <h4 align="center">
     <p>
         <a href="https://github.com/huggingface/transformers/">English</a> |
-        <a href="https://github.com/huggingface/transformers/blob/master/README_zh-hans.md">简体中文</a> |
+        <a href="https://github.com/huggingface/transformers/blob/main/README_zh-hans.md">简体中文</a> |
         <b>繁體中文</b> |
-        <a href="https://github.com/huggingface/transformers/blob/master/README_ko.md">한국어</a>
+        <a href="https://github.com/huggingface/transformers/blob/main/README_ko.md">한국어</a>
     <p>
 </h4>
 
@@ -203,7 +203,7 @@ Tokenizer 為所有的預訓練模型提供了預處理，並可以直接轉換
 
 - 本函式庫並不是模組化的神經網絡工具箱。模型文件中的程式碼並未做額外的抽象封裝，以便研究人員快速地翻閱及修改程式碼，而不會深陷複雜的類別包裝之中。
 - `Trainer` API 並非相容任何模型，它只為本函式庫中的模型最佳化。對於一般的機器學習用途，請使用其他函式庫。
-- 儘管我們已盡力而為，[examples 目錄](https://github.com/huggingface/transformers/tree/master/examples)中的腳本也僅為範例而已。對於特定問題，它們並不一定隨選即用，可能需要修改幾行程式碼以符合需求。
+- 儘管我們已盡力而為，[examples 目錄](https://github.com/huggingface/transformers/tree/main/examples)中的腳本也僅為範例而已。對於特定問題，它們並不一定隨選即用，可能需要修改幾行程式碼以符合需求。
 
 ## 安裝
 
@@ -263,27 +263,33 @@ conda install -c huggingface transformers
 1. **[CANINE](https://huggingface.co/docs/transformers/model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
 1. **[CLIP](https://huggingface.co/docs/transformers/model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
+1. **[ConvNeXT](https://huggingface.co/docs/transformers/main/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
 1. **[CPM](https://huggingface.co/docs/transformers/model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
 1. **[CTRL](https://huggingface.co/docs/transformers/model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
+1. **[Data2Vec](https://huggingface.co/docs/transformers/main/model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
 1. **[DeBERTa](https://huggingface.co/docs/transformers/model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
 1. **[DeBERTa-v2](https://huggingface.co/docs/transformers/model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
+1. **[Decision Transformer](https://huggingface.co/docs/transformers/model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
 1. **[DeiT](https://huggingface.co/docs/transformers/model_doc/deit)** (from Facebook) released with the paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) by Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou.
 1. **[DETR](https://huggingface.co/docs/transformers/model_doc/detr)** (from Facebook) released with the paper [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) by Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko.
 1. **[DialoGPT](https://huggingface.co/docs/transformers/model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
-1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/master/examples/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/master/examples/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/master/examples/distillation) and a German version of DistilBERT.
+1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/distillation) and a German version of DistilBERT.
+1. **[DiT](https://huggingface.co/docs/transformers/model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
 1. **[DPR](https://huggingface.co/docs/transformers/model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
+1. **[DPT](https://huggingface.co/docs/transformers/master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by René Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
 1. **[ELECTRA](https://huggingface.co/docs/transformers/model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
 1. **[EncoderDecoder](https://huggingface.co/docs/transformers/model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
 1. **[FlauBERT](https://huggingface.co/docs/transformers/model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab.
 1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
 1. **[Funnel Transformer](https://huggingface.co/docs/transformers/model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
+1. **[GLPN](https://huggingface.co/docs/transformers/main/model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
 1. **[GPT](https://huggingface.co/docs/transformers/model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
 1. **[GPT Neo](https://huggingface.co/docs/transformers/model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
 1. **[GPT-2](https://huggingface.co/docs/transformers/model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
 1. **[GPT-J](https://huggingface.co/docs/transformers/model_doc/gptj)** (from EleutherAI) released with the paper [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
 1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (from Facebook) released with the paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) by Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
 1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (from Berkeley) released with the paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) by Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
-1. **[ImageGPT](https://huggingface.co/docs/transformers/master/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
+1. **[ImageGPT](https://huggingface.co/docs/transformers/main/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
 1. **[LayoutLM](https://huggingface.co/docs/transformers/model_doc/layoutlm)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
 1. **[LayoutLMv2](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) by Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
 1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
@@ -293,6 +299,7 @@ conda install -c huggingface transformers
 1. **[LXMERT](https://huggingface.co/docs/transformers/model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
 1. **[M2M100](https://huggingface.co/docs/transformers/model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
 1. **[MarianMT](https://huggingface.co/docs/transformers/model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jörg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
+1. **[MaskFormer](https://huggingface.co/docs/transformers/main/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov 
 1. **[MBart](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
 1. **[MBart-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
 1. **[Megatron-BERT](https://huggingface.co/docs/transformers/model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
@@ -300,15 +307,18 @@ conda install -c huggingface transformers
 1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
 1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
 1. **[MT5](https://huggingface.co/docs/transformers/model_doc/mt5)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
-1. **[Nyströmformer](https://huggingface.co/docs/transformers/master/model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
+1. **[Nyströmformer](https://huggingface.co/docs/transformers/main/model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
 1. **[Pegasus](https://huggingface.co/docs/transformers/model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
 1. **[Perceiver IO](https://huggingface.co/docs/transformers/model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira.
 1. **[PhoBERT](https://huggingface.co/docs/transformers/model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
+1. **[PLBart](https://huggingface.co/docs/transformers/main/model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
+1. **[PoolFormer](https://huggingface.co/docs/transformers/main/model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
 1. **[ProphetNet](https://huggingface.co/docs/transformers/model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
 1. **[QDQBert](https://huggingface.co/docs/transformers/model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
-1. **[REALM](https://huggingface.co/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
+1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/pdf/2010.12821.pdf) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
+1. **[ResNet](https://huggingface.co/docs/transformers/main/model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun. 
 1. **[RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta)** (from Facebook), released together with the paper a [Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
 1. **[RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper a [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/pdf/2104.09864v1.pdf) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
 1. **[SegFormer](https://huggingface.co/docs/transformers/model_doc/segformer)** (from NVIDIA) released with the paper [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) by Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo.
@@ -318,7 +328,7 @@ conda install -c huggingface transformers
 1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (from Facebook) released with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
 1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (from Tel Aviv University) released with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
 1. **[SqueezeBert](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
-1. **[Swin Transformer](https://huggingface.co/docs/transformers/master/model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
+1. **[Swin Transformer](https://huggingface.co/docs/transformers/main/model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
 1. **[T5](https://huggingface.co/docs/transformers/model_doc/t5)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
 1. **[T5v1.1](https://huggingface.co/docs/transformers/model_doc/t5v1.1)** (from Google AI) released with the paper [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
 1. **[TAPAS](https://huggingface.co/docs/transformers/model_doc/tapas)** (from Google AI) released with the paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) by Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos.
@@ -326,20 +336,23 @@ conda install -c huggingface transformers
 1. **[TrOCR](https://huggingface.co/docs/transformers/model_doc/trocr)** (from Microsoft) released with the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
 1. **[UniSpeech](https://huggingface.co/docs/transformers/model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
 1. **[UniSpeechSat](https://huggingface.co/docs/transformers/model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
-1. **[ViLT)](https://huggingface.co/docs/transformers/master/model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
+1. **[VAN](https://huggingface.co/docs/transformers/main/model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/pdf/2202.09741.pdf) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
+1. **[ViLT](https://huggingface.co/docs/transformers/main/model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
 1. **[Vision Transformer (ViT)](https://huggingface.co/docs/transformers/model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby. 
 1. **[VisualBERT](https://huggingface.co/docs/transformers/model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
-1. **[ViTMAE)](https://huggingface.co/docs/transformers/master/model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
+1. **[ViTMAE](https://huggingface.co/docs/transformers/main/model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
 1. **[Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
-1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/master/transformers/model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
-1. **[WavLM](https://huggingface.co/docs/transformers/master/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/transformers/model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
+1. **[WavLM](https://huggingface.co/docs/transformers/main/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[XGLM](https://huggingface.co/docs/transformers/model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li. 
 1. **[XLM](https://huggingface.co/docs/transformers/model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
 1. **[XLM-ProphetNet](https://huggingface.co/docs/transformers/model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
 1. **[XLM-RoBERTa](https://huggingface.co/docs/transformers/model_doc/xlm-roberta)** (from Facebook AI), released together with the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov.
+1. **[XLM-RoBERTa-XL](https://huggingface.co/docs/transformers/model_doc/xlm-roberta-xl)** (from Facebook AI) released with the paper [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) by Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau.
 1. **[XLNet](https://huggingface.co/docs/transformers/model_doc/xlnet)** (from Google/CMU) released with the paper [​XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
-1. **[XLS-R](https://huggingface.co/docs/master/transformers/model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
+1. **[XLS-R](https://huggingface.co/docs/transformers/model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
 1. **[XLSR-Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
-1. **[YOSO](https://huggingface.co/docs/transformers/master/model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
+1. **[YOSO](https://huggingface.co/docs/transformers/main/model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
 1. 想要貢獻新的模型？我們這裡有一份**詳細指引和模板**來引導你加入新的模型。你可以在 [`templates`](./templates) 目錄中找到它們。記得查看[貢獻指引](./CONTRIBUTING.md)並在開始寫 PR 前聯繫維護人員或開一個新的 issue 來獲得 feedbacks。
 
 要檢查某個模型是否已有 Flax、PyTorch 或 TensorFlow 的實作，或其是否在🤗 Tokenizers 函式庫中有對應的 tokenizer，敬請參閱[此表](https://huggingface.co/docs/transformers/index#supported-frameworks)。
@@ -355,7 +368,7 @@ conda install -c huggingface transformers
 | [任務概覽](https://huggingface.co/docs/transformers/task_summary) | 🤗 Transformers 支援的任務 |
 | [預處理教學](https://huggingface.co/docs/transformers/preprocessing) | 使用 `Tokenizer` 來為模型準備資料 |
 | [訓練和微調](https://huggingface.co/docs/transformers/training) | 使用 PyTorch/TensorFlow 的內建的訓練方式或於 `Trainer` API 中使用 🤗 Transformers 提供的模型 |
-| [快速上手：微調和範例腳本](https://github.com/huggingface/transformers/tree/master/examples) | 為各種任務提供的範例腳本 |
+| [快速上手：微調和範例腳本](https://github.com/huggingface/transformers/tree/main/examples) | 為各種任務提供的範例腳本 |
 | [模型分享和上傳](https://huggingface.co/docs/transformers/model_sharing) | 上傳並與社群分享你微調的模型 |
 | [遷移](https://huggingface.co/docs/transformers/migration) | 從 `pytorch-transformers` 或 `pytorch-pretrained-bert` 遷移到 🤗 Transformers |
 

conftest.py

@@ -15,6 +15,7 @@
 # tests directory-specific settings - this file is run automatically
 # by pytest before any tests are run
 
+import doctest
 import sys
 import warnings
 from os.path import abspath, dirname, join
@@ -22,7 +23,7 @@ from os.path import abspath, dirname, join
 
 # allow having multiple repository checkouts and not needing to remember to rerun
 # 'pip install -e .[dev]' when switching between checkouts and running tests.
-git_repo_path = abspath(join(dirname(dirname(__file__)), "src"))
+git_repo_path = abspath(join(dirname(__file__), "src"))
 sys.path.insert(1, git_repo_path)
 
 # silence FutureWarning warnings in tests since often we can't act on them until
@@ -59,3 +60,19 @@ def pytest_sessionfinish(session, exitstatus):
     # If no tests are collected, pytest exists with code 5, which makes the CI fail.
     if exitstatus == 5:
         session.exitstatus = 0
+
+
+# Doctest custom flag to ignore output.
+IGNORE_RESULT = doctest.register_optionflag('IGNORE_RESULT')
+
+OutputChecker = doctest.OutputChecker
+
+
+class CustomOutputChecker(OutputChecker):
+    def check_output(self, want, got, optionflags):
+        if IGNORE_RESULT & optionflags:
+            return True
+        return OutputChecker.check_output(self, want, got, optionflags)
+
+
+doctest.OutputChecker = CustomOutputChecker

docker/transformers-all-latest-gpu/Dockerfile

@@ -0,0 +1,22 @@
+FROM nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04
+LABEL maintainer="Hugging Face"
+
+ARG DEBIAN_FRONTEND=noninteractive
+
+RUN apt update
+RUN apt install -y git libsndfile1-dev tesseract-ocr espeak-ng python3 python3-pip ffmpeg
+RUN python3 -m pip install --no-cache-dir --upgrade pip
+
+ARG REF=main
+RUN git clone https://github.com/huggingface/transformers && cd transformers && git checkout $REF
+RUN python3 -m pip install --no-cache-dir -e ./transformers[dev,onnxruntime]
+
+RUN python3 -m pip install --no-cache-dir -U torch tensorflow
+RUN python3 -m pip uninstall -y flax jax
+RUN python3 -m pip install --no-cache-dir torch-scatter -f https://data.pyg.org/whl/torch-$(python3 -c "from torch import version; print(version.__version__.split('+')[0])")+cu102.html
+RUN python3 -m pip install --no-cache-dir git+https://github.com/facebookresearch/detectron2.git pytesseract https://github.com/kpu/kenlm/archive/master.zip
+RUN python3 -m pip install -U "itsdangerous<2.1.0"
+
+# When installing in editable mode, `transformers` is not recognized as a package.
+# this line must be added in order for python to be aware of transformers.
+RUN cd transformers && python3 setup.py develop

docker/transformers-doc-builder/Dockerfile

@@ -0,0 +1,19 @@
+FROM python:3.8
+LABEL maintainer="Hugging Face"
+
+RUN apt update
+RUN git clone https://github.com/huggingface/transformers
+
+RUN python3 -m pip install --no-cache-dir --upgrade pip && python3 -m pip install --no-cache-dir git+https://github.com/huggingface/doc-builder ./transformers[dev]
+RUN apt-get -y update && apt-get install -y libsndfile1-dev && apt install -y tesseract-ocr
+
+# Torch needs to be installed before deepspeed
+RUN python3 -m pip install --no-cache-dir ./transformers[deepspeed]
+
+RUN python3 -m pip install --no-cache-dir torch-scatter -f https://data.pyg.org/whl/torch-$(python -c "from torch import version; print(version.__version__.split('+')[0])")+cpu.html
+RUN python3 -m pip install --no-cache-dir torchvision git+https://github.com/facebookresearch/detectron2.git pytesseract https://github.com/kpu/kenlm/archive/master.zip
+RUN python3 -m pip install --no-cache-dir pytorch-quantization --extra-index-url https://pypi.ngc.nvidia.com
+RUN python3 -m pip install -U "itsdangerous<2.1.0"
+
+RUN doc-builder build transformers transformers/docs/source --build_dir doc-build-dev --notebook_dir notebooks/transformers_doc --clean --version pr_$PR_NUMBER
+RUN rm -rf doc-build-dev

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

@@ -0,0 +1,21 @@
+FROM nvcr.io/nvidia/pytorch:21.03-py3
+LABEL maintainer="Hugging Face"
+
+ARG DEBIAN_FRONTEND=noninteractive
+
+RUN apt -y update
+RUN apt install -y libaio-dev
+RUN python3 -m pip install --no-cache-dir --upgrade pip
+
+ARG REF=main
+RUN git clone https://github.com/huggingface/transformers && cd transformers && git checkout $REF
+RUN python3 -m pip install --no-cache-dir -e ./transformers[testing,deepspeed]
+
+RUN git clone https://github.com/microsoft/DeepSpeed && cd DeepSpeed && rm -rf build && \
+    DS_BUILD_CPU_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install -e . --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check 2>&1
+
+# When installing in editable mode, `transformers` is not recognized as a package.
+# this line must be added in order for python to be aware of transformers.
+RUN cd transformers && python3 setup.py develop
+
+RUN python3 -c "from deepspeed.launcher.runner import main"

docker/transformers-pytorch-gpu/Dockerfile

@@ -1,30 +1,26 @@
-FROM nvidia/cuda:10.2-cudnn7-devel-ubuntu18.04
+FROM nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04
 LABEL maintainer="Hugging Face"
-LABEL repository="transformers"
 
-RUN apt update && \
-    apt install -y bash \
-                   build-essential \
-                   git \
-                   curl \
-                   ca-certificates \
-                   python3 \
-                   python3-pip && \
-    rm -rf /var/lib/apt/lists
+ARG DEBIAN_FRONTEND=noninteractive
 
-RUN python3 -m pip install --no-cache-dir --upgrade pip && \
-    python3 -m pip install --no-cache-dir \
-    mkl \
-    torch
+RUN apt update
+RUN apt install -y git libsndfile1-dev tesseract-ocr espeak-ng python3 python3-pip ffmpeg
+RUN python3 -m pip install --no-cache-dir --upgrade pip
 
-RUN git clone https://github.com/NVIDIA/apex
-RUN cd apex && \
-    python3 setup.py install && \
-    pip install -v --no-cache-dir --global-option="--cpp_ext" --global-option="--cuda_ext" ./
+ARG REF=main
+RUN git clone https://github.com/huggingface/transformers && cd transformers && git checkout $REF
+RUN python3 -m pip install --no-cache-dir -e ./transformers[dev-torch,testing]
 
-WORKDIR /workspace
-COPY . transformers/
-RUN cd transformers/ && \
-    python3 -m pip install --no-cache-dir .
+# If set to nothing, will install the latest version
+ARG PYTORCH=''
 
-CMD ["/bin/bash"]
+RUN [ ${#PYTORCH} -gt 0 ] && VERSION='torch=='$PYTORCH'.*' ||  VERSION='torch'; python3 -m pip install --no-cache-dir -U $VERSION
+RUN python3 -m pip uninstall -y tensorflow flax
+
+RUN python3 -m pip install --no-cache-dir torch-scatter -f https://data.pyg.org/whl/torch-$(python3 -c "from torch import version; print(version.__version__.split('+')[0])")+cu102.html
+RUN python3 -m pip install --no-cache-dir git+https://github.com/facebookresearch/detectron2.git pytesseract https://github.com/kpu/kenlm/archive/master.zip
+RUN python3 -m pip install -U "itsdangerous<2.1.0"
+
+# When installing in editable mode, `transformers` is not recognized as a package.
+# this line must be added in order for python to be aware of transformers.
+RUN cd transformers && python3 setup.py develop

docker/transformers-pytorch-tpu/Dockerfile

@@ -1,7 +1,7 @@
 FROM google/cloud-sdk:slim
 
 # Build args.
-ARG GITHUB_REF=refs/heads/master
+ARG GITHUB_REF=refs/heads/main
 
 # TODO: This Dockerfile installs pytorch/xla 3.6 wheels. There are also 3.7
 # wheels available; see below.

docker/transformers-tensorflow-gpu/Dockerfile

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

docs/README.md

@@ -39,8 +39,8 @@ check how they look like before committing for instance). You don't have to comm
 
 ## Building the documentation
 
-Once you have setup the `doc-builder` and additional packages, you can generate the documentation by typing th
-following command:
+Once you have setup the `doc-builder` and additional packages, you can generate the documentation by 
+typing the following command:
 
 ```bash
 doc-builder build transformers docs/source/ --build_dir ~/tmp/test-build
@@ -63,7 +63,7 @@ will see a bot add a comment to a link where the documentation with your changes
 Accepted files are Markdown (.md or .mdx).
 
 Create a file with its extension and put it in the source directory. You can then link it to the toc-tree by putting
-the filename without the extension in the [`_toctree.yml`](https://github.com/huggingface/transformers/blob/master/docs/source/_toctree.yml) file.
+the filename without the extension in the [`_toctree.yml`](https://github.com/huggingface/transformers/blob/main/docs/source/_toctree.yml) file.
 
 ## Renaming section headers and moving sections
 
@@ -88,7 +88,7 @@ Sections that were moved:
 
 Use the relative style to link to the new file so that the versioned docs continue to work.
 
-For an example of a rich moved sections set please see the very end of [the Trainer doc](https://github.com/huggingface/transformers/blob/master/docs/source/main_classes/trainer.mdx).
+For an example of a rich moved sections set please see the very end of [the Trainer doc](https://github.com/huggingface/transformers/blob/main/docs/source/main_classes/trainer.mdx).
 
 
 ## Writing Documentation - Specification
@@ -172,11 +172,11 @@ adds a link to its documentation with this syntax: \[\`XXXClass\`\] or \[\`funct
 function to be in the main package.
 
 If you want to create a link to some internal class or function, you need to
-provide its path. For instance: \[\`file_utils.ModelOutput\`\]. This will be converted into a link with
-`file_utils.ModelOutput` in the description. To get rid of the path and only keep the name of the object you are
-linking to in the description, add a ~: \[\`~file_utils.ModelOutput\`\] will generate a link with `ModelOutput` in the description.
+provide its path. For instance: \[\`utils.ModelOutput\`\]. This will be converted into a link with
+`utils.ModelOutput` in the description. To get rid of the path and only keep the name of the object you are
+linking to in the description, add a ~: \[\`~utils.ModelOutput\`\] will generate a link with `ModelOutput` in the description.
 
-The same wroks for methods so you can either use \[\`XXXClass.method\`\] or \[~\`XXXClass.method\`\].
+The same works for methods so you can either use \[\`XXXClass.method\`\] or \[~\`XXXClass.method\`\].
 
 #### Defining arguments in a method
 
@@ -283,3 +283,123 @@ We have an automatic script running with the `make style` comment that will make
 This script may have some weird failures if you made a syntax mistake or if you uncover a bug. Therefore, it's
 recommended to commit your changes before running `make style`, so you can revert the changes done by that script
 easily.
+
+# Testing documentation examples
+
+Good documentation oftens comes with an example of how a specific function or class should be used. 
+Each model class should contain at least one example showcasing
+how to use this model class in inference. *E.g.* the class [Wav2Vec2ForCTC](https://huggingface.co/docs/transformers/model_doc/wav2vec2#transformers.Wav2Vec2ForCTC) 
+includes an example of how to transcribe speech to text in the 
+[docstring of its forward function](https://huggingface.co/docs/transformers/model_doc/wav2vec2#transformers.Wav2Vec2ForCTC.forward).
+
+## Writing documenation examples
+
+The syntax for Example docstrings can look as follows:
+
+```
+    Example:
+
+    ```python
+    >>> from transformers import Wav2Vec2Processor, Wav2Vec2ForCTC
+    >>> from datasets import load_dataset
+    >>> import torch
+
+    >>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation")
+    >>> dataset = dataset.sort("id")
+    >>> sampling_rate = dataset.features["audio"].sampling_rate
+
+    >>> processor = Wav2Vec2Processor.from_pretrained("facebook/wav2vec2-base-960h")
+    >>> model = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-base-960h")
+
+    >>> # audio file is decoded on the fly
+    >>> inputs = processor(dataset[0]["audio"]["array"], sampling_rate=sampling_rate, return_tensors="pt")
+    >>> with torch.no_grad():
+    ...     logits = model(**inputs).logits
+    >>> predicted_ids = torch.argmax(logits, dim=-1)
+
+    >>> # transcribe speech
+    >>> transcription = processor.batch_decode(predicted_ids)
+    >>> transcription[0]
+    'MISTER QUILTER IS THE APOSTLE OF THE MIDDLE CLASSES AND WE ARE GLAD TO WELCOME HIS GOSPEL'
+    ```
+```
+
+The docstring should give a minimal, clear example of how the respective model 
+is to be used in inference and also include the expected (ideally sensible)
+output.
+Often, readers will try out the example before even going through the function 
+or class definitions. Therefore it is of utmost importance that the example 
+works as expected.
+
+## Docstring testing
+
+To do so each example should be included in the doctests. 
+We use pytests' [doctest integration](https://docs.pytest.org/doctest.html) to verify that all of our examples run correctly. 
+For Transformers, the doctests are run on a daily basis via GitHub Actions as can be 
+seen [here](https://github.com/huggingface/transformers/actions/workflows/doctests.yml).
+
+To include your example in the daily doctests, you need add the filename that
+contains the example docstring to the [documentation_tests.txt](../utils/documentation_tests.txt).
+
+### For Python files
+
+You will first need to run the following command (from the root of the repository) to prepare the doc file (doc-testing needs to add additional lines that we don't include in the doc source files):
+
+```bash
+python utils/prepare_for_doc_test.py src docs
+```
+
+If you work on a specific python module, say `modeling_wav2vec2.py`, you can run the command as follows (to avoid the unnecessary temporary changes in irrelevant files):
+
+```bash
+python utils/prepare_for_doc_test.py src/transformers/utils/doc.py src/transformers/models/wav2vec2/modeling_wav2vec2.py
+```
+(`utils/doc.py` should always be included)
+
+Then you can run all the tests in the docstrings of a given file with the following command, here is how we test the modeling file of Wav2Vec2 for instance:
+
+```bash
+pytest --doctest-modules src/transformers/models/wav2vec2/modeling_wav2vec2.py -sv --doctest-continue-on-failure
+```
+
+If you want to isolate a specific docstring, just add `::` after the file name then type the whole path of the function/class/method whose docstring you want to test. For instance, here is how to just test the forward method of `Wav2Vec2ForCTC`:
+
+```bash
+pytest --doctest-modules src/transformers/models/wav2vec2/modeling_wav2vec2.py::transformers.models.wav2vec2.modeling_wav2vec2.Wav2Vec2ForCTC.forward -sv --doctest-continue-on-failure
+```
+
+Once you're done, you can run the following command (still from the root of the repository) to undo the changes made by the first command before committing:
+
+```bash
+python utils/prepare_for_doc_test.py src docs --remove_new_line
+```
+
+### For Markdown files
+
+You will first need to run the following command (from the root of the repository) to prepare the doc file (doc-testing needs to add additional lines that we don't include in the doc source files):
+
+```bash
+python utils/prepare_for_doc_test.py src docs
+```
+
+Then you can test locally a given file with this command (here testing the quicktour):
+
+```bash
+pytest --doctest-modules docs/source/quicktour.mdx -sv --doctest-continue-on-failure --doctest-glob="*.mdx"
+```
+
+Once you're done, you can run the following command (still from the root of the repository) to undo the changes made by the first command before committing:
+
+```bash
+python utils/prepare_for_doc_test.py src docs --remove_new_line
+```
+
+### Writing doctests
+
+Here are a few tips to help you debug the doctests and make them pass:
+
+- The outputs of the code need to match the expected output **exactly**, so make sure you have the same outputs. In particular doctest will see a difference between single quotes and double quotes, or a missing parenthesis. The only exceptions to that rule are:
+  * whitespace: one give whitespace (space, tabulation, new line) is equivalent to any number of whitespace, so you can add new lines where there are spaces to make your output more readable.
+  * numerical values: you should never put more than 4 or 5 digits to expected results as different setups or library versions might get you slightly different results. `doctest` is configure to ignore any difference lower than the precision to which you wrote (so 1e-4 if you write 4 digits).
+- Don't leave a block of code that is very long to execute. If you can't make it fast, you can either not use the doctest syntax on it (so that it's ignored), or if you want to use the doctest syntax to show the results, you can add a comment `# doctest: +SKIP` at the end of the lines of code too long to execute
+- Each line of code that produces a result needs to have that result written below. You can ignore an output if you don't want to show it in your code example by adding a comment ` # doctest: +IGNORE_RESULT` at the end of the line of code produing it.

docs/source/_config.py

@@ -6,4 +6,9 @@ INSTALL_CONTENT = """
 # ! pip install git+https://github.com/huggingface/transformers.git
 """
 
-notebook_first_cells = [{"type": "code", "content": INSTALL_CONTENT}]
+notebook_first_cells = [{"type": "code", "content": INSTALL_CONTENT}]
+black_avoid_patterns = {
+    "{processor_class}": "FakeProcessorClass",
+    "{model_class}": "FakeModelClass",
+    "{object_class}": "FakeObjectClass",    
+}

docs/source/contributing.md

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

docs/source/custom_datasets.mdx

@@ -1,702 +0,0 @@
-<!--Copyright 2020 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
--->
-
-# How to fine-tune a model for common downstream tasks
-
-[[open-in-colab]]
-
-This guide will show you how to fine-tune 🤗 Transformers models for common downstream tasks. You will use the 🤗
-Datasets library to quickly load and preprocess the datasets, getting them ready for training with PyTorch and
-TensorFlow.
-
-Before you begin, make sure you have the 🤗 Datasets library installed. For more detailed installation instructions,
-refer to the 🤗 Datasets [installation page](https://huggingface.co/docs/datasets/installation.html). All of the
-examples in this guide will use 🤗 Datasets to load and preprocess a dataset.
-
-```bash
-pip install datasets
-```
-
-Learn how to fine-tune a model for:
-
-- [seq_imdb](#seq_imdb)
-- [tok_ner](#tok_ner)
-- [qa_squad](#qa_squad)
-
-<a id='seq_imdb'></a>
-
-## Sequence classification with IMDb reviews
-
-Sequence classification refers to the task of classifying sequences of text according to a given number of classes. In
-this example, learn how to fine-tune a model on the [IMDb dataset](https://huggingface.co/datasets/imdb) to determine
-whether a review is positive or negative.
-
-<Tip>
-
-For a more in-depth example of how to fine-tune a model for text classification, take a look at the corresponding
-[PyTorch notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/master/examples/text_classification.ipynb)
-or [TensorFlow notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/master/examples/text_classification-tf.ipynb).
-
-</Tip>
-
-### Load IMDb dataset
-
-The 🤗 Datasets library makes it simple to load a dataset:
-
-```python
-from datasets import load_dataset
-
-imdb = load_dataset("imdb")
-```
-
-This loads a `DatasetDict` object which you can index into to view an example:
-
-```python
-imdb["train"][0]
-{
-    "label": 1,
-    "text": "Bromwell High is a cartoon comedy. It ran at the same time as some other programs about school life, such as \"Teachers\". My 35 years in the teaching profession lead me to believe that Bromwell High's satire is much closer to reality than is \"Teachers\". The scramble to survive financially, the insightful students who can see right through their pathetic teachers' pomp, the pettiness of the whole situation, all remind me of the schools I knew and their students. When I saw the episode in which a student repeatedly tried to burn down the school, I immediately recalled ......... at .......... High. A classic line: INSPECTOR: I'm here to sack one of your teachers. STUDENT: Welcome to Bromwell High. I expect that many adults of my age think that Bromwell High is far fetched. What a pity that it isn't!",
-}
-```
-
-### Preprocess
-
-The next step is to tokenize the text into a readable format by the model. It is important to load the same tokenizer a
-model was trained with to ensure appropriately tokenized words. Load the DistilBERT tokenizer with the
-[`AutoTokenizer`] because we will eventually train a classifier using a pretrained [DistilBERT](https://huggingface.co/distilbert-base-uncased) model:
-
-```python
-from transformers import AutoTokenizer
-
-tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased")
-```
-
-Now that you have instantiated a tokenizer, create a function that will tokenize the text. You should also truncate
-longer sequences in the text to be no longer than the model's maximum input length:
-
-```python
-def preprocess_function(examples):
-    return tokenizer(examples["text"], truncation=True)
-```
-
-Use 🤗 Datasets `map` function to apply the preprocessing function to the entire dataset. You can also set
-`batched=True` to apply the preprocessing function to multiple elements of the dataset at once for faster
-preprocessing:
-
-```python
-tokenized_imdb = imdb.map(preprocess_function, batched=True)
-```
-
-Lastly, pad your text so they are a uniform length. While it is possible to pad your text in the `tokenizer` function
-by setting `padding=True`, it is more efficient to only pad the text to the length of the longest element in its
-batch. This is known as **dynamic padding**. You can do this with the `DataCollatorWithPadding` function:
-
-```python
-from transformers import DataCollatorWithPadding
-
-data_collator = DataCollatorWithPadding(tokenizer=tokenizer)
-```
-
-### Fine-tune with the Trainer API
-
-Now load your model with the [`AutoModelForSequenceClassification`] class along with the number of expected labels:
-
-```python
-from transformers import AutoModelForSequenceClassification
-
-model = AutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased", num_labels=2)
-```
-
-At this point, only three steps remain:
-
-1. Define your training hyperparameters in [`TrainingArguments`].
-2. Pass the training arguments to a [`Trainer`] along with the model, dataset, tokenizer, and data collator.
-3. Call [`Trainer.train()`] to fine-tune your model.
-
-```python
-from transformers import TrainingArguments, Trainer
-
-training_args = TrainingArguments(
-    output_dir="./results",
-    learning_rate=2e-5,
-    per_device_train_batch_size=16,
-    per_device_eval_batch_size=16,
-    num_train_epochs=5,
-    weight_decay=0.01,
-)
-
-trainer = Trainer(
-    model=model,
-    args=training_args,
-    train_dataset=tokenized_imdb["train"],
-    eval_dataset=tokenized_imdb["test"],
-    tokenizer=tokenizer,
-    data_collator=data_collator,
-)
-
-trainer.train()
-```
-
-### Fine-tune with TensorFlow
-
-Fine-tuning with TensorFlow is just as easy, with only a few differences.
-
-Start by batching the processed examples together with dynamic padding using the [`DataCollatorWithPadding`] function.
-Make sure you set `return_tensors="tf"` to return `tf.Tensor` outputs instead of PyTorch tensors!
-
-```python
-from transformers import DataCollatorWithPadding
-
-data_collator = DataCollatorWithPadding(tokenizer, return_tensors="tf")
-```
-
-Next, convert your datasets to the `tf.data.Dataset` format with `to_tf_dataset`. Specify inputs and labels in the
-`columns` argument:
-
-```python
-tf_train_dataset = tokenized_imdb["train"].to_tf_dataset(
-    columns=["attention_mask", "input_ids", "label"],
-    shuffle=True,
-    batch_size=16,
-    collate_fn=data_collator,
-)
-
-tf_validation_dataset = tokenized_imdb["train"].to_tf_dataset(
-    columns=["attention_mask", "input_ids", "label"],
-    shuffle=False,
-    batch_size=16,
-    collate_fn=data_collator,
-)
-```
-
-Set up an optimizer function, learning rate schedule, and some training hyperparameters:
-
-```python
-from transformers import create_optimizer
-import tensorflow as tf
-
-batch_size = 16
-num_epochs = 5
-batches_per_epoch = len(tokenized_imdb["train"]) // batch_size
-total_train_steps = int(batches_per_epoch * num_epochs)
-optimizer, schedule = create_optimizer(init_lr=2e-5, num_warmup_steps=0, num_train_steps=total_train_steps)
-```
-
-Load your model with the [`TFAutoModelForSequenceClassification`] class along with the number of expected labels:
-
-```python
-from transformers import TFAutoModelForSequenceClassification
-
-model = TFAutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased", num_labels=2)
-```
-
-Compile the model:
-
-```python
-import tensorflow as tf
-
-model.compile(optimizer=optimizer)
-```
-
-Finally, fine-tune the model by calling `model.fit`:
-
-```python
-model.fit(
-    tf_train_set,
-    validation_data=tf_validation_set,
-    epochs=num_train_epochs,
-)
-```
-
-<a id='tok_ner'></a>
-
-## Token classification with WNUT emerging entities
-
-Token classification refers to the task of classifying individual tokens in a sentence. One of the most common token
-classification tasks is Named Entity Recognition (NER). NER attempts to find a label for each entity in a sentence,
-such as a person, location, or organization. In this example, learn how to fine-tune a model on the [WNUT 17](https://huggingface.co/datasets/wnut_17) dataset to detect new entities.
-
-<Tip>
-
-For a more in-depth example of how to fine-tune a model for token classification, take a look at the corresponding
-[PyTorch notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/master/examples/token_classification.ipynb)
-or [TensorFlow notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/master/examples/token_classification-tf.ipynb).
-
-</Tip>
-
-### Load WNUT 17 dataset
-
-Load the WNUT 17 dataset from the 🤗 Datasets library:
-
-```python
->>> from datasets import load_dataset
-
->>> wnut = load_dataset("wnut_17")
-```
-
-A quick look at the dataset shows the labels associated with each word in the sentence:
-
-```python
->>> wnut["train"][0]
-{'id': '0',
- 'ner_tags': [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 7, 8, 8, 0, 7, 0, 0, 0, 0, 0, 0, 0, 0],
- 'tokens': ['@paulwalk', 'It', "'s", 'the', 'view', 'from', 'where', 'I', "'m", 'living', 'for', 'two', 'weeks', '.', 'Empire', 'State', 'Building', '=', 'ESB', '.', 'Pretty', 'bad', 'storm', 'here', 'last', 'evening', '.']
-}
-```
-
-View the specific NER tags by:
-
-```python
->>> label_list = wnut["train"].features[f"ner_tags"].feature.names
->>> label_list
-[
-    "O",
-    "B-corporation",
-    "I-corporation",
-    "B-creative-work",
-    "I-creative-work",
-    "B-group",
-    "I-group",
-    "B-location",
-    "I-location",
-    "B-person",
-    "I-person",
-    "B-product",
-    "I-product",
-]
-```
-
-A letter prefixes each NER tag which can mean:
-
-- `B-` indicates the beginning of an entity.
-- `I-` indicates a token is contained inside the same entity (e.g., the `State` token is a part of an entity like
-  `Empire State Building`).
-- `0` indicates the token doesn't correspond to any entity.
-
-### Preprocess
-
-Now you need to tokenize the text. Load the DistilBERT tokenizer with an [`AutoTokenizer`]:
-
-```python
-from transformers import AutoTokenizer
-
-tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased")
-```
-
-Since the input has already been split into words, set `is_split_into_words=True` to tokenize the words into
-subwords:
-
-```python
->>> tokenized_input = tokenizer(example["tokens"], is_split_into_words=True)
->>> tokens = tokenizer.convert_ids_to_tokens(tokenized_input["input_ids"])
->>> tokens
-['[CLS]', '@', 'paul', '##walk', 'it', "'", 's', 'the', 'view', 'from', 'where', 'i', "'", 'm', 'living', 'for', 'two', 'weeks', '.', 'empire', 'state', 'building', '=', 'es', '##b', '.', 'pretty', 'bad', 'storm', 'here', 'last', 'evening', '.', '[SEP]']
-```
-
-The addition of the special tokens `[CLS]` and `[SEP]` and subword tokenization creates a mismatch between the
-input and labels. Realign the labels and tokens by:
-
-1. Mapping all tokens to their corresponding word with the `word_ids` method.
-2. Assigning the label `-100` to the special tokens `[CLS]` and ``[SEP]``` so the PyTorch loss function ignores
-   them.
-3. Only labeling the first token of a given word. Assign `-100` to the other subtokens from the same word.
-
-Here is how you can create a function that will realign the labels and tokens:
-
-```python
-def tokenize_and_align_labels(examples):
-    tokenized_inputs = tokenizer(examples["tokens"], truncation=True, is_split_into_words=True)
-
-    labels = []
-    for i, label in enumerate(examples[f"ner_tags"]):
-        word_ids = tokenized_inputs.word_ids(batch_index=i)  # Map tokens to their respective word.
-        previous_word_idx = None
-        label_ids = []
-        for word_idx in word_ids:  # Set the special tokens to -100.
-            if word_idx is None:
-                label_ids.append(-100)
-            elif word_idx != previous_word_idx:  # Only label the first token of a given word.
-                label_ids.append(label[word_idx])
-            else:
-                label_ids.append(-100)
-            previous_word_idx = word_idx
-        labels.append(label_ids)
-
-    tokenized_inputs["labels"] = labels
-    return tokenized_inputs
-```
-
-Now tokenize and align the labels over the entire dataset with 🤗 Datasets `map` function:
-
-```python
-tokenized_wnut = wnut.map(tokenize_and_align_labels, batched=True)
-```
-
-Finally, pad your text and labels, so they are a uniform length:
-
-```python
-from transformers import DataCollatorForTokenClassification
-
-data_collator = DataCollatorForTokenClassification(tokenizer)
-```
-
-### Fine-tune with the Trainer API
-
-Load your model with the [`AutoModelForTokenClassification`] class along with the number of expected labels:
-
-```python
-from transformers import AutoModelForTokenClassification, TrainingArguments, Trainer
-
-model = AutoModelForTokenClassification.from_pretrained("distilbert-base-uncased", num_labels=len(label_list))
-```
-
-Gather your training arguments in [`TrainingArguments`]:
-
-```python
-training_args = TrainingArguments(
-    output_dir="./results",
-    evaluation_strategy="epoch",
-    learning_rate=2e-5,
-    per_device_train_batch_size=16,
-    per_device_eval_batch_size=16,
-    num_train_epochs=3,
-    weight_decay=0.01,
-)
-```
-
-Collect your model, training arguments, dataset, data collator, and tokenizer in [`Trainer`]:
-
-```python
-trainer = Trainer(
-    model=model,
-    args=training_args,
-    train_dataset=tokenized_wnut["train"],
-    eval_dataset=tokenized_wnut["test"],
-    data_collator=data_collator,
-    tokenizer=tokenizer,
-)
-```
-
-Fine-tune your model:
-
-```python
-trainer.train()
-```
-
-### Fine-tune with TensorFlow
-
-Batch your examples together and pad your text and labels, so they are a uniform length:
-
-```python
-from transformers import DataCollatorForTokenClassification
-
-data_collator = DataCollatorForTokenClassification(tokenizer, return_tensors="tf")
-```
-
-Convert your datasets to the `tf.data.Dataset` format with `to_tf_dataset`:
-
-```python
-tf_train_set = tokenized_wnut["train"].to_tf_dataset(
-    columns=["attention_mask", "input_ids", "labels"],
-    shuffle=True,
-    batch_size=16,
-    collate_fn=data_collator,
-)
-
-tf_validation_set = tokenized_wnut["validation"].to_tf_dataset(
-    columns=["attention_mask", "input_ids", "labels"],
-    shuffle=False,
-    batch_size=16,
-    collate_fn=data_collator,
-)
-```
-
-Load the model with the [`TFAutoModelForTokenClassification`] class along with the number of expected labels:
-
-```python
-from transformers import TFAutoModelForTokenClassification
-
-model = TFAutoModelForTokenClassification.from_pretrained("distilbert-base-uncased", num_labels=len(label_list))
-```
-
-Set up an optimizer function, learning rate schedule, and some training hyperparameters:
-
-```python
-from transformers import create_optimizer
-
-batch_size = 16
-num_train_epochs = 3
-num_train_steps = (len(tokenized_datasets["train"]) // batch_size) * num_train_epochs
-optimizer, lr_schedule = create_optimizer(
-    init_lr=2e-5,
-    num_train_steps=num_train_steps,
-    weight_decay_rate=0.01,
-    num_warmup_steps=0,
-)
-```
-
-Compile the model:
-
-```python
-import tensorflow as tf
-
-model.compile(optimizer=optimizer)
-```
-
-Call `model.fit` to fine-tune your model:
-
-```python
-model.fit(
-    tf_train_set,
-    validation_data=tf_validation_set,
-    epochs=num_train_epochs,
-)
-```
-
-<a id='qa_squad'></a>
-
-## Question Answering with SQuAD
-
-There are many types of question answering (QA) tasks. Extractive QA focuses on identifying the answer from the text
-given a question. In this example, learn how to fine-tune a model on the [SQuAD](https://huggingface.co/datasets/squad) dataset.
-
-<Tip>
-
-For a more in-depth example of how to fine-tune a model for question answering, take a look at the corresponding
-[PyTorch notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/master/examples/question_answering.ipynb)
-or [TensorFlow notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/master/examples/question_answering-tf.ipynb).
-
-</Tip>
-
-### Load SQuAD dataset
-
-Load the SQuAD dataset from the 🤗 Datasets library:
-
-```python
-from datasets import load_dataset
-
-squad = load_dataset("squad")
-```
-
-Take a look at an example from the dataset:
-
-```python
->>> squad["train"][0]
-{'answers': {'answer_start': [515], 'text': ['Saint Bernadette Soubirous']},
- 'context': 'Architecturally, the school has a Catholic character. Atop the Main Building\'s gold dome is a golden statue of the Virgin Mary. Immediately in front of the Main Building and facing it, is a copper statue of Christ with arms upraised with the legend "Venite Ad Me Omnes". Next to the Main Building is the Basilica of the Sacred Heart. Immediately behind the basilica is the Grotto, a Marian place of prayer and reflection. It is a replica of the grotto at Lourdes, France where the Virgin Mary reputedly appeared to Saint Bernadette Soubirous in 1858. At the end of the main drive (and in a direct line that connects through 3 statues and the Gold Dome), is a simple, modern stone statue of Mary.',
- 'id': '5733be284776f41900661182',
- 'question': 'To whom did the Virgin Mary allegedly appear in 1858 in Lourdes France?',
- 'title': 'University_of_Notre_Dame'
-}
-```
-
-### Preprocess
-
-Load the DistilBERT tokenizer with an [`AutoTokenizer`]:
-
-```python
-from transformers import AutoTokenizer
-
-tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased")
-```
-
-There are a few things to be aware of when preprocessing text for question answering:
-
-1. Some examples in a dataset may have a very long `context` that exceeds the maximum input length of the model. You
-   can deal with this by truncating the `context` and set `truncation="only_second"`.
-2. Next, you need to map the start and end positions of the answer to the original context. Set
-   `return_offset_mapping=True` to handle this.
-3. With the mapping in hand, you can find the start and end tokens of the answer. Use the `sequence_ids` method to
-   find which part of the offset corresponds to the question, and which part of the offset corresponds to the context.
-
-Assemble everything in a preprocessing function as shown below:
-
-```python
-def preprocess_function(examples):
-    questions = [q.strip() for q in examples["question"]]
-    inputs = tokenizer(
-        questions,
-        examples["context"],
-        max_length=384,
-        truncation="only_second",
-        return_offsets_mapping=True,
-        padding="max_length",
-    )
-
-    offset_mapping = inputs.pop("offset_mapping")
-    answers = examples["answers"]
-    start_positions = []
-    end_positions = []
-
-    for i, offset in enumerate(offset_mapping):
-        answer = answers[i]
-        start_char = answer["answer_start"][0]
-        end_char = answer["answer_start"][0] + len(answer["text"][0])
-        sequence_ids = inputs.sequence_ids(i)
-
-        # Find the start and end of the context
-        idx = 0
-        while sequence_ids[idx] != 1:
-            idx += 1
-        context_start = idx
-        while sequence_ids[idx] == 1:
-            idx += 1
-        context_end = idx - 1
-
-        # If the answer is not fully inside the context, label it (0, 0)
-        if offset[context_start][0] > end_char or offset[context_end][1] < start_char:
-            start_positions.append(0)
-            end_positions.append(0)
-        else:
-            # Otherwise it's the start and end token positions
-            idx = context_start
-            while idx <= context_end and offset[idx][0] <= start_char:
-                idx += 1
-            start_positions.append(idx - 1)
-
-            idx = context_end
-            while idx >= context_start and offset[idx][1] >= end_char:
-                idx -= 1
-            end_positions.append(idx + 1)
-
-    inputs["start_positions"] = start_positions
-    inputs["end_positions"] = end_positions
-    return inputs
-```
-
-Apply the preprocessing function over the entire dataset with 🤗 Datasets `map` function:
-
-```python
-tokenized_squad = squad.map(preprocess_function, batched=True, remove_columns=squad["train"].column_names)
-```
-
-Batch the processed examples together:
-
-```python
-from transformers import default_data_collator
-
-data_collator = default_data_collator
-```
-
-### Fine-tune with the Trainer API
-
-Load your model with the [`AutoModelForQuestionAnswering`] class:
-
-```python
-from transformers import AutoModelForQuestionAnswering, TrainingArguments, Trainer
-
-model = AutoModelForQuestionAnswering.from_pretrained("distilbert-base-uncased")
-```
-
-Gather your training arguments in [`TrainingArguments`]:
-
-```python
-training_args = TrainingArguments(
-    output_dir="./results",
-    evaluation_strategy="epoch",
-    learning_rate=2e-5,
-    per_device_train_batch_size=16,
-    per_device_eval_batch_size=16,
-    num_train_epochs=3,
-    weight_decay=0.01,
-)
-```
-
-Collect your model, training arguments, dataset, data collator, and tokenizer in [`Trainer`]:
-
-```python
-trainer = Trainer(
-    model=model,
-    args=training_args,
-    train_dataset=tokenized_squad["train"],
-    eval_dataset=tokenized_squad["validation"],
-    data_collator=data_collator,
-    tokenizer=tokenizer,
-)
-```
-
-Fine-tune your model:
-
-```python
-trainer.train()
-```
-
-### Fine-tune with TensorFlow
-
-Batch the processed examples together with a TensorFlow default data collator:
-
-```python
-from transformers.data.data_collator import tf_default_collator
-
-data_collator = tf_default_collator
-```
-
-Convert your datasets to the `tf.data.Dataset` format with the `to_tf_dataset` function:
-
-```python
-tf_train_set = tokenized_squad["train"].to_tf_dataset(
-    columns=["attention_mask", "input_ids", "start_positions", "end_positions"],
-    dummy_labels=True,
-    shuffle=True,
-    batch_size=16,
-    collate_fn=data_collator,
-)
-
-tf_validation_set = tokenized_squad["validation"].to_tf_dataset(
-    columns=["attention_mask", "input_ids", "start_positions", "end_positions"],
-    dummy_labels=True,
-    shuffle=False,
-    batch_size=16,
-    collate_fn=data_collator,
-)
-```
-
-Set up an optimizer function, learning rate schedule, and some training hyperparameters:
-
-```python
-from transformers import create_optimizer
-
-batch_size = 16
-num_epochs = 2
-total_train_steps = (len(tokenized_squad["train"]) // batch_size) * num_epochs
-optimizer, schedule = create_optimizer(
-    init_lr=2e-5,
-    num_warmup_steps=0,
-    num_train_steps=total_train_steps,
-)
-```
-
-Load your model with the [`TFAutoModelForQuestionAnswering`] class:
-
-```python
-from transformers import TFAutoModelForQuestionAnswering
-
-model = TFAutoModelForQuestionAnswering("distilbert-base-uncased")
-```
-
-Compile the model:
-
-```python
-import tensorflow as tf
-
-model.compile(optimizer=optimizer)
-```
-
-Call `model.fit` to fine-tune the model:
-
-```python
-model.fit(
-    tf_train_set,
-    validation_data=tf_validation_set,
-    epochs=num_train_epochs,
-)
-```

docs/source/en/_config.py

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

docs/source/en/_toctree.yml

@@ -5,75 +5,105 @@
     title: Quick tour
   - local: installation
     title: Installation
-  - local: philosophy
-    title: Philosophy
-  - local: glossary
-    title: Glossary
   title: Get started
 - sections:
-  - local: task_summary
-    title: Summary of the tasks
-  - local: model_summary
-    title: Summary of the models
+  - local: pipeline_tutorial
+    title: Pipelines for inference
+  - local: autoclass_tutorial
+    title: Load pretrained instances with an AutoClass
   - local: preprocessing
-    title: Preprocessing data
+    title: Preprocess
   - local: training
-    title: Fine-tuning a pretrained model
+    title: Fine-tune a pretrained model
   - local: accelerate
     title: Distributed training with 🤗 Accelerate
   - local: model_sharing
-    title: Model sharing and uploading
-  - local: tokenizer_summary
-    title: Summary of the tokenizers
-  - local: multilingual
-    title: Multi-lingual models
-  title: "Using 🤗 Transformers"
+    title: Share a model
+  title: Tutorials
 - sections:
-  - local: examples
-    title: Examples
-  - local: troubleshooting
-    title: Troubleshooting
-  - local: custom_datasets
-    title: Fine-tuning with custom datasets
-  - local: notebooks
-    title: "🤗 Transformers Notebooks"
+  - local: fast_tokenizers
+    title: "Use tokenizers from 🤗 Tokenizers"
+  - local: create_a_model
+    title: Create a custom architecture
+  - local: custom_models
+    title: Sharing custom models
+  - sections:
+    - local: tasks/sequence_classification
+      title: Text classification
+    - local: tasks/token_classification
+      title: Token classification
+    - local: tasks/question_answering
+      title: Question answering
+    - local: tasks/language_modeling
+      title: Language modeling
+    - local: tasks/translation
+      title: Translation
+    - local: tasks/summarization
+      title: Summarization
+    - local: tasks/multiple_choice
+      title: Multiple choice
+    - local: tasks/audio_classification
+      title: Audio classification
+    - local: tasks/asr
+      title: Automatic speech recognition
+    - local: tasks/image_classification
+      title: Image classification
+    title: Fine-tune for downstream tasks
+  - local: run_scripts
+    title: Train with a script
   - local: sagemaker
     title: Run training on Amazon SageMaker
-  - local: community
-    title: Community
+  - local: multilingual
+    title: Inference for multilingual models
   - local: converting_tensorflow_models
-    title: Converting Tensorflow Checkpoints
+    title: Converting TensorFlow Checkpoints
+  - local: serialization
+    title: Export 🤗 Transformers models
+  - local: performance
+    title: 'Performance and Scalability: How To Fit a Bigger Model and Train It Faster'
+  - local: parallelism
+    title: Model Parallelism
+  - local: benchmarks
+    title: Benchmarks
   - local: migration
     title: Migrating from previous packages
+  - local: troubleshooting
+    title: Troubleshoot
+  - local: debugging
+    title: Debugging
+  - local: notebooks
+    title: "🤗 Transformers Notebooks"
+  - local: community
+    title: Community
   - local: contributing
     title: How to contribute to transformers?
   - local: add_new_model
     title: "How to add a model to 🤗 Transformers?"
   - local: add_new_pipeline
     title: "How to add a pipeline to 🤗 Transformers?"
-  - local: fast_tokenizers
-    title: "Using tokenizers from 🤗 Tokenizers"
-  - local: performance
-    title: 'Performance and Scalability: How To Fit a Bigger Model and Train It Faster'
-  - local: parallelism
-    title: Model Parallelism
   - local: testing
     title: Testing
-  - local: debugging
-    title: Debugging
-  - local: serialization
-    title: Exporting 🤗 Transformers models
   - local: pr_checks
     title: Checks on a Pull Request
-  title: Advanced guides
+  title: How-to guides
 - sections:
+  - local: philosophy
+    title: Philosophy
+  - local: glossary
+    title: Glossary
+  - local: task_summary
+    title: Summary of the tasks
+  - local: model_summary
+    title: Summary of the models
+  - local: tokenizer_summary
+    title: Summary of the tokenizers
+  - local: pad_truncation
+    title: Padding and truncation
   - local: bertology
     title: BERTology
   - local: perplexity
     title: Perplexity of fixed-length models
-  - local: benchmarks
-    title: Benchmarks
-  title: Research
+  title: Conceptual guides
 - sections:
   - sections:
     - local: main_classes/callback
@@ -88,6 +118,8 @@
       title: Logging
     - local: main_classes/model
       title: Models
+    - local: main_classes/text_generation
+      title: Text Generation
     - local: main_classes/onnx
       title: ONNX
     - local: main_classes/optimizer_schedules
@@ -144,6 +176,8 @@
       title: CamemBERT
     - local: model_doc/canine
       title: CANINE
+    - local: model_doc/convnext
+      title: ConvNeXT
     - local: model_doc/clip
       title: CLIP
     - local: model_doc/convbert
@@ -152,10 +186,14 @@
       title: CPM
     - local: model_doc/ctrl
       title: CTRL
+    - local: model_doc/data2vec
+      title: Data2Vec
     - local: model_doc/deberta
       title: DeBERTa
     - local: model_doc/deberta-v2
       title: DeBERTa-v2
+    - local: model_doc/decision_transformer
+      title: Decision Transformer
     - local: model_doc/deit
       title: DeiT
     - local: model_doc/detr
@@ -164,8 +202,12 @@
       title: DialoGPT
     - local: model_doc/distilbert
       title: DistilBERT
+    - local: model_doc/dit
+      title: DiT
     - local: model_doc/dpr
       title: DPR
+    - local: model_doc/dpt
+      title: DPT
     - local: model_doc/electra
       title: ELECTRA
     - local: model_doc/encoder-decoder
@@ -178,6 +220,8 @@
       title: FSMT
     - local: model_doc/funnel
       title: Funnel Transformer
+    - local: model_doc/glpn
+      title: GLPN
     - local: model_doc/herbert
       title: HerBERT
     - local: model_doc/ibert
@@ -200,6 +244,8 @@
       title: LXMERT
     - local: model_doc/marian
       title: MarianMT
+    - local: model_doc/maskformer
+      title: MaskFormer
     - local: model_doc/m2m_100
       title: M2M100
     - local: model_doc/mbart
@@ -209,11 +255,9 @@
     - local: model_doc/megatron_gpt2
       title: MegatronGPT2
     - local: model_doc/mluke
-      title: MLUKE
+      title: mLUKE
     - local: model_doc/mobilebert
       title: MobileBERT
-    - local: model_doc/mluke
-      title: mLUKE
     - local: model_doc/mpnet
       title: MPNet
     - local: model_doc/mt5
@@ -236,6 +280,10 @@
       title: Pegasus
     - local: model_doc/phobert
       title: PhoBERT
+    - local: model_doc/plbart
+      title: PLBart
+    - local: model_doc/poolformer
+      title: PoolFormer
     - local: model_doc/prophetnet
       title: ProphetNet
     - local: model_doc/qdqbert
@@ -248,6 +296,8 @@
       title: Reformer
     - local: model_doc/rembert
       title: RemBERT
+    - local: model_doc/resnet
+      title: ResNet
     - local: model_doc/retribert
       title: RetriBERT
     - local: model_doc/roberta
@@ -286,6 +336,8 @@
       title: UniSpeech
     - local: model_doc/unispeech-sat
       title: UniSpeech-SAT
+    - local: model_doc/van
+      title: VAN
     - local: model_doc/vilt
       title: ViLT
     - local: model_doc/vision-encoder-decoder
@@ -304,12 +356,16 @@
       title: Wav2Vec2Phoneme
     - local: model_doc/wavlm
       title: WavLM
+    - local: model_doc/xglm
+      title: XGLM
     - local: model_doc/xlm
       title: XLM
     - local: model_doc/xlm-prophetnet
       title: XLM-ProphetNet
     - local: model_doc/xlm-roberta
       title: XLM-RoBERTa
+    - local: model_doc/xlm-roberta-xl
+      title: XLM-RoBERTa-XL
     - local: model_doc/xlnet
       title: XLNet
     - local: model_doc/xlsr_wav2vec2

docs/source/en/accelerate.mdx

@@ -22,7 +22,7 @@ Get started by installing 🤗 Accelerate:
 pip install accelerate
 ```
 
-Then import and create an [`Accelerator`](https://huggingface.co/docs/accelerate/accelerator.html#accelerate.Accelerator) object. [`Accelerator`] will automatically detect your type of distributed setup and initialize all the necessary components for training. You don't need to explicitly place your model on a device.
+Then import and create an [`Accelerator`](https://huggingface.co/docs/accelerate/accelerator.html#accelerate.Accelerator) object. `Accelerator` will automatically detect your type of distributed setup and initialize all the necessary components for training. You don't need to explicitly place your model on a device.
 
 ```py
 >>> from accelerate import Accelerator
@@ -32,7 +32,7 @@ Then import and create an [`Accelerator`](https://huggingface.co/docs/accelerate
 
 ## Prepare to accelerate
 
-The next step is to pass all the relevant training objects to [`prepare`](https://huggingface.co/docs/accelerate/accelerator.html#accelerate.Accelerator.prepare). This includes your training and evaluation DataLoaders, a model and an optimizer:
+The next step is to pass all the relevant training objects to the [`prepare`](https://huggingface.co/docs/accelerate/accelerator.html#accelerate.Accelerator.prepare) method. This includes your training and evaluation DataLoaders, a model and an optimizer:
 
 ```py
 >>> train_dataloader, eval_dataloader, model, optimizer = accelerator.prepare(
@@ -42,7 +42,7 @@ The next step is to pass all the relevant training objects to [`prepare`](https:
 
 ## Backward
 
-The last addition is to replace the typical `loss.backward()` in your training loop with 🤗 Accelerate's [`backward`](https://huggingface.co/docs/accelerate/accelerator.html#accelerate.Accelerator.backward):
+The last addition is to replace the typical `loss.backward()` in your training loop with 🤗 Accelerate's [`backward`](https://huggingface.co/docs/accelerate/accelerator.html#accelerate.Accelerator.backward) method:
 
 ```py
 >>> for epoch in range(num_epochs):
@@ -57,9 +57,49 @@ The last addition is to replace the typical `loss.backward()` in your training l
 ...         progress_bar.update(1)
 ```
 
-As you can see in the following image, you only need to add four additional lines of code to your training loop to enable distributed training!
+As you can see in the following code, you only need to add four additional lines of code to your training loop to enable distributed training!
 
-![accelerate](https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/accelerate.png)
+```diff
++ from accelerate import Accelerator
+  from transformers import AdamW, AutoModelForSequenceClassification, get_scheduler
+
++ accelerator = Accelerator()
+
+  model = AutoModelForSequenceClassification.from_pretrained(checkpoint, num_labels=2)
+  optimizer = AdamW(model.parameters(), lr=3e-5)
+
+- device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
+- model.to(device)
+
++ train_dataloader, eval_dataloader, model, optimizer = accelerator.prepare(
++     train_dataloader, eval_dataloader, model, optimizer
++ )
+
+  num_epochs = 3
+  num_training_steps = num_epochs * len(train_dataloader)
+  lr_scheduler = get_scheduler(
+      "linear",
+      optimizer=optimizer,
+      num_warmup_steps=0,
+      num_training_steps=num_training_steps
+  )
+
+  progress_bar = tqdm(range(num_training_steps))
+
+  model.train()
+  for epoch in range(num_epochs):
+      for batch in train_dataloader:
+-         batch = {k: v.to(device) for k, v in batch.items()}
+          outputs = model(**batch)
+          loss = outputs.loss
+-         loss.backward()
++         accelerator.backward(loss)
+
+          optimizer.step()
+          lr_scheduler.step()
+          optimizer.zero_grad()
+          progress_bar.update(1)
+```
 
 ## Train
 

docs/source/en/add_new_model.mdx

@@ -19,7 +19,7 @@ independently. Thus, for some new models that the community wants to be added to
 model to 🤗 Transformers.
 
 If this sounds like something you would be interested in, feel free to check out the currently open
-“calls-for-model-addition” [here](https://github.com/huggingface/transformers/tree/master/templates/adding_a_new_model/open_model_proposals/README.md)
+“calls-for-model-addition” [here](https://github.com/huggingface/transformers/tree/main/templates/adding_a_new_model/open_model_proposals/README.md)
 and to contact us.
 
 If selected, you will then work closely with one member of the Hugging Face team to integrate the model into 🤗
@@ -95,6 +95,24 @@ different formats - the model to a *pytorch_model.bin* file and the configuratio
 [`~PretrainedConfig.save_pretrained`], so that both model and configuration are saved.
 
 
+### Code style
+
+When coding your new model, keep in mind that Transformers is an opinionated library and we have a few quirks of our
+own regarding how code should be written :-)
+
+1. The forward pass of your model should be fully written in the modeling file while being fully independent of other
+   models in the library. If you want to reuse a block from another model, copy the code and paste it with a
+   `# Copied from` comment on top (see [here](https://github.com/huggingface/transformers/blob/v4.17.0/src/transformers/models/roberta/modeling_roberta.py#L160)
+   for a good example).
+2. The code should be fully understandable, even by a non-native English speaker. This means you should pick
+   descriptive variable names and avoid abbreviations. As an example, `activation` is preferred to `act`.
+   One-letter variable names are strongly discouraged unless it's an index in a for loop. 
+3. More generally we prefer longer explicit code to short magical one.
+4. Avoid subclassing `nn.Sequential` in PyTorch but subclass `nn.Module` and write the forward pass, so that anyone
+   using your code can quickly debug it by adding print statements or breaking points.
+5. Your function signature should be type-annotated. For the rest, good variable names are way more readable and
+   understandable than type annotations.
+
 ### Overview of tokenizers
 
 Not quite ready yet :-( This section will be added soon!
@@ -363,7 +381,7 @@ important. Here is some advice is to make your debugging environment as efficien
   original code so that you can directly input the ids instead of an input string.
 - Make sure that the model in your debugging setup is **not** in training mode, which often causes the model to yield
   random outputs due to multiple dropout layers in the model. Make sure that the forward pass in your debugging
-  environment is **deterministic** so that the dropout layers are not used. Or use *transformers.file_utils.set_seed*
+  environment is **deterministic** so that the dropout layers are not used. Or use *transformers.utils.set_seed*
   if the old and new implementations are in the same framework.
 
 The following section gives you more specific details/tips on how you can do this for *brand_new_bert*.
@@ -380,15 +398,12 @@ In the special case that you are adding a model whose architecture exactly match
 existing model you only have to add a conversion script as described in [this section](#write-a-conversion-script).
 In this case, you can just re-use the whole model architecture of the already existing model.
 
-Otherwise, let's start generating a new model with the amazing Cookiecutter!
+Otherwise, let's start generating a new model. You have two choices here:
 
-**Use the Cookiecutter to automatically generate the model's code**
+- `transformers-cli add-new-model-like` to add a new model like an existing one
+- `transformers-cli add-new-model` to add a new model from our template (will look like BERT or Bart depending on the type of model you select)
 
-To begin with head over to the [🤗 Transformers templates](https://github.com/huggingface/transformers/tree/master/templates/adding_a_new_model) to make use of our
-`cookiecutter` implementation to automatically generate all the relevant files for your model. Again, we recommend
-only adding the PyTorch version of the model at first. Make sure you follow the instructions of the `README.md` on
-the [🤗 Transformers templates](https://github.com/huggingface/transformers/tree/master/templates/adding_a_new_model)
-carefully.
+In both cases, you will be prompted with a questionnaire to fill the basic information of your model. The second command requires to install `cookiecutter`, you can find more information on it [here](https://github.com/huggingface/transformers/tree/main/templates/adding_a_new_model).
 
 **Open a Pull Request on the main huggingface/transformers repo**
 
@@ -398,7 +413,7 @@ side-by-side on integrating the model into 🤗 Transformers.
 
 You should do the following:
 
-1. Create a branch with a descriptive name from your master branch
+1. Create a branch with a descriptive name from your main branch
 
 ```bash
 git checkout -b add_brand_new_bert
@@ -411,11 +426,11 @@ git add .
 git commit
 ```
 
-3. Fetch and rebase to current master
+3. Fetch and rebase to current main
 
 ```bash
 git fetch upstream
-git rebase upstream/master
+git rebase upstream/main
 ```
 
 4. Push the changes to your account using:
@@ -431,12 +446,12 @@ git push -u origin a-descriptive-name-for-my-changes
 6. Change the PR into a draft by clicking on “Convert to draft” on the right of the GitHub pull request web page.
 
 In the following, whenever you have done some progress, don't forget to commit your work and push it to your account so
-that it shows in the pull request. Additionally, you should make sure to update your work with the current master from
+that it shows in the pull request. Additionally, you should make sure to update your work with the current main from
 time to time by doing:
 
 ```bash
 git fetch upstream
-git merge upstream/master
+git merge upstream/main
 ```
 
 In general, all questions you might have regarding the model or your implementation should be asked in your PR and
@@ -494,7 +509,7 @@ slightly adapt it for your use case. Don't hesitate to ask the Hugging Face team
 existing conversion script for your model.
 
 - If you are porting a model from TensorFlow to PyTorch, a good starting point might be BERT's conversion script [here](https://github.com/huggingface/transformers/blob/7acfa95afb8194f8f9c1f4d2c6028224dbed35a2/src/transformers/models/bert/modeling_bert.py#L91)
-- If you are porting a model from PyTorch to PyTorch, a good starting point might be BART's conversion script [here](https://github.com/huggingface/transformers/blob/master/src/transformers/models/bart/convert_bart_original_pytorch_checkpoint_to_pytorch.py)
+- If you are porting a model from PyTorch to PyTorch, a good starting point might be BART's conversion script [here](https://github.com/huggingface/transformers/blob/main/src/transformers/models/bart/convert_bart_original_pytorch_checkpoint_to_pytorch.py)
 
 In the following, we'll quickly explain how PyTorch models store layer weights and define layer names. In PyTorch, the
 name of a layer is defined by the name of the class attribute you give the layer. Let's define a dummy model in
@@ -762,7 +777,7 @@ the community to add some *Tips* to show how the model should be used. Don't hes
 regarding the docstrings.
 
 Next, make sure that the docstring added to `src/transformers/models/brand_new_bert/modeling_brand_new_bert.py` is
-correct and included all necessary inputs and outputs. It is always to good to remind oneself that documentation should
+correct and included all necessary inputs and outputs. We have a detailed guide about writing documentation and our docstring format [here](writing-documentation). It is always to good to remind oneself that documentation should
 be treated at least as carefully as the code in 🤗 Transformers since the documentation is usually the first contact
 point of the community with the model.
 
@@ -793,9 +808,19 @@ You have now finished the coding part, congratulation! 🎉 You are Awesome!
 **12. Upload the models to the model hub**
 
 In this final part, you should convert and upload all checkpoints to the model hub and add a model card for each
-uploaded model checkpoint. You should work alongside the Hugging Face team here to decide on a fitting name for each
+uploaded model checkpoint. You can get familiar with the hub functionalities by reading our [Model sharing and uploading Page](model_sharing). You should work alongside the Hugging Face team here to decide on a fitting name for each
 checkpoint and to get the required access rights to be able to upload the model under the author's organization of
-*brand_new_bert*.
+*brand_new_bert*. The `push_to_hub` method, present in all models in `transformers`, is a quick and efficient way to push your checkpoint to the hub. A little snippet is pasted below:
+
+```python
+brand_new_bert.push_to_hub(
+    repo_path_or_name="brand_new_bert",
+    # Uncomment the following line to push to an organization
+    # organization="<ORGANIZATION>",
+    commit_message="Add model",
+    use_temp_dir=True,
+)
+```
 
 It is worth spending some time to create fitting model cards for each checkpoint. The model cards should highlight the
 specific characteristics of this particular checkpoint, *e.g.* On which dataset was the checkpoint
@@ -809,7 +834,7 @@ fine-tuned on a downstream task. This is not mandatory to merge your PR, but ver
 
 **14. Submit your finished PR**
 
-You're done programming now and can move to the last step, which is getting your PR merged into master. Usually, the
+You're done programming now and can move to the last step, which is getting your PR merged into main. Usually, the
 Hugging Face team should have helped you already at this point, but it is worth taking some time to give your finished
 PR a nice description and eventually add comments to your code, if you want to point out certain design choices to your
 reviewer.

docs/source/en/autoclass_tutorial.mdx

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

docs/source/en/benchmarks.mdx

@@ -12,11 +12,18 @@ specific language governing permissions and limitations under the License.
 
 # Benchmarks
 
+<Tip warning={true}>
+
+Hugging Face's Benchmarking tools are deprecated and it is advised to use external Benchmarking libraries to measure the speed 
+and memory complexity of Transformer models.
+
+</Tip>
+
 [[open-in-colab]]
 
 Let's take a look at how 🤗 Transformers models can be benchmarked, best practices, and already available benchmarks.
 
-A notebook explaining in more detail how to benchmark 🤗 Transformers models can be found [here](https://github.com/huggingface/notebooks/tree/master/examples/benchmark.ipynb).
+A notebook explaining in more detail how to benchmark 🤗 Transformers models can be found [here](https://github.com/huggingface/notebooks/tree/main/examples/benchmark.ipynb).
 
 ## How to benchmark 🤗 Transformers models
 
@@ -32,12 +39,17 @@ backward pass.
 The benchmark classes [`PyTorchBenchmark`] and [`TensorFlowBenchmark`] expect an object of type [`PyTorchBenchmarkArguments`] and
 [`TensorFlowBenchmarkArguments`], respectively, for instantiation. [`PyTorchBenchmarkArguments`] and [`TensorFlowBenchmarkArguments`] are data classes and contain all relevant configurations for their corresponding benchmark class. In the following example, it is shown how a BERT model of type _bert-base-cased_ can be benchmarked.
 
+<frameworkcontent>
+<pt>
 ```py
 >>> from transformers import PyTorchBenchmark, PyTorchBenchmarkArguments
 
 >>> args = PyTorchBenchmarkArguments(models=["bert-base-uncased"], batch_sizes=[8], sequence_lengths=[8, 32, 128, 512])
 >>> benchmark = PyTorchBenchmark(args)
-===PT-TF-SPLIT===
+```
+</pt>
+<tf>
+```py
 >>> from transformers import TensorFlowBenchmark, TensorFlowBenchmarkArguments
 
 >>> args = TensorFlowBenchmarkArguments(
@@ -45,6 +57,8 @@ The benchmark classes [`PyTorchBenchmark`] and [`TensorFlowBenchmark`] expect an
 ... )
 >>> benchmark = TensorFlowBenchmark(args)
 ```
+</tf>
+</frameworkcontent>
 
 Here, three arguments are given to the benchmark argument data classes, namely `models`, `batch_sizes`, and
 `sequence_lengths`. The argument `models` is required and expects a `list` of model identifiers from the
@@ -56,11 +70,10 @@ and `src/transformers/benchmark/benchmark_args_tf.py` (for Tensorflow). Alternat
 commands from root will print out a descriptive list of all configurable parameters for PyTorch and Tensorflow
 respectively.
 
+<frameworkcontent>
+<pt>
 ```bash
 python examples/pytorch/benchmarking/run_benchmark.py --help
-
-===PT-TF-SPLIT===
-python examples/tensorflow/benchmarking/run_benchmark_tf.py --help
 ```
 
 An instantiated benchmark object can then simply be run by calling `benchmark.run()`.
@@ -111,8 +124,18 @@ bert-base-uncased          8              512            1539
 - gpu_power_watts: 280.0
 - gpu_performance_state: 2
 - use_tpu: False
+```
+</pt>
+<tf>
+```bash
+python examples/tensorflow/benchmarking/run_benchmark_tf.py --help
+```
 
-===PT-TF-SPLIT===
+An instantiated benchmark object can then simply be run by calling `benchmark.run()`.
+
+```py
+>>> results = benchmark.run()
+>>> print(results)
 >>> results = benchmark.run()
 >>> print(results)
 ====================       INFERENCE - SPEED - RESULT       ====================
@@ -159,6 +182,8 @@ bert-base-uncased          8              512            1770
 - gpu_performance_state: 2
 - use_tpu: False
 ```
+</tf>
+</frameworkcontent>
 
 By default, the _time_ and the _required memory_ for _inference_ are benchmarked. In the example output above the first
 two sections show the result corresponding to _inference time_ and _inference memory_. In addition, all relevant
@@ -172,6 +197,8 @@ Instead of benchmarking pre-trained models via their model identifier, _e.g._ `b
 alternatively benchmark an arbitrary configuration of any available model class. In this case, a `list` of
 configurations must be inserted with the benchmark args as follows.
 
+<frameworkcontent>
+<pt>
 ```py
 >>> from transformers import PyTorchBenchmark, PyTorchBenchmarkArguments, BertConfig
 
@@ -243,8 +270,10 @@ bert-6-lay                 8              512            1359
 - gpu_power_watts: 280.0
 - gpu_performance_state: 2
 - use_tpu: False
-
-===PT-TF-SPLIT===
+```
+</pt>
+<tf>
+```py
 >>> from transformers import TensorFlowBenchmark, TensorFlowBenchmarkArguments, BertConfig
 
 >>> args = TensorFlowBenchmarkArguments(
@@ -316,6 +345,8 @@ bert-6-lay                 8              512            1540
 - gpu_performance_state: 2
 - use_tpu: False
 ```
+</tf>
+</frameworkcontent>
 
 Again, _inference time_ and _required memory_ for _inference_ are measured, but this time for customized configurations
 of the `BertModel` class. This feature can especially be helpful when deciding for which configuration the model
@@ -348,5 +379,5 @@ available [here](https://docs.google.com/spreadsheets/d/1sryqufw2D0XlUH4sq3e9Wnx
 
 With the new _benchmark_ tools, it is easier than ever to share your benchmark results with the community
 
-- [PyTorch Benchmarking Results](https://github.com/huggingface/transformers/tree/master/examples/pytorch/benchmarking/README.md).
-- [TensorFlow Benchmarking Results](https://github.com/huggingface/transformers/tree/master/examples/tensorflow/benchmarking/README.md).
+- [PyTorch Benchmarking Results](https://github.com/huggingface/transformers/tree/main/examples/pytorch/benchmarking/README.md).
+- [TensorFlow Benchmarking Results](https://github.com/huggingface/transformers/tree/main/examples/tensorflow/benchmarking/README.md).

docs/source/en/bertology.mdx

@@ -32,5 +32,5 @@ help people access the inner representations, mainly adapted from the great work
 - retrieving heads output values and gradients to be able to compute head importance score and prune head as explained
   in https://arxiv.org/abs/1905.10650.
 
-To help you understand and use these features, we have added a specific example script: [bertology.py](https://github.com/huggingface/transformers/tree/master/examples/research_projects/bertology/run_bertology.py) while extract information and prune a model pre-trained on
+To help you understand and use these features, we have added a specific example script: [bertology.py](https://github.com/huggingface/transformers/tree/main/examples/research_projects/bertology/run_bertology.py) while extract information and prune a model pre-trained on
 GLUE.

docs/source/en/community.mdx

@@ -62,3 +62,4 @@ This page regroups resources around 🤗 Transformers developed by the community
 | [Speech Emotion Classification with Wav2Vec2](https://github/m3hrdadfi/soxan/blob/main/notebooks/Emotion_recognition_in_Greek_speech_using_Wav2Vec2.ipynb) | How to leverage a pretrained Wav2Vec2 model for Emotion Classification on the MEGA dataset | [Mehrdad Farahani](https://github.com/m3hrdadfi) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/m3hrdadfi/soxan/blob/main/notebooks/Emotion_recognition_in_Greek_speech_using_Wav2Vec2.ipynb) |
 | [Detect objects in an image with DETR](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/DETR/DETR_minimal_example_(with_DetrFeatureExtractor).ipynb) | How to use a trained *DetrForObjectDetection* model to detect objects in an image and visualize attention | [Niels Rogge](https://github.com/NielsRogge) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/DETR/DETR_minimal_example_(with_DetrFeatureExtractor).ipynb) |
 | [Fine-tune DETR on a custom object detection dataset](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/DETR/Fine_tuning_DetrForObjectDetection_on_custom_dataset_(balloon).ipynb) | How to fine-tune *DetrForObjectDetection* on a custom object detection dataset | [Niels Rogge](https://github.com/NielsRogge) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/DETR/Fine_tuning_DetrForObjectDetection_on_custom_dataset_(balloon).ipynb) |
+| [Finetune T5 for Named Entity Recognition](https://github.com/ToluClassics/Notebooks/blob/main/T5_Ner_Finetuning.ipynb) | How to fine-tune *T5* on a Named Entity Recognition Task | [Ogundepo Odunayo](https://github.com/ToluClassics) | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1obr78FY_cBmWY5ODViCmzdY6O1KB65Vc?usp=sharing) |

docs/source/en/contributing.md

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

docs/source/en/converting_tensorflow_models.mdx

@@ -27,12 +27,12 @@ The documentation below reflects the **transformers-cli convert** command format
 ## BERT
 
 You can convert any TensorFlow checkpoint for BERT (in particular [the pre-trained models released by Google](https://github.com/google-research/bert#pre-trained-models)) in a PyTorch save file by using the
-[convert_bert_original_tf_checkpoint_to_pytorch.py](https://github.com/huggingface/transformers/tree/master/src/transformers/models/bert/convert_bert_original_tf_checkpoint_to_pytorch.py) script.
+[convert_bert_original_tf_checkpoint_to_pytorch.py](https://github.com/huggingface/transformers/tree/main/src/transformers/models/bert/convert_bert_original_tf_checkpoint_to_pytorch.py) script.
 
 This CLI takes as input a TensorFlow checkpoint (three files starting with `bert_model.ckpt`) and the associated
 configuration file (`bert_config.json`), and creates a PyTorch model for this configuration, loads the weights from
 the TensorFlow checkpoint in the PyTorch model and saves the resulting model in a standard PyTorch save file that can
-be imported using `from_pretrained()` (see example in [quicktour](quicktour) , [run_glue.py](https://github.com/huggingface/transformers/tree/master/examples/pytorch/text-classification/run_glue.py) ).
+be imported using `from_pretrained()` (see example in [quicktour](quicktour) , [run_glue.py](https://github.com/huggingface/transformers/tree/main/examples/pytorch/text-classification/run_glue.py) ).
 
 You only need to run this conversion script **once** to get a PyTorch model. You can then disregard the TensorFlow
 checkpoint (the three files starting with `bert_model.ckpt`) but be sure to keep the configuration file (\
@@ -56,7 +56,7 @@ You can download Google's pre-trained models for the conversion [here](https://g
 ## ALBERT
 
 Convert TensorFlow model checkpoints of ALBERT to PyTorch using the
-[convert_albert_original_tf_checkpoint_to_pytorch.py](https://github.com/huggingface/transformers/tree/master/src/transformers/models/albert/convert_albert_original_tf_checkpoint_to_pytorch.py) script.
+[convert_albert_original_tf_checkpoint_to_pytorch.py](https://github.com/huggingface/transformers/tree/main/src/transformers/models/albert/convert_albert_original_tf_checkpoint_to_pytorch.py) script.
 
 The CLI takes as input a TensorFlow checkpoint (three files starting with `model.ckpt-best`) and the accompanying
 configuration file (`albert_config.json`), then creates and saves a PyTorch model. To run this conversion you will

docs/source/en/create_a_model.mdx

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

docs/source/en/custom_models.mdx

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

docs/source/en/debugging.mdx

@@ -12,6 +12,35 @@ specific language governing permissions and limitations under the License.
 
 # Debugging
 
+## Multi-GPU Network Issues Debug
+
+When training or inferencing with `DistributedDataParallel` and multiple GPU, if you run into issue of inter-communication between processes and/or nodes, you can use the following script to diagnose network issues.
+
+```bash
+wget https://raw.githubusercontent.com/huggingface/transformers/main/scripts/distributed/torch-distributed-gpu-test.py
+```
+
+For example to test how 2 GPUs interact do:
+
+```bash
+python -m torch.distributed.run --nproc_per_node 2 --nnodes 1 torch-distributed-gpu-test.py
+```
+If both processes can talk to each and allocate GPU memory each will print an OK status.
+
+For more GPUs or nodes adjust the arguments in the script.
+
+You will find a lot more details inside the diagnostics script and even a recipe to how you could run it in a SLURM environment.
+
+An additional level of debug is to add `NCCL_DEBUG=INFO` environment variable as follows:
+
+```bash
+NCCL_DEBUG=INFO python -m torch.distributed.run --nproc_per_node 2 --nnodes 1 torch-distributed-gpu-test.py
+```
+
+This will dump a lot of NCCL-related debug information, which you can then search online if you find that some problems are reported. Or if you're not sure how to interpret the output you can share the log file in an Issue.
+
+
+
 ## Underflow and Overflow Detection
 
 <Tip>

docs/source/en/fast_tokenizers.mdx

@@ -10,7 +10,7 @@ an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express o
 specific language governing permissions and limitations under the License.
 -->
 
-# Using tokenizers from 🤗 Tokenizers
+# Use tokenizers from 🤗 Tokenizers
 
 The [`PreTrainedTokenizerFast`] depends on the [🤗 Tokenizers](https://huggingface.co/docs/tokenizers) library. The tokenizers obtained from the 🤗 Tokenizers library can be
 loaded very simply into 🤗 Transformers.

docs/source/en/index.mdx

@@ -12,25 +12,18 @@ specific language governing permissions and limitations under the License.
 
 # 🤗 Transformers
 
-State-of-the-art Machine Learning for Jax, Pytorch and TensorFlow
+State-of-the-art Machine Learning for PyTorch, TensorFlow and JAX.
 
-🤗 Transformers (formerly known as _pytorch-transformers_ and _pytorch-pretrained-bert_) provides thousands of pretrained models to perform tasks on different modalities such as text, vision, and audio.
+🤗 Transformers provides APIs to easily download and train state-of-the-art pretrained models. Using pretrained models can reduce your compute costs, carbon footprint, and save you time from training a model from scratch. The models can be used across different modalities such as:
 
-These models can applied on:
+* 📝 Text: text classification, information extraction, question answering, summarization, translation, and text generation in over 100 languages.
+* 🖼️ Images: image classification, object detection, and segmentation.
+* 🗣️ Audio: speech recognition and audio classification.
+* 🐙 Multimodal: table question answering, optical character recognition, information extraction from scanned documents, video classification, and visual question answering.
 
-* 📝 Text, for tasks like text classification, information extraction, question answering, summarization, translation, text generation, in over 100 languages.
-* 🖼️ Images, for tasks like image classification, object detection, and segmentation.
-* 🗣️ Audio, for tasks like speech recognition and audio classification.
+Our library supports seamless integration between three of the most popular deep learning libraries: [PyTorch](https://pytorch.org/), [TensorFlow](https://www.tensorflow.org/) and [JAX](https://jax.readthedocs.io/en/latest/). Train your model in three lines of code in one framework, and load it for inference with another.
 
-Transformer models can also perform tasks on **several modalities combined**, such as table question answering, optical character recognition, information extraction from scanned documents, video classification, and visual question answering.
-
-🤗 Transformers provides APIs to quickly download and use those pretrained models on a given text, fine-tune them on your own datasets and then share them with the community on our [model hub](https://huggingface.co/models). At the same time, each python module defining an architecture is fully standalone and can be modified to enable quick research experiments.
-
-🤗 Transformers is backed by the three most popular deep learning libraries — [Jax](https://jax.readthedocs.io/en/latest/), [PyTorch](https://pytorch.org/) and [TensorFlow](https://www.tensorflow.org/) — with a seamless integration between them. It's straightforward to train your models with one before loading them for inference with the other.
-
-This is the documentation of our repository [transformers](https://github.com/huggingface/transformers). You can
-also follow our [online course](https://huggingface.co/course) that teaches how to use this library, as well as the
-other libraries developed by Hugging Face and the Hub.
+Each 🤗 Transformers architecture is defined in a standalone Python module so they can be easily customized for research and experiments.
 
 ## If you are looking for custom support from the Hugging Face team
 
@@ -38,60 +31,28 @@ other libraries developed by Hugging Face and the Hub.
 <img alt="HuggingFace Expert Acceleration Program" src="https://huggingface.co/front/thumbnails/support.png" style="max-width: 600px; border: 1px solid #eee; border-radius: 4px; box-shadow: 0 1px 2px 0 rgba(0, 0, 0, 0.05);">
 </a><br>
 
-## Features
-
-1. Easy-to-use state-of-the-art models:
-    - High performance on natural language understanding & generation, computer vision, and audio tasks.
-    - Low barrier to entry for educators and practitioners.
-    - Few user-facing abstractions with just three classes to learn.
-    - A unified API for using all our pretrained models.
-
-1. Lower compute costs, smaller carbon footprint:
-    - Researchers can share trained models instead of always retraining.
-    - Practitioners can reduce compute time and production costs.
-    - Dozens of architectures with over 20,000 pretrained models, some in more than 100 languages.
-
-1. Choose the right framework for every part of a model's lifetime:
-    - Train state-of-the-art models in 3 lines of code.
-    - Move a single model between TF2.0/PyTorch/JAX frameworks at will.
-    - Seamlessly pick the right framework for training, evaluation and production.
-
-1. Easily customize a model or an example to your needs:
-    - We provide examples for each architecture to reproduce the results published by its original authors.
-    - Model internals are exposed as consistently as possible.
-    - Model files can be used independently of the library for quick experiments.
-
-[All the model checkpoints](https://huggingface.co/models) are seamlessly integrated from the huggingface.co [model
-hub](https://huggingface.co) where they are uploaded directly by [users](https://huggingface.co/users) and
-[organizations](https://huggingface.co/organizations).
-
-Current number of checkpoints: <img src="https://img.shields.io/endpoint?url=https://huggingface.co/api/shields/models&color=brightgreen">
-
 ## Contents
 
 The documentation is organized in five parts:
 
-- **GET STARTED** contains a quick tour, the installation instructions and some useful information about our philosophy
-  and a glossary.
-- **USING 🤗 TRANSFORMERS** contains general tutorials on how to use the library.
-- **ADVANCED GUIDES** contains more advanced guides that are more specific to a given script or part of the library.
-- **RESEARCH** focuses on tutorials that have less to do with how to use the library but more about general research in
-  transformers model
-- **API** contains the documentation of each public class and function, grouped in:
+- **GET STARTED** contains a quick tour and installation instructions to get up and running with 🤗 Transformers.
+- **TUTORIALS** are a great place to begin if you are new to our library. This section will help you gain the basic skills you need to start using 🤗 Transformers.
+- **HOW-TO GUIDES** will show you how to achieve a specific goal like fine-tuning a pretrained model for language modeling or how to create a custom model head.
+- **CONCEPTUAL GUIDES** provides more discussion and explanation of the underlying concepts and ideas behind models, tasks, and the design philosophy of 🤗 Transformers. 
+- **API** describes each class and function, grouped in:
 
   - **MAIN CLASSES** for the main classes exposing the important APIs of the library.
   - **MODELS** for the classes and functions related to each model implemented in the library.
   - **INTERNAL HELPERS** for the classes and functions we use internally.
 
-The library currently contains Jax, PyTorch and Tensorflow implementations, pretrained model weights, usage scripts and
-conversion utilities for the following models.
+The library currently contains JAX, PyTorch and TensorFlow implementations, pretrained model weights, usage scripts and conversion utilities for the following models.
 
 ### Supported models
 
 <!--This list is updated automatically from the README with _make fix-copies_. Do not update manually! -->
 
 1. **[ALBERT](model_doc/albert)** (from Google Research and the Toyota Technological Institute at Chicago) released with the paper [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://arxiv.org/abs/1909.11942), by Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut.
-1. **[BART](model_doc/bart)** (from Facebook) released with the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/pdf/1910.13461.pdf) by Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Ves Stoyanov and Luke Zettlemoyer.
+1. **[BART](model_doc/bart)** (from Facebook) released with the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/abs/1910.13461) by Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Ves Stoyanov and Luke Zettlemoyer.
 1. **[BARThez](model_doc/barthez)** (from École polytechnique) released with the paper [BARThez: a Skilled Pretrained French Sequence-to-Sequence Model](https://arxiv.org/abs/2010.12321) by Moussa Kamal Eddine, Antoine J.-P. Tixier, Michalis Vazirgiannis.
 1. **[BARTpho](model_doc/bartpho)** (from VinAI Research) released with the paper [BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese](https://arxiv.org/abs/2109.09701) by Nguyen Luong Tran, Duong Minh Le and Dat Quoc Nguyen.
 1. **[BEiT](model_doc/beit)** (from Microsoft) released with the paper [BEiT: BERT Pre-Training of Image Transformers](https://arxiv.org/abs/2106.08254) by Hangbo Bao, Li Dong, Furu Wei.
@@ -106,22 +67,28 @@ conversion utilities for the following models.
 1. **[ByT5](model_doc/byt5)** (from Google Research) released with the paper [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) by Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel.
 1. **[CamemBERT](model_doc/camembert)** (from Inria/Facebook/Sorbonne) released with the paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) by Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suárez*, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah and Benoît Sagot.
 1. **[CANINE](model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
+1. **[ConvNeXT](model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
 1. **[CLIP](model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
 1. **[ConvBERT](model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
 1. **[CPM](model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
 1. **[CTRL](model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
+1. **[Data2Vec](model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
 1. **[DeBERTa](model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
 1. **[DeBERTa-v2](model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
+1. **[Decision Transformer](model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
+1. **[DiT](model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
 1. **[DeiT](model_doc/deit)** (from Facebook) released with the paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) by Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou.
 1. **[DETR](model_doc/detr)** (from Facebook) released with the paper [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) by Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko.
 1. **[DialoGPT](model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
-1. **[DistilBERT](model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/master/examples/research_projects/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/master/examples/research_projects/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/master/examples/research_projects/distillation) and a German version of DistilBERT.
+1. **[DistilBERT](model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation) and a German version of DistilBERT.
 1. **[DPR](model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
+1. **[DPT](master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by René Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
 1. **[EncoderDecoder](model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
 1. **[ELECTRA](model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
 1. **[FlauBERT](model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab.
 1. **[FNet](model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
 1. **[Funnel Transformer](model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
+1. **[GLPN](model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
 1. **[GPT](model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
 1. **[GPT-2](model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
 1. **[GPT-J](model_doc/gptj)** (from EleutherAI) released in the repository [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
@@ -139,6 +106,7 @@ conversion utilities for the following models.
 1. **[LXMERT](model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
 1. **[M2M100](model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
 1. **[MarianMT](model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jörg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
+1. **[MaskFormer](model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
 1. **[MBart](model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
 1. **[MBart-50](model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
 1. **[Megatron-BERT](model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
@@ -149,13 +117,16 @@ conversion utilities for the following models.
 1. **[Pegasus](model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
 1. **[Perceiver IO](model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira.
 1. **[PhoBERT](model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
+1. **[PLBart](model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
+1. **[PoolFormer](model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
 1. **[ProphetNet](model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
 1. **[QDQBert](model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
-1. **[REALM](https://huggingface.co/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
+1. **[REALM](model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
 1. **[Reformer](model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
-1. **[RemBERT](model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/pdf/2010.12821.pdf) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
-1. **[RoBERTa](model_doc/roberta)** (from Facebook), released together with the paper a [Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
-1. **[RoFormer](model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper a [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/pdf/2104.09864v1.pdf) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
+1. **[RemBERT](model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
+1. **[ResNet](model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun.
+1. **[RoBERTa](model_doc/roberta)** (from Facebook), released together with the paper [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
+1. **[RoFormer](model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/abs/2104.09864) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
 1. **[SegFormer](model_doc/segformer)** (from NVIDIA) released with the paper [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) by Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo.
 1. **[SEW](model_doc/sew)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
 1. **[SEW-D](model_doc/sew_d)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
@@ -171,19 +142,22 @@ conversion utilities for the following models.
 1. **[TrOCR](model_doc/trocr)** (from Microsoft), released together with the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
 1. **[UniSpeech](model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
 1. **[UniSpeechSat](model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
-1. **[ViLT)](model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
+1. **[VAN](model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/abs/2202.09741) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
+1. **[ViLT](model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
 1. **[Vision Transformer (ViT)](model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby.
-1. **[ViTMAE)](model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
+1. **[ViTMAE](model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
 1. **[VisualBERT](model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
 1. **[WavLM](model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
 1. **[Wav2Vec2](model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
-1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/master/transformers/model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
+1. **[Wav2Vec2Phoneme](model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
+1. **[XGLM](model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li.
 1. **[XLM](model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
 1. **[XLM-ProphetNet](model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
 1. **[XLM-RoBERTa](model_doc/xlm-roberta)** (from Facebook AI), released together with the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov.
+1. **[XLM-RoBERTa-XL](model_doc/xlm-roberta-xl)** (from Facebook AI), released together with the paper [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) by Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau.
 1. **[XLNet](model_doc/xlnet)** (from Google/CMU) released with the paper [​XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
 1. **[XLSR-Wav2Vec2](model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
-1. **[XLS-R](https://huggingface.co/docs/master/transformers/model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
+1. **[XLS-R](model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
 1. **[YOSO](model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) Once: Linear Cost Self-Attention Via Bernoulli Sampling](https://arxiv.org/abs/2111.09714) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
 
 
@@ -210,21 +184,27 @@ Flax), PyTorch, and/or TensorFlow.
 |           Canine            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            CLIP             |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |          ConvBERT           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|          ConvNext           |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
 |            CTRL             |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
+|        Data2VecAudio        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|        Data2VecText         |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           DeBERTa           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
 |         DeBERTa-v2          |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
+|    Decision Transformer     |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            DeiT             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            DETR             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |         DistilBERT          |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |             DPR             |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|             DPT             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           ELECTRA           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |       Encoder decoder       |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
 | FairSeq Machine-Translation |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |          FlauBERT           |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
 |            FNet             |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
 |     Funnel Transformer      |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|            GLPN             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           GPT Neo           |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
-|            GPT-J            |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
+|            GPT-J            |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
 |           Hubert            |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
 |           I-BERT            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |          ImageGPT           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
@@ -236,6 +216,7 @@ Flax), PyTorch, and/or TensorFlow.
 |           LXMERT            |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
 |           M2M100            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           Marian            |       ✅       |       ❌       |       ✅        |         ✅         |      ✅      |
+|         MaskFormer          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            mBART            |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |        MegatronBert         |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |         MobileBERT          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
@@ -246,20 +227,23 @@ Flax), PyTorch, and/or TensorFlow.
 |        OpenAI GPT-2         |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |           Pegasus           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |          Perceiver          |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           PLBart            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
+|         PoolFormer          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |         ProphetNet          |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           QDQBert           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |             RAG             |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
 |            Realm            |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
 |          Reformer           |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
 |           RemBERT           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|           ResNet            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |          RetriBERT          |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
 |           RoBERTa           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |          RoFormer           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |          SegFormer          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |             SEW             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            SEW-D            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|   Speech Encoder decoder    |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|         Speech2Text         |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
+|   Speech Encoder decoder    |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
+|         Speech2Text         |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
 |        Speech2Text2         |       ✅       |       ❌       |       ❌        |         ❌         |      ❌      |
 |          Splinter           |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
 |         SqueezeBERT         |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
@@ -270,16 +254,19 @@ Flax), PyTorch, and/or TensorFlow.
 |            TrOCR            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |          UniSpeech          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |        UniSpeechSat         |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|             VAN             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            ViLT             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |   Vision Encoder decoder    |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
 |    VisionTextDualEncoder    |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
 |         VisualBert          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |             ViT             |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
-|           ViTMAE            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           ViTMAE            |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
 |          Wav2Vec2           |       ✅       |       ❌       |       ✅        |         ✅         |      ✅      |
 |            WavLM            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            XGLM             |       ✅       |       ✅       |       ✅        |         ❌         |      ✅      |
 |             XLM             |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
-|         XLM-RoBERTa         |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|         XLM-RoBERTa         |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|       XLM-RoBERTa-XL        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |        XLMProphetNet        |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            XLNet            |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
 |            YOSO             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |

docs/source/en/installation.mdx

@@ -0,0 +1,235 @@
+<!---
+Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+-->
+
+# Installation
+
+Install 🤗 Transformers for whichever deep learning library you're working with, setup your cache, and optionally configure 🤗 Transformers to run offline.
+
+🤗 Transformers is tested on Python 3.6+, PyTorch 1.1.0+, TensorFlow 2.0+, and Flax. Follow the installation instructions below for the deep learning library you are using:
+
+* [PyTorch](https://pytorch.org/get-started/locally/) installation instructions.
+* [TensorFlow 2.0](https://www.tensorflow.org/install/pip) installation instructions.
+* [Flax](https://flax.readthedocs.io/en/latest/) installation instructions.
+
+## Install with pip
+
+You should install 🤗 Transformers in a [virtual environment](https://docs.python.org/3/library/venv.html). If you're unfamiliar with Python virtual environments, take a look at this [guide](https://packaging.python.org/guides/installing-using-pip-and-virtual-environments/). A virtual environment makes it easier to manage different projects, and avoid compatibility issues between dependencies.
+
+Start by creating a virtual environment in your project directory:
+
+```bash
+python -m venv .env
+```
+
+Activate the virtual environment:
+
+```bash
+source .env/bin/activate
+```
+
+Now you're ready to install 🤗 Transformers with the following command:
+
+```bash
+pip install transformers
+```
+
+For CPU-support only, you can conveniently install 🤗 Transformers and a deep learning library in one line. For example, install 🤗 Transformers and PyTorch with:
+
+```bash
+pip install transformers[torch]
+```
+
+🤗 Transformers and TensorFlow 2.0:
+
+```bash
+pip install transformers[tf-cpu]
+```
+
+🤗 Transformers and Flax:
+
+```bash
+pip install transformers[flax]
+```
+
+Finally, check if 🤗 Transformers has been properly installed by running the following command. It will download a pretrained model:
+
+```bash
+python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('we love you'))"
+```
+
+Then print out the label and score:
+
+```bash
+[{'label': 'POSITIVE', 'score': 0.9998704791069031}]
+```
+
+## Install from source
+
+Install 🤗 Transformers from source with the following command:
+
+```bash
+pip install git+https://github.com/huggingface/transformers
+```
+
+This command installs the bleeding edge `main` version rather than the latest `stable` version. The `main` version is useful for staying up-to-date with the latest developments. For instance, if a bug has been fixed since the last official release but a new release hasn't been rolled out yet. However, this means the `main` version may not always be stable. We strive to keep the `main` version operational, and most issues are usually resolved within a few hours or a day. If you run into a problem, please open an [Issue](https://github.com/huggingface/transformers/issues) so we can fix it even sooner!
+
+Check if 🤗 Transformers has been properly installed by running the following command:
+
+```bash
+python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('I love you'))"
+```
+
+## Editable install
+
+You will need an editable install if you'd like to:
+
+* Use the `main` version of the source code.
+* Contribute to 🤗 Transformers and need to test changes in the code.
+
+Clone the repository and install 🤗 Transformers with the following commands:
+
+```bash
+git clone https://github.com/huggingface/transformers.git
+cd transformers
+pip install -e .
+```
+
+These commands will link the folder you cloned the repository to and your Python library paths. Python will now look inside the folder you cloned to in addition to the normal library paths. For example, if your Python packages are typically installed in `~/anaconda3/envs/main/lib/python3.7/site-packages/`, Python will also search the folder you cloned to: `~/transformers/`.
+
+<Tip warning={true}>
+
+You must keep the `transformers` folder if you want to keep using the library.
+
+</Tip>
+
+Now you can easily update your clone to the latest version of 🤗 Transformers with the following command:
+
+```bash
+cd ~/transformers/
+git pull
+```
+
+Your Python environment will find the `main` version of 🤗 Transformers on the next run.
+
+## Install with conda
+
+Install from the conda channel `huggingface`:
+
+```bash
+conda install -c huggingface transformers
+```
+
+## Cache setup
+
+Pretrained models are downloaded and locally cached at: `~/.cache/huggingface/transformers/`. This is the default directory given by the shell environment variable `TRANSFORMERS_CACHE`. On Windows, the default directory is given by `C:\Users\username\.cache\huggingface\transformers`. You can change the shell environment variables shown below - in order of priority - to specify a different cache directory:
+
+1. Shell environment variable (default): `TRANSFORMERS_CACHE`.
+2. Shell environment variable: `HF_HOME` + `transformers/`.
+3. Shell environment variable: `XDG_CACHE_HOME` + `/huggingface/transformers`.
+
+<Tip>
+
+🤗 Transformers will use the shell environment variables `PYTORCH_TRANSFORMERS_CACHE` or `PYTORCH_PRETRAINED_BERT_CACHE` if you are coming from an earlier iteration of this library and have set those environment variables, unless you specify the shell environment variable `TRANSFORMERS_CACHE`.
+
+</Tip>
+
+## Offline mode
+
+🤗 Transformers is able to run in a firewalled or offline environment by only using local files. Set the environment variable `TRANSFORMERS_OFFLINE=1` to enable this behavior.
+
+<Tip>
+
+Add [🤗 Datasets](https://huggingface.co/docs/datasets/) to your offline training workflow by setting the environment variable `HF_DATASETS_OFFLINE=1`.
+
+</Tip>
+
+For example, you would typically run a program on a normal network firewalled to external instances with the following command:
+
+```bash
+python examples/pytorch/translation/run_translation.py --model_name_or_path t5-small --dataset_name wmt16 --dataset_config ro-en ...
+```
+
+Run this same program in an offline instance with:
+
+```bash
+HF_DATASETS_OFFLINE=1 TRANSFORMERS_OFFLINE=1 \
+python examples/pytorch/translation/run_translation.py --model_name_or_path t5-small --dataset_name wmt16 --dataset_config ro-en ...
+```
+
+The script should now run without hanging or waiting to timeout because it knows it should only look for local files.
+
+### Fetch models and tokenizers to use offline
+
+Another option for using 🤗 Transformers offline is to download the files ahead of time, and then point to their local path when you need to use them offline. There are three ways to do this:
+
+* Download a file through the user interface on the [Model Hub](https://huggingface.co/models) by clicking on the ↓ icon.
+
+    ![download-icon](https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/download-icon.png)
+
+* Use the [`PreTrainedModel.from_pretrained`] and [`PreTrainedModel.save_pretrained`] workflow:
+
+    1. Download your files ahead of time with [`PreTrainedModel.from_pretrained`]:
+
+    ```py
+    >>> from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+
+    >>> tokenizer = AutoTokenizer.from_pretrained("bigscience/T0_3B")
+    >>> model = AutoModelForSeq2SeqLM.from_pretrained("bigscience/T0_3B")
+    ```
+
+    2. Save your files to a specified directory with [`PreTrainedModel.save_pretrained`]:
+
+    ```py
+    >>> tokenizer.save_pretrained("./your/path/bigscience_t0")
+    >>> model.save_pretrained("./your/path/bigscience_t0")
+    ```
+
+    3. Now when you're offline, reload your files with [`PreTrainedModel.from_pretrained`] from the specified directory:
+
+    ```py
+    >>> tokenizer = AutoTokenizer.from_pretrained("./your/path/bigscience_t0")
+    >>> model = AutoModel.from_pretrained("./your/path/bigscience_t0")
+    ```
+
+* Programmatically download files with the [huggingface_hub](https://github.com/huggingface/huggingface_hub/tree/main/src/huggingface_hub) library:
+
+    1. Install the `huggingface_hub` library in your virtual environment:
+
+    ```bash
+    python -m pip install huggingface_hub
+    ```
+
+    2. Use the [`hf_hub_download`](https://huggingface.co/docs/hub/adding-a-library#download-files-from-the-hub) function to download a file to a specific path. For example, the following command downloads the `config.json` file from the [T0](https://huggingface.co/bigscience/T0_3B) model to your desired path:
+
+    ```py
+    >>> from huggingface_hub import hf_hub_download
+
+    >>> hf_hub_download(repo_id="bigscience/T0_3B", filename="config.json", cache_dir="./your/path/bigscience_t0")
+    ```
+
+Once your file is downloaded and locally cached, specify it's local path to load and use it:
+
+```py
+>>> from transformers import AutoConfig
+
+>>> config = AutoConfig.from_pretrained("./your/path/bigscience_t0/config.json")
+```
+
+<Tip>
+
+See the [How to download files from the Hub](https://huggingface.co/docs/hub/how-to-downstream) section for more details on downloading files stored on the Hub.
+
+</Tip>

docs/source/en/internal/file_utils.mdx

@@ -12,35 +12,35 @@ specific language governing permissions and limitations under the License.
 
 # General Utilities
 
-This page lists all of Transformers general utility functions that are found in the file `file_utils.py`.
+This page lists all of Transformers general utility functions that are found in the file `utils.py`.
 
 Most of those are only useful if you are studying the general code in the library.
 
 
 ## Enums and namedtuples
 
-[[autodoc]] file_utils.ExplicitEnum
+[[autodoc]] utils.ExplicitEnum
 
-[[autodoc]] file_utils.PaddingStrategy
+[[autodoc]] utils.PaddingStrategy
 
-[[autodoc]] file_utils.TensorType
+[[autodoc]] utils.TensorType
 
 ## Special Decorators
 
-[[autodoc]] file_utils.add_start_docstrings
+[[autodoc]] utils.add_start_docstrings
 
-[[autodoc]] file_utils.add_start_docstrings_to_model_forward
+[[autodoc]] utils.add_start_docstrings_to_model_forward
 
-[[autodoc]] file_utils.add_end_docstrings
+[[autodoc]] utils.add_end_docstrings
 
-[[autodoc]] file_utils.add_code_sample_docstrings
+[[autodoc]] utils.add_code_sample_docstrings
 
-[[autodoc]] file_utils.replace_return_docstrings
+[[autodoc]] utils.replace_return_docstrings
 
 ## Special Properties
 
-[[autodoc]] file_utils.cached_property
+[[autodoc]] utils.cached_property
 
 ## Other Utilities
 
-[[autodoc]] file_utils._LazyModule
+[[autodoc]] utils._LazyModule

docs/source/en/internal/generation_utils.mdx

@@ -16,15 +16,16 @@ This page lists all the utility functions used by [`~generation_utils.Generation
 [`~generation_utils.GenerationMixin.greedy_search`],
 [`~generation_utils.GenerationMixin.sample`],
 [`~generation_utils.GenerationMixin.beam_search`],
-[`~generation_utils.GenerationMixin.beam_sample`], and
-[`~generation_utils.GenerationMixin.group_beam_search`].
+[`~generation_utils.GenerationMixin.beam_sample`],
+[`~generation_utils.GenerationMixin.group_beam_search`], and
+[`~generation_utils.GenerationMixin.constrained_beam_search`].
 
 Most of those are only useful if you are studying the code of the generate methods in the library.
 
 ## Generate Outputs
 
 The output of [`~generation_utils.GenerationMixin.generate`] is an instance of a subclass of
-[`~file_utils.ModelOutput`]. This output is a data structure containing all the information returned
+[`~utils.ModelOutput`]. This output is a data structure containing all the information returned
 by [`~generation_utils.GenerationMixin.generate`], but that can also be used as tuple or dictionary.
 
 Here's an example:
@@ -147,6 +148,36 @@ generation.
 [[autodoc]] InfNanRemoveLogitsProcessor
     - __call__
 
+[[autodoc]] TFLogitsProcessor
+    - __call__
+
+[[autodoc]] TFLogitsProcessorList
+    - __call__
+
+[[autodoc]] TFLogitsWarper
+    - __call__
+
+[[autodoc]] TFTemperatureLogitsWarper
+    - __call__
+
+[[autodoc]] TFTopPLogitsWarper
+    - __call__
+
+[[autodoc]] TFTopKLogitsWarper
+    - __call__
+
+[[autodoc]] TFMinLengthLogitsProcessor
+    - __call__
+
+[[autodoc]] TFNoBadWordsLogitsProcessor
+    - __call__
+
+[[autodoc]] TFNoRepeatNGramLogitsProcessor
+    - __call__
+
+[[autodoc]] TFRepetitionPenaltyLogitsProcessor
+    - __call__
+
 [[autodoc]] FlaxLogitsProcessor
     - __call__
 
@@ -190,6 +221,18 @@ A [`StoppingCriteria`] can be used to change when to stop generation (other than
 [[autodoc]] MaxTimeCriteria
     - __call__
 
+## Constraints
+
+A [`Constraint`] can be used to force the generation to include specific tokens or sequences in the output.
+
+[[autodoc]] Constraint
+
+[[autodoc]] PhrasalConstraint
+
+[[autodoc]] DisjunctiveConstraint
+
+[[autodoc]] ConstraintListState
+
 ## BeamSearch
 
 [[autodoc]] BeamScorer
@@ -200,6 +243,10 @@ A [`StoppingCriteria`] can be used to change when to stop generation (other than
     - process
     - finalize
 
+[[autodoc]] ConstrainedBeamSearchScorer
+    - process
+    - finalize
+
 ## Utilities
 
 [[autodoc]] top_k_top_p_filtering

docs/source/en/main_classes/callback.mdx

@@ -34,6 +34,8 @@ By default a [`Trainer`] will use the following callbacks:
 - [`~integrations.MLflowCallback`] if [mlflow](https://www.mlflow.org/) is installed.
 - [`~integrations.AzureMLCallback`] if [azureml-sdk](https://pypi.org/project/azureml-sdk/) is
   installed.
+- [`~integrations.CodeCarbonCallback`] if [codecarbon](https://pypi.org/project/codecarbon/) is
+  installed.
 
 The main class that implements callbacks is [`TrainerCallback`]. It gets the
 [`TrainingArguments`] used to instantiate the [`Trainer`], can access that
@@ -66,6 +68,8 @@ Here is the list of the available [`TrainerCallback`] in the library:
 
 [[autodoc]] integrations.AzureMLCallback
 
+[[autodoc]] integrations.CodeCarbonCallback
+
 ## TrainerCallback
 
 [[autodoc]] TrainerCallback

docs/source/en/main_classes/deepspeed.mdx

@@ -367,7 +367,7 @@ cat <<'EOT' > ds_config_zero3.json
         "stage3_param_persistence_threshold": "auto",
         "stage3_max_live_parameters": 1e9,
         "stage3_max_reuse_distance": 1e9,
-        "stage3_gather_fp16_weights_on_model_save": true
+        "stage3_gather_16bit_weights_on_model_save": true
     },
 
     "gradient_accumulation_steps": "auto",
@@ -613,6 +613,17 @@ The following is an example of configuration for ZeRO stage 2:
   the slower the communication gets, and the more GPU RAM will be available to other tasks. So if a bigger batch size is
   important, getting a slightly slower training time could be a good trade.
 
+Additionally, `deepspeed==0.4.4` added a new option `round_robin_gradients` which you can enable with:
+
+```json
+{
+    "zero_optimization": {
+        "round_robin_gradients": true
+    }
+}
+```
+
+This is a stage 2 optimization for CPU offloading that parallelizes gradient copying to CPU memory among ranks by fine-grained gradient partitioning. Performance benefit grows with gradient accumulation steps (more copying between optimizer steps) or GPU count (increased parallelism).
 
 
 <a id='deepspeed-zero3-config'></a>
@@ -641,7 +652,7 @@ The following is an example of configuration for ZeRO stage 3:
         "stage3_param_persistence_threshold": "auto",
         "stage3_max_live_parameters": 1e9,
         "stage3_max_reuse_distance": 1e9,
-        "stage3_gather_fp16_weights_on_model_save": true
+        "stage3_gather_16bit_weights_on_model_save": true
     }
 }
 ```
@@ -680,7 +691,7 @@ The following configuration values depend on the model's hidden size:
 therefore set these values to `auto` and the [`Trainer`] will automatically assign the recommended
 values. But, of course, feel free to set these explicitly as well.
 
-`stage3_gather_fp16_weights_on_model_save` enables model fp16 weights consolidation when model gets saved. With large
+`stage3_gather_16bit_weights_on_model_save` enables model fp16 weights consolidation when model gets saved. With large
 models and multiple GPUs this is an expensive operation both in terms of memory and speed. It's currently required if
 you plan to resume the training. Watch out for future updates that will remove this limitation and make things more
 flexible.
@@ -733,14 +744,14 @@ The following configuration example enables NVMe to offload both optimizer state
             "buffer_count": 5,
             "buffer_size": 1e8,
             "max_in_cpu": 1e9
-        }
+        },
         "aio": {
             "block_size": 262144,
             "queue_depth": 32,
             "thread_count": 1,
             "single_submit": false,
             "overlap_events": true
-        }
+        },
         "overlap_comm": true,
         "contiguous_gradients": true,
         "sub_group_size": 1e9,
@@ -749,7 +760,7 @@ The following configuration example enables NVMe to offload both optimizer state
         "stage3_param_persistence_threshold": "auto",
         "stage3_max_live_parameters": 1e9,
         "stage3_max_reuse_distance": 1e9,
-        "stage3_gather_fp16_weights_on_model_save": true
+        "stage3_gather_16bit_weights_on_model_save": true
     },
 }
 ```
@@ -955,7 +966,7 @@ Here is a full ZeRO-3 auto-configuration file `ds_config_zero3.json`:
         "stage3_param_persistence_threshold": "auto",
         "stage3_max_live_parameters": 1e9,
         "stage3_max_reuse_distance": 1e9,
-        "stage3_gather_fp16_weights_on_model_save": true
+        "stage3_gather_16bit_weights_on_model_save": true
     },
 
     "gradient_accumulation_steps": "auto",
@@ -1018,7 +1029,7 @@ values look like, but we highly recommend using the one with multiple `auto` set
         "stage3_param_persistence_threshold": 1e4,
         "stage3_max_live_parameters": 1e9,
         "stage3_max_reuse_distance": 1e9,
-        "stage3_gather_fp16_weights_on_model_save": true
+        "stage3_gather_16bit_weights_on_model_save": true
     },
 
     "steps_per_print": 2000,
@@ -1221,6 +1232,7 @@ the much more efficient tf32 format for some operations, but the results will st
 benchmarks, please, see [TensorFloat-32(TF32) on Ampere devices](https://pytorch.org/docs/stable/notes/cuda.html#tensorfloat-32-tf32-on-ampere-devices). The document includes
 instructions on how to disable this automatic conversion if for some reason you prefer not to use it.
 
+With the 🤗 Trainer you can use `--tf32` to enable it, or disable it with `--tf32 0` or `--no_tf32`. By default the PyTorch default is used.
 
 
 
@@ -1230,7 +1242,9 @@ instructions on how to disable this automatic conversion if for some reason you
 
 You can use automatic mixed precision with either a pytorch-like AMP way or the apex-like way:
 
-To configure pytorch AMP-like mode set:
+### fp16
+
+To configure pytorch AMP-like mode with fp16 (float16) set:
 
 ```json
 {
@@ -1248,7 +1262,7 @@ To configure pytorch AMP-like mode set:
 and the [`Trainer`] will automatically enable or disable it based on the value of
 `args.fp16_backend`. The rest of config values are up to you.
 
-This mode gets enabled when `--fp16 --fp16_backend amp` command line args are passed.
+This mode gets enabled when `--fp16 --fp16_backend amp` or `--fp16_full_eval` command line args are passed.
 
 You can also enable/disable this mode explicitly:
 
@@ -1270,6 +1284,43 @@ configuration.
 
 Here is the [documentation](https://www.deepspeed.ai/docs/config-json/#fp16-training-options).
 
+### bf16
+
+If bf16 (bfloat16) is desired instead of fp16 then the following configuration section is to be used:
+
+```json
+{
+    "bf16": {
+        "enabled": "auto"
+    }
+}
+```
+
+bf16 has the same dynamic range as fp32 and thus doesn't require loss scaling.
+
+This mode gets enabled when `--bf16` or `--bf16_full_eval` command line args are passed.
+
+You can also enable/disable this mode explicitly:
+
+```json
+{
+    "bf16": {
+        "enabled": true
+    }
+}
+```
+
+<Tip>
+
+As of `deepspeed==0.6.0` the bf16 support is new and experimental.
+
+If you use [gradient accumulation](#gradient-accumulation) with bf16-enabled, you need to be aware that it'll accumulate gradients in bf16, which may not be what you want due to this format's low precision, as it may lead to a lossy accumulation.
+
+</Tip>
+
+
+### apex
+
 To configure apex AMP-like mode set:
 
 ```json
@@ -1400,15 +1451,14 @@ When a model is saved under ZeRO-2, you end up having the normal `pytorch_model.
 they are only the fp16 version of the weights.
 
 Under ZeRO-3, things are much more complicated, since the model weights are partitioned out over multiple GPUs,
-therefore `"stage3_gather_fp16_weights_on_model_save": true` is required to get the `Trainer` to save the fp16
-version of the weights. If this setting is `False` ``pytorch_model.bin` won't be created. This is because by default DeepSpeed's `state_dict` contains a placeholder and not the real weights. If we were to save this `state_dict`` it
-won't be possible to load it back.
+therefore `"stage3_gather_16bit_weights_on_model_save": true` is required to get the `Trainer` to save the fp16
+version of the weights. If this setting is `False` `pytorch_model.bin` won't be created. This is because by default DeepSpeed's `state_dict` contains a placeholder and not the real weights. If we were to save this `state_dict` it won't be possible to load it back.
 
 
 ```json
 {
     "zero_optimization": {
-        "stage3_gather_fp16_weights_on_model_save": true
+        "stage3_gather_16bit_weights_on_model_save": true
     }
 }
 ```
@@ -1623,12 +1673,68 @@ deepspeed examples/pytorch/translation/run_translation.py \
 Since for inference there is no need for additional large memory used by the optimizer states and the gradients you
 should be able to fit much larger batches and/or sequence length onto the same hardware.
 
-
 Additionally DeepSpeed is currently developing a related product called Deepspeed-Inference which has no relationship
 to the ZeRO technology, but instead uses tensor parallelism to scale models that can't fit onto a single GPU. This is a
 work in progress and we will provide the integration once that product is complete.
 
 
+### Memory Requirements
+
+Since Deepspeed ZeRO can offload memory to CPU (and NVMe) the framework provides utils that allow one to tell how much CPU and GPU memory will be needed depending on the number of GPUs being used.
+
+Let's estimate how much memory is needed to finetune "bigscience/T0_3B" on a single GPU:
+
+```bash
+$ python -c 'from transformers import AutoModel; \
+from deepspeed.runtime.zero.stage3 import estimate_zero3_model_states_mem_needs_all_live; \
+model = AutoModel.from_pretrained("bigscience/T0_3B"); \
+estimate_zero3_model_states_mem_needs_all_live(model, num_gpus_per_node=1, num_nodes=1)'
+[...]
+Estimated memory needed for params, optim states and gradients for a:
+HW: Setup with 1 node, 1 GPU per node.
+SW: Model with 2783M total params, 65M largest layer params.
+  per CPU  |  per GPU |   Options
+   70.00GB |   0.25GB | offload_param=cpu , offload_optimizer=cpu , zero_init=1
+   70.00GB |   0.25GB | offload_param=cpu , offload_optimizer=cpu , zero_init=0
+   62.23GB |   5.43GB | offload_param=none, offload_optimizer=cpu , zero_init=1
+   62.23GB |   5.43GB | offload_param=none, offload_optimizer=cpu , zero_init=0
+    0.37GB |  46.91GB | offload_param=none, offload_optimizer=none, zero_init=1
+   15.56GB |  46.91GB | offload_param=none, offload_optimizer=none, zero_init=0
+```
+
+So you can fit it on a single 80GB GPU and no CPU offload, or a tiny 8GB GPU but then need ~60GB of CPU memory. (Remember this is just the memory for params, optimizer states and gradients - you will need a bit more memory for cuda kernels, activations and temps.)
+
+Then it's a tradeoff of cost vs speed. It'll be cheaper to buy/rent a smaller GPU (or less GPUs since you can use multiple GPUs with Deepspeed ZeRO. But then it'll be slower, so even if you don't care about how fast something will be done, the slowdown has a direct impact on the duration of using the GPU and thus bigger cost. So experiment and compare which works the best.
+
+If you have enough GPU memory make sure to disable the CPU/NVMe offload as it'll make everything faster.
+
+For example, let's repeat the same for 2 GPUs:
+
+```bash
+$ python -c 'from transformers import AutoModel; \
+from deepspeed.runtime.zero.stage3 import estimate_zero3_model_states_mem_needs_all_live; \
+model = AutoModel.from_pretrained("bigscience/T0_3B"); \
+estimate_zero3_model_states_mem_needs_all_live(model, num_gpus_per_node=2, num_nodes=1)'
+[...]
+Estimated memory needed for params, optim states and gradients for a:
+HW: Setup with 1 node, 2 GPUs per node.
+SW: Model with 2783M total params, 65M largest layer params.
+  per CPU  |  per GPU |   Options
+   70.00GB |   0.25GB | offload_param=cpu , offload_optimizer=cpu , zero_init=1
+   70.00GB |   0.25GB | offload_param=cpu , offload_optimizer=cpu , zero_init=0
+   62.23GB |   2.84GB | offload_param=none, offload_optimizer=cpu , zero_init=1
+   62.23GB |   2.84GB | offload_param=none, offload_optimizer=cpu , zero_init=0
+    0.74GB |  23.58GB | offload_param=none, offload_optimizer=none, zero_init=1
+   31.11GB |  23.58GB | offload_param=none, offload_optimizer=none, zero_init=0
+
+```
+
+So here you'd want 2x 32GB GPUs or higher without offloading to CPU.
+
+For full information please see [memory estimators](https://deepspeed.readthedocs.io/en/latest/memory.html).
+
+
+
 ### Filing Issues
 
 Here is how to file an issue so that we could quickly get to the bottom of the issue and help you to unblock your work.
@@ -1655,7 +1761,7 @@ In your report please always include:
 
 5. Unless it's impossible please always use a standard dataset that we can use and not something custom.
 
-6. If possible try to use one of the existing [examples](https://github.com/huggingface/transformers/tree/master/examples/pytorch) to reproduce the problem with.
+6. If possible try to use one of the existing [examples](https://github.com/huggingface/transformers/tree/main/examples/pytorch) to reproduce the problem with.
 
 Things to consider:
 
@@ -1677,18 +1783,58 @@ Things to consider:
 
 ### Troubleshooting
 
-- `deepspeed` process gets killed at startup without a traceback
+#### the `deepspeed` process gets killed at startup without a traceback
 
 If the `deepspeed` process gets killed at launch time without a traceback, that usually means that the program tried
 to allocate more CPU memory than your system has or your process is allowed to allocate and the OS kernel killed that
 process. This is because your configuration file most likely has either `offload_optimizer` or `offload_param` or
 both configured to offload to `cpu`. If you have NVMe, experiment with offloading to NVMe if you're running under
-ZeRO-3.
-
-Work is being done to enable estimating how much memory is needed for a specific model: [PR](https://github.com/microsoft/DeepSpeed/pull/965).
+ZeRO-3. Here is how you can [estimate how much memory is needed for a specific model](https://deepspeed.readthedocs.io/en/latest/memory.html).
 
 
+#### training and/or eval/predict loss is `NaN`
 
+This often happens when one takes a model pre-trained in bf16 mixed precision mode and tries to use it under fp16 (with or without mixed precision). Most models trained on TPU and often the ones released by Google are in this category (e.g. almost all t5-based models). Here the solution is to either use fp32 or bf16 if your hardware supports it (TPU, Ampere GPUs or newer).
+
+The other problem may have to do with using fp16. When you configure this section:
+
+```json
+{
+    "fp16": {
+        "enabled": "auto",
+        "loss_scale": 0,
+        "loss_scale_window": 1000,
+        "initial_scale_power": 16,
+        "hysteresis": 2,
+        "min_loss_scale": 1
+    }
+}
+```
+
+and you see in your log that Deepspeed reports `OVERFLOW!` as follows:
+
+```
+0%|                                                                                                                             | 0/189 [00:00<?, ?it/s]
+ [deepscale] OVERFLOW! Rank 0 Skipping step. Attempted loss scale: 262144, reducing to 262144
+  1%|▌                                                                                                                    | 1/189 [00:00<01:26,  2.17it/s]
+ [deepscale] OVERFLOW! Rank 0 Skipping step. Attempted loss scale: 262144, reducing to 131072.0
+  1%|█▏
+ [...]
+ [deepscale] OVERFLOW! Rank 0 Skipping step. Attempted loss scale: 1, reducing to 1
+ 14%|████████████████▌                                                                                                   | 27/189 [00:14<01:13,  2.21it/s]
+ [deepscale] OVERFLOW! Rank 0 Skipping step. Attempted loss scale: 1, reducing to 1
+ 15%|█████████████████▏                                                                                                  | 28/189 [00:14<01:13,  2.18it/s]
+ [deepscale] OVERFLOW! Rank 0 Skipping step. Attempted loss scale: 1, reducing to 1
+ 15%|█████████████████▊                                                                                                  | 29/189 [00:15<01:13,  2.18it/s]
+ [deepscale] OVERFLOW! Rank 0 Skipping step. Attempted loss scale: 1, reducing to 1
+[...]
+```
+
+that means that the Deepspeed loss scaler can't figure out a scaling co-efficient that overcomes loss overflow.
+
+(the log was massaged to be more readable here.)
+
+In this case you usually need to raise the value of `initial_scale_power`. Setting it to `"initial_scale_power": 32` will typically resolve the problem.
 
 
 
@@ -1708,21 +1854,24 @@ Work is being done to enable estimating how much memory is needed for a specific
 ## Non-Trainer Deepspeed Integration
 
 The [`~deepspeed.HfDeepSpeedConfig`] is used to integrate Deepspeed into the 🤗 Transformers core
-functionality, when [`Trainer`] is not used.
+functionality, when [`Trainer`] is not used. The only thing that it does is handling Deepspeed ZeRO-3 param gathering and automatically splitting the model onto multiple gpus during `from_pretrained` call. Everything else you have to do by yourself.
 
 When using [`Trainer`] everything is automatically taken care of.
 
-When not using [`Trainer`], to efficiently deploy DeepSpeed stage 3, you must instantiate the
-[`~deepspeed.HfDeepSpeedConfig`] object before instantiating the model.
+When not using [`Trainer`], to efficiently deploy DeepSpeed ZeRO-3, you must instantiate the
+[`~deepspeed.HfDeepSpeedConfig`] object before instantiating the model and keep that object alive.
+
+If you're using Deepspeed ZeRO-1 or ZeRO-2 you don't need to use `HfDeepSpeedConfig` at all.
 
 For example for a pretrained model:
 
 ```python
 from transformers.deepspeed import HfDeepSpeedConfig
-from transformers import AutoModel, deepspeed
+from transformers import AutoModel
+import deepspeed
 
 ds_config = {...}  # deepspeed config object or path to the file
-# must run before instantiating the model
+# must run before instantiating the model to detect zero 3
 dschf = HfDeepSpeedConfig(ds_config)  # keep this object alive
 model = AutoModel.from_pretrained("gpt2")
 engine = deepspeed.initialize(model=model, config_params=ds_config, ...)
@@ -1732,21 +1881,188 @@ or for non-pretrained model:
 
 ```python
 from transformers.deepspeed import HfDeepSpeedConfig
-from transformers import AutoModel, AutoConfig, deepspeed
+from transformers import AutoModel, AutoConfig
+import deepspeed
 
 ds_config = {...}  # deepspeed config object or path to the file
-# must run before instantiating the model
+# must run before instantiating the model to detect zero 3
 dschf = HfDeepSpeedConfig(ds_config)  # keep this object alive
 config = AutoConfig.from_pretrained("gpt2")
 model = AutoModel.from_config(config)
 engine = deepspeed.initialize(model=model, config_params=ds_config, ...)
 ```
 
+Please note that if you're not using the [`Trainer`] integration, you're completely on your own. Basically follow the documentation on the [Deepspeed](https://www.deepspeed.ai/) website. Also you have to configure explicitly the config file - you can't use `"auto"` values and you will have to put real values instead.
+
 ## HfDeepSpeedConfig
 
 [[autodoc]] deepspeed.HfDeepSpeedConfig
     - all
 
+### Custom DeepSpeed ZeRO Inference
+
+Here is an example of how one could do DeepSpeed ZeRO Inference without using [`Trainer`] when one can't fit a model onto a single GPU. The solution includes using additional GPUs or/and offloading GPU memory to CPU memory.
+
+The important nuance to understand here is that the way ZeRO is designed you can process different inputs on different GPUs in parallel.
+
+The example has copious notes and is self-documenting.
+
+Make sure to:
+
+1. disable CPU offload if you have enough GPU memory (since it slows things down)
+2. enable bf16 if you own an Ampere or a newer GPU to make things faster. If you don't have that hardware you may enable fp16 as long as you don't use any model that was pre-trained in bf16 mixed precision (such as most t5 models). These usually overflow in fp16 and you will see garbage as output.
+
+```python
+#!/usr/bin/env python
+
+# This script demonstrates how to use Deepspeed ZeRO in an inference mode when one can't fit a model
+# into a single GPU
+#
+# 1. Use 1 GPU with CPU offload
+# 2. Or use multiple GPUs instead
+#
+# First you need to install deepspeed: pip install deepspeed
+#
+# Here we use a 3B "bigscience/T0_3B" model which needs about 15GB GPU RAM - so 1 largish or 2
+# small GPUs can handle it. or 1 small GPU and a lot of CPU memory.
+#
+# To use a larger model like "bigscience/T0" which needs about 50GB, unless you have an 80GB GPU -
+# you will need 2-4 gpus. And then you can adapt the script to handle more gpus if you want to
+# process multiple inputs at once.
+#
+# The provided deepspeed config also activates CPU memory offloading, so chances are that if you
+# have a lot of available CPU memory and you don't mind a slowdown you should be able to load a
+# model that doesn't normally fit into a single GPU. If you have enough GPU memory the program will
+# run faster if you don't want offload to CPU - so disable that section then.
+#
+# To deploy on 1 gpu:
+#
+# deepspeed --num_gpus 1 t0.py
+# or:
+# python -m torch.distributed.run --nproc_per_node=1 t0.py
+#
+# To deploy on 2 gpus:
+#
+# deepspeed --num_gpus 2 t0.py
+# or:
+# python -m torch.distributed.run --nproc_per_node=2 t0.py
+
+
+from transformers import AutoTokenizer, AutoConfig, AutoModelForSeq2SeqLM
+from transformers.deepspeed import HfDeepSpeedConfig
+import deepspeed
+import os
+import torch
+
+os.environ["TOKENIZERS_PARALLELISM"] = "false"  # To avoid warnings about parallelism in tokenizers
+
+# distributed setup
+local_rank = int(os.getenv("LOCAL_RANK", "0"))
+world_size = int(os.getenv("WORLD_SIZE", "1"))
+torch.cuda.set_device(local_rank)
+deepspeed.init_distributed()
+
+model_name = "bigscience/T0_3B"
+
+config = AutoConfig.from_pretrained(model_name)
+model_hidden_size = config.d_model
+
+# batch size has to be divisible by world_size, but can be bigger than world_size
+train_batch_size = 1 * world_size
+
+# ds_config notes
+#
+# - enable bf16 if you use Ampere or higher GPU - this will run in mixed precision and will be
+# faster.
+#
+# - for older GPUs you can enable fp16, but it'll only work for non-bf16 pretrained models - e.g.
+# all official t5 models are bf16-pretrained
+#
+# - set offload_param.device to "none" or completely remove the `offload_param` section if you don't
+# - want CPU offload
+#
+# - if using `offload_param` you can manually finetune stage3_param_persistence_threshold to control
+# - which params should remain on gpus - the larger the value the smaller the offload size
+#
+# For indepth info on Deepspeed config see
+# https://huggingface.co/docs/transformers/main/main_classes/deepspeed
+
+# keeping the same format as json for consistency, except it uses lower case for true/false
+# fmt: off
+ds_config = {
+    "fp16": {
+        "enabled": False
+    },
+    "bf16": {
+        "enabled": False
+    },
+    "zero_optimization": {
+        "stage": 3,
+        "offload_param": {
+            "device": "cpu",
+            "pin_memory": True
+        },
+        "overlap_comm": True,
+        "contiguous_gradients": True,
+        "reduce_bucket_size": model_hidden_size * model_hidden_size,
+        "stage3_prefetch_bucket_size": 0.9 * model_hidden_size * model_hidden_size,
+        "stage3_param_persistence_threshold": 10 * model_hidden_size
+    },
+    "steps_per_print": 2000,
+    "train_batch_size": train_batch_size,
+    "train_micro_batch_size_per_gpu": 1,
+    "wall_clock_breakdown": False
+}
+# fmt: on
+
+# next line instructs transformers to partition the model directly over multiple gpus using
+# deepspeed.zero.Init when model's `from_pretrained` method is called.
+#
+# **it has to be run before loading the model AutoModelForSeq2SeqLM.from_pretrained(model_name)**
+#
+# otherwise the model will first be loaded normally and only partitioned at forward time which is
+# less efficient and when there is little CPU RAM may fail
+dschf = HfDeepSpeedConfig(ds_config)  # keep this object alive
+
+# now a model can be loaded.
+model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
+
+# initialise Deepspeed ZeRO and store only the engine object
+ds_engine = deepspeed.initialize(model=model, config_params=ds_config)[0]
+ds_engine.module.eval()  # inference
+
+# Deepspeed ZeRO can process unrelated inputs on each GPU. So for 2 gpus you process 2 inputs at once.
+# If you use more GPUs adjust for more.
+# And of course if you have just one input to process you then need to pass the same string to both gpus
+# If you use only one GPU, then you will have only rank 0.
+rank = torch.distributed.get_rank()
+if rank == 0:
+    text_in = "Is this review positive or negative? Review: this is the best cast iron skillet you will ever buy"
+elif rank == 1:
+    text_in = "Is this review positive or negative? Review: this is the worst restaurant ever"
+
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+inputs = tokenizer.encode(text_in, return_tensors="pt").to(device=local_rank)
+with torch.no_grad():
+    outputs = ds_engine.module.generate(inputs, synced_gpus=True)
+text_out = tokenizer.decode(outputs[0], skip_special_tokens=True)
+print(f"rank{rank}:\n   in={text_in}\n  out={text_out}")
+```
+
+Let's save it as `t0.py` and run it:
+```
+$ deepspeed --num_gpus 2 t0.py
+rank0:
+   in=Is this review positive or negative? Review: this is the best cast iron skillet you will ever buy
+  out=Positive
+rank1:
+   in=Is this review positive or negative? Review: this is the worst restaurant ever
+  out=negative
+```
+
+This was a very basic example and you will want to adapt it to your needs.
+
+
 ## Main DeepSpeed Resources
 
 - [Project's github](https://github.com/microsoft/deepspeed)

docs/source/en/main_classes/logging.mdx

@@ -36,11 +36,34 @@ Additionally, some `warnings` can be disabled by setting the environment variabl
 `TRANSFORMERS_NO_ADVISORY_WARNINGS` to a true value, like *1*. This will disable any warning that is logged using
 [`logger.warning_advice`]. For example:
 
-
 ```bash
 TRANSFORMERS_NO_ADVISORY_WARNINGS=1 ./myprogram.py
 ```
 
+Here is an example of how to use `logging` in a module:
+
+```python
+from transformers.utils import logging
+
+logging.set_verbosity_info()
+logger = logging.get_logger(__name__)
+logger.info("INFO")
+logger.warning("WARN")
+```
+
+Above, a `logger` instance is created from `logging.get_logger(__name__)`. If you want to use `logging` in a script, you shouldn't pass `__name__` to `logging.get_logger`. For example:
+
+```python
+from transformers.utils import logging
+
+if __name__ == "__main__":
+    logging.set_verbosity_info()
+    # leave it empy or use a string
+    logger = logging.get_logger()
+    logger.info("INFO")
+    logger.warning("WARN")
+```
+
 All the methods of this logging module are documented below, the main ones are
 [`logging.get_verbosity`] to get the current level of verbosity in the logger and
 [`logging.set_verbosity`] to set the verbosity to the level of your choice. In order (from the least
@@ -54,7 +77,7 @@ verbose to the most verbose), those levels (with their corresponding int values
 - `transformers.logging.INFO` (int value, 20): reports error, warnings and basic information.
 - `transformers.logging.DEBUG` (int value, 10): report all information.
 
-By default, `tqdm` progress bars will be displayed during model download. [`logging.disable_progress_bar`] and [`logging.enable_progress_bar`] can be used to suppress or unsuppress this behavior. 
+By default, `tqdm` progress bars will be displayed during model download. [`logging.disable_progress_bar`] and [`logging.enable_progress_bar`] can be used to suppress or unsuppress this behavior.
 
 ## Base setters
 

docs/source/en/main_classes/model.mdx

@@ -86,14 +86,6 @@ Due to Pytorch design, this functionality is only available for floating dtypes.
     - push_to_hub
     - all
 
-## Generation
-
-[[autodoc]] generation_utils.GenerationMixin
-
-[[autodoc]] generation_tf_utils.TFGenerationMixin
-
-[[autodoc]] generation_flax_utils.FlaxGenerationMixin
-
 ## Pushing to the Hub
 
-[[autodoc]] file_utils.PushToHubMixin
+[[autodoc]] utils.PushToHubMixin

docs/source/en/main_classes/output.mdx

@@ -12,7 +12,7 @@ specific language governing permissions and limitations under the License.
 
 # Model outputs
 
-All models have outputs that are instances of subclasses of [`~file_utils.ModelOutput`]. Those are
+All models have outputs that are instances of subclasses of [`~utils.ModelOutput`]. Those are
 data structures containing all the information returned by the model, but that can also be used as tuples or
 dictionaries.
 
@@ -57,7 +57,7 @@ documented on their corresponding model page.
 
 ## ModelOutput
 
-[[autodoc]] file_utils.ModelOutput
+[[autodoc]] utils.ModelOutput
     - to_tuple
 
 ## BaseModelOutput

docs/source/en/main_classes/pipelines.mdx

@@ -39,6 +39,7 @@ There are two categories of pipeline abstractions to be aware about:
   - [`TokenClassificationPipeline`]
   - [`TranslationPipeline`]
   - [`ZeroShotClassificationPipeline`]
+  - [`ZeroShotImageClassificationPipeline`]
 
 ## The pipeline abstraction
 
@@ -78,7 +79,7 @@ GPU. If it doesn't don't hesitate to create an issue.
 ```python
 import datasets
 from transformers import pipeline
-from transformers.pipelines.base import KeyDataset
+from transformers.pipelines.pt_utils import KeyDataset
 from tqdm.auto import tqdm
 
 pipe = pipeline("automatic-speech-recognition", model="facebook/wav2vec2-base-960h", device=0)
@@ -128,7 +129,7 @@ whenever the pipeline uses its streaming ability (so when passing lists or `Data
 
 ```python
 from transformers import pipeline
-from transformers.pipelines.base import KeyDataset
+from transformers.pipelines.pt_utils import KeyDataset
 import datasets
 
 dataset = datasets.load_dataset("imdb", name="plain_text", split="unsupervised")
@@ -428,6 +429,12 @@ See [`TokenClassificationPipeline`] for all details.
     - __call__
     - all
 
+### ZeroShotImageClassificationPipeline
+
+[[autodoc]] ZeroShotImageClassificationPipeline
+    - __call__
+    - all
+
 ## Parent class: `Pipeline`
 
 [[autodoc]] Pipeline

docs/source/en/main_classes/processors.mdx

@@ -12,10 +12,22 @@ specific language governing permissions and limitations under the License.
 
 # Processors
 
-This library includes processors for several traditional tasks. These processors can be used to process a dataset into
-examples that can be fed to a model.
+Processors can mean two different things in the Transformers library:
+- the objects that pre-process inputs for multi-modal models such as [Wav2Vec2](../model_doc/wav2vec2) (speech and text)
+  or [CLIP](../model_doc/clip) (text and vision)
+- deprecated objects that were used in older versions of the library to preprocess data for GLUE or SQUAD.
 
-## Processors
+## Multi-modal processors
+
+Any multi-modal model will require an object to encode or decode the data that groups several modalities (among text,
+vision and audio). This is handled by objects called processors, which group tokenizers (for the text modality) and
+feature extractors (for vision and audio).
+
+Those processors inherit from the following base class that implements the saving and loading functionality:
+
+[[autodoc]] ProcessorMixin
+
+## Deprecated processors
 
 All processors follow the same architecture which is that of the
 [`~data.processors.utils.DataProcessor`]. The processor returns a list of
@@ -53,12 +65,7 @@ Those processors are:
 Additionally, the following method can be used to load values from a data file and convert them to a list of
 [`~data.processors.utils.InputExample`].
 
-automethod,transformers.data.processors.glue.glue_convert_examples_to_features
-
-
-### Example usage
-
-An example using these processors is given in the [run_glue.py](https://github.com/huggingface/transformers/tree/master/examples/legacy/text-classification/run_glue.py) script.
+[[autodoc]] data.processors.glue.glue_convert_examples_to_features
 
 
 ## XNLI
@@ -75,7 +82,7 @@ This library hosts the processor to load the XNLI data:
 
 Please note that since the gold labels are available on the test set, evaluation is performed on the test set.
 
-An example using these processors is given in the [run_xnli.py](https://github.com/huggingface/transformers/tree/master/examples/legacy/text-classification/run_xnli.py) script.
+An example using these processors is given in the [run_xnli.py](https://github.com/huggingface/transformers/tree/main/examples/legacy/text-classification/run_xnli.py) script.
 
 
 ## SQuAD
@@ -102,7 +109,7 @@ They both inherit from the abstract class [`~data.processors.utils.SquadProcesso
 Additionally, the following method can be used to convert SQuAD examples into
 [`~data.processors.utils.SquadFeatures`] that can be used as model inputs.
 
-automethod,transformers.data.processors.squad.squad_convert_examples_to_features
+[[autodoc]] data.processors.squad.squad_convert_examples_to_features
 
 
 These processors as well as the aforementionned method can be used with files containing the data as well as with the
@@ -149,4 +156,4 @@ features = squad_convert_examples_to_features(
 )
 ```
 
-Another example using these processors is given in the [run_squad.py](https://github.com/huggingface/transformers/tree/master/examples/legacy/question-answering/run_squad.py) script.
+Another example using these processors is given in the [run_squad.py](https://github.com/huggingface/transformers/tree/main/examples/legacy/question-answering/run_squad.py) script.

docs/source/en/main_classes/text_generation.mdx

@@ -0,0 +1,40 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# Generation
+
+Each framework has a generate method for auto-regressive text generation implemented in their respective `GenerationMixin` class:
+
+- PyTorch [`~generation_utils.GenerationMixin.generate`] is implemented in [`~generation_utils.GenerationMixin`].
+- TensorFlow [`~generation_tf_utils.TFGenerationMixin.generate`] is implemented in [`~generation_tf_utils.TFGenerationMixin`].
+- Flax/JAX [`~generation_flax_utils.FlaxGenerationMixin.generate`] is implemented in [`~generation_flax_utils.FlaxGenerationMixin`].
+
+## GenerationMixin
+
+[[autodoc]] generation_utils.GenerationMixin
+	- generate
+	- greedy_search
+	- sample
+	- beam_search
+	- beam_sample
+	- group_beam_search
+	- constrained_beam_search
+
+## TFGenerationMixin
+
+[[autodoc]] generation_tf_utils.TFGenerationMixin
+	- generate
+
+## FlaxGenerationMixin
+
+[[autodoc]] generation_flax_utils.FlaxGenerationMixin
+	- generate

docs/source/en/main_classes/trainer.mdx

@@ -40,7 +40,7 @@ The [`Trainer`] contains the basic training loop which supports the above featur
 The [`Trainer`] class is optimized for 🤗 Transformers models and can have surprising behaviors
 when you use it on other models. When using it on your own model, make sure:
 
-- your model always return tuples or subclasses of [`~file_utils.ModelOutput`].
+- your model always return tuples or subclasses of [`~utils.ModelOutput`].
 - your model can compute the loss if a `labels` argument is provided and that loss is returned as the first
   element of the tuple (if your model returns tuples)
 - your model can accept multiple label arguments (use the `label_names` in your [`TrainingArguments`] to indicate their name to the [`Trainer`]) but none of them should be named `"label"`.
@@ -189,6 +189,103 @@ that make things deterministic (.e.g., `torch.backends.cudnn.deterministic`) may
 can't be done by default, but you can enable those yourself if needed.
 
 
+## Specific GPUs Selection
+
+Let's discuss how you can tell your program which GPUs are to be used and in what order.
+
+When using [`DistributedDataParallel`](https://pytorch.org/docs/stable/generated/torch.nn.parallel.DistributedDataParallel.html) to use only a subset of your GPUs, you simply specify the number of GPUs to use. For example, if you have 4 GPUs, but you wish to use the first 2 you can do:
+
+```bash
+python -m torch.distributed.launch --nproc_per_node=2  trainer-program.py ...
+```
+
+if you have either [`accelerate`](https://github.com/huggingface/accelerate) or [`deepspeed`](https://github.com/microsoft/DeepSpeed) installed you can also accomplish the same by using one of:
+```bash
+accelerate launch --num_processes 2 trainer-program.py ...
+```
+
+```bash
+deepspeed --num_gpus 2 trainer-program.py ...
+```
+
+You don't need to use the Accelerate or [the Deepspeed integration](Deepspeed) features to use these launchers.
+
+
+Until now you were able to tell the program how many GPUs to use. Now let's discuss how to select specific GPUs and control their order.
+
+The following environment variables help you control which GPUs to use and their order.
+
+**`CUDA_VISIBLE_DEVICES`**
+
+If you have multiple GPUs and you'd like to use only 1 or a few of those GPUs, set the environment variable `CUDA_VISIBLE_DEVICES` to a list of the GPUs to be used.
+
+For example, let's say you have 4 GPUs: 0, 1, 2 and 3. To run only on the physical GPUs 0 and 2, you can do:
+
+```bash
+CUDA_VISIBLE_DEVICES=0,2 python -m torch.distributed.launch trainer-program.py ...
+```
+
+So now pytorch will see only 2 GPUs, where your physical GPUs 0 and 2 are mapped to `cuda:0` and `cuda:1` correspondingly.
+
+You can even change their order:
+
+```bash
+CUDA_VISIBLE_DEVICES=2,0 python -m torch.distributed.launch trainer-program.py ...
+```
+
+Here your physical GPUs 0 and 2 are mapped to `cuda:1` and `cuda:0` correspondingly.
+
+The above examples were all for `DistributedDataParallel` use pattern, but the same method works for [`DataParallel`](https://pytorch.org/docs/stable/generated/torch.nn.DataParallel.html) as well:
+```bash
+CUDA_VISIBLE_DEVICES=2,0 python trainer-program.py ...
+```
+
+To emulate an environment without GPUs simply set this environment variable to an empty value like so:
+
+```bash
+CUDA_VISIBLE_DEVICES= python trainer-program.py ...
+```
+
+As with any environment variable you can, of course, export those instead of adding these to the command line, as in:
+
+
+```bash
+export CUDA_VISIBLE_DEVICES=0,2
+python -m torch.distributed.launch trainer-program.py ...
+```
+
+but this approach can be confusing since you may forget you set up the environment variable earlier and not understand why the wrong GPUs are used. Therefore, it's a common practice to set the environment variable just for a specific run on the same command line as it's shown in most examples of this section.
+
+**`CUDA_DEVICE_ORDER`**
+
+There is an additional environment variable `CUDA_DEVICE_ORDER` that controls how the physical devices are ordered. The two choices are:
+
+1. ordered by PCIe bus IDs (matches `nvidia-smi`'s order) - this is the default.
+
+```bash
+export CUDA_DEVICE_ORDER=PCI_BUS_ID
+```
+
+2. ordered by GPU compute capabilities
+
+```bash
+export CUDA_DEVICE_ORDER=FASTEST_FIRST
+```
+
+Most of the time you don't need to care about this environment variable, but it's very helpful if you have a lopsided setup where you have an old and a new GPUs physically inserted in such a way so that the slow older card appears to be first. One way to fix that is to swap the cards. But if you can't swap the cards (e.g., if the cooling of the devices gets impacted) then setting `CUDA_DEVICE_ORDER=FASTEST_FIRST` will always put the newer faster card first. It'll be somewhat confusing though since `nvidia-smi` will still report them in the PCIe order.
+
+The other solution to swapping the order is to use:
+
+```bash
+export CUDA_VISIBLE_DEVICES=1,0
+```
+In this example we are working with just 2 GPUs, but of course the same would apply to as many GPUs as your computer has.
+
+Also if you do set this environment variable it's the best to set it in your `~/.bashrc` file or some other startup config file and forget about it.
+
+
+
+
 ## Trainer Integrations
 
 The [`Trainer`] has been extended to support libraries that may dramatically improve your training

docs/source/en/model_doc/auto.mdx

@@ -142,6 +142,10 @@ Likewise, if your `NewModel` is a subclass of [`PreTrainedModel`], make sure its
 
 [[autodoc]] AutoModelForAudioXVector
 
+## AutoModelForMaskedImageModeling
+
+[[autodoc]] AutoModelForMaskedImageModeling
+
 ## AutoModelForObjectDetection
 
 [[autodoc]] AutoModelForObjectDetection
@@ -150,6 +154,14 @@ Likewise, if your `NewModel` is a subclass of [`PreTrainedModel`], make sure its
 
 [[autodoc]] AutoModelForImageSegmentation
 
+## AutoModelForSemanticSegmentation
+
+[[autodoc]] AutoModelForSemanticSegmentation
+
+## AutoModelForInstanceSegmentation
+
+[[autodoc]] AutoModelForInstanceSegmentation
+
 ## TFAutoModel
 
 [[autodoc]] TFAutoModel
@@ -198,6 +210,10 @@ Likewise, if your `NewModel` is a subclass of [`PreTrainedModel`], make sure its
 
 [[autodoc]] TFAutoModelForVision2Seq
 
+## TFAutoModelForSpeechSeq2Seq
+
+[[autodoc]] TFAutoModelForSpeechSeq2Seq
+
 ## FlaxAutoModel
 
 [[autodoc]] FlaxAutoModel

docs/source/en/model_doc/bart.mdx

@@ -38,7 +38,7 @@ This model was contributed by [sshleifer](https://huggingface.co/sshleifer). The
 ### Examples
 
 - Examples and scripts for fine-tuning BART and other models for sequence to sequence tasks can be found in
-  [examples/pytorch/summarization/](https://github.com/huggingface/transformers/tree/master/examples/pytorch/summarization/README.md).
+  [examples/pytorch/summarization/](https://github.com/huggingface/transformers/tree/main/examples/pytorch/summarization/README.md).
 - An example of how to train [`BartForConditionalGeneration`] with a Hugging Face `datasets`
   object can be found in this [forum discussion](https://discuss.huggingface.co/t/train-bart-for-conditional-generation-e-g-summarization/1904).
 - [Distilled checkpoints](https://huggingface.co/models?search=distilbart) are described in this [paper](https://arxiv.org/abs/2010.13002).
@@ -51,7 +51,7 @@ This model was contributed by [sshleifer](https://huggingface.co/sshleifer). The
 - The forward pass of [`BartModel`] will create the `decoder_input_ids` if they are not passed.
   This is different than some other modeling APIs. A typical use case of this feature is mask filling.
 - Model predictions are intended to be identical to the original implementation when
-  `force_bos_token_to_be_generated=True`. This only works, however, if the string you pass to
+  `forced_bos_token_id=0`. This only works, however, if the string you pass to
   [`fairseq.encode`] starts with a space.
 - [`~generation_utils.GenerationMixin.generate`] should be used for conditional generation tasks like
   summarization, see the example in that docstrings.
@@ -152,3 +152,8 @@ assert tok.batch_decode(generated_ids, skip_special_tokens=True) == [
     - __call__
     - encode
     - decode
+
+## FlaxBartForCausalLM
+
+[[autodoc]] FlaxBartForCausalLM
+    - __call__

docs/source/en/model_doc/barthez.mdx

@@ -38,7 +38,7 @@ This model was contributed by [moussakam](https://huggingface.co/moussakam). The
 ### Examples
 
 - BARThez can be fine-tuned on sequence-to-sequence tasks in a similar way as BART, check:
-  [examples/pytorch/summarization/](https://github.com/huggingface/transformers/tree/master/examples/pytorch/summarization/README.md).
+  [examples/pytorch/summarization/](https://github.com/huggingface/transformers/tree/main/examples/pytorch/summarization/README.md).
 
 
 ## BarthezTokenizer

docs/source/en/model_doc/bigbird_pegasus.mdx

@@ -46,7 +46,7 @@ Tips:
 - Sequence length must be divisible by block size.
 - Current implementation supports only **ITC**.
 - Current implementation doesn't support **num_random_blocks = 0**.
-- BigBirdPegasus uses the [PegasusTokenizer](https://github.com/huggingface/transformers/blob/master/src/transformers/models/pegasus/tokenization_pegasus.py).
+- BigBirdPegasus uses the [PegasusTokenizer](https://github.com/huggingface/transformers/blob/main/src/transformers/models/pegasus/tokenization_pegasus.py).
 
 The original code can be found [here](https://github.com/google-research/bigbird).
 

docs/source/en/model_doc/camembert.mdx

@@ -85,6 +85,10 @@ This model was contributed by [camembert](https://huggingface.co/camembert). The
 
 [[autodoc]] TFCamembertModel
 
+## TFCamembertForCasualLM
+
+[[autodoc]] TFCamembertForCausalLM
+
 ## TFCamembertForMaskedLM
 
 [[autodoc]] TFCamembertForMaskedLM

docs/source/en/model_doc/convnext.mdx

@@ -0,0 +1,74 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# ConvNeXT
+
+## Overview
+
+The ConvNeXT model was proposed in [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
+ConvNeXT is a pure convolutional model (ConvNet), inspired by the design of Vision Transformers, that claims to outperform them.
+
+The abstract from the paper is the following:
+
+*The "Roaring 20s" of visual recognition began with the introduction of Vision Transformers (ViTs), which quickly superseded ConvNets as the state-of-the-art image classification model.
+A vanilla ViT, on the other hand, faces difficulties when applied to general computer vision tasks such as object detection and semantic segmentation. It is the hierarchical Transformers
+(e.g., Swin Transformers) that reintroduced several ConvNet priors, making Transformers practically viable as a generic vision backbone and demonstrating remarkable performance on a wide
+variety of vision tasks. However, the effectiveness of such hybrid approaches is still largely credited to the intrinsic superiority of Transformers, rather than the inherent inductive
+biases of convolutions. In this work, we reexamine the design spaces and test the limits of what a pure ConvNet can achieve. We gradually "modernize" a standard ResNet toward the design
+of a vision Transformer, and discover several key components that contribute to the performance difference along the way. The outcome of this exploration is a family of pure ConvNet models
+dubbed ConvNeXt. Constructed entirely from standard ConvNet modules, ConvNeXts compete favorably with Transformers in terms of accuracy and scalability, achieving 87.8% ImageNet top-1 accuracy
+and outperforming Swin Transformers on COCO detection and ADE20K segmentation, while maintaining the simplicity and efficiency of standard ConvNets.*
+
+Tips:
+
+- See the code examples below each model regarding usage.
+
+<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/convnext_architecture.jpg"
+alt="drawing" width="600"/> 
+
+<small> ConvNeXT architecture. Taken from the <a href="https://arxiv.org/abs/2201.03545">original paper</a>.</small>
+
+This model was contributed by [nielsr](https://huggingface.co/nielsr). TensorFlow version of the model was contributed by [ariG23498](https://github.com/ariG23498),
+[gante](https://github.com/gante), and [sayakpaul](https://github.com/sayakpaul) (equal contribution). The original code can be found [here](https://github.com/facebookresearch/ConvNeXt).
+
+## ConvNextConfig
+
+[[autodoc]] ConvNextConfig
+
+
+## ConvNextFeatureExtractor
+
+[[autodoc]] ConvNextFeatureExtractor
+
+
+## ConvNextModel
+
+[[autodoc]] ConvNextModel
+    - forward
+
+
+## ConvNextForImageClassification
+
+[[autodoc]] ConvNextForImageClassification
+    - forward
+
+
+## TFConvNextModel
+
+[[autodoc]] TFConvNextModel
+    - call
+
+
+## TFConvNextForImageClassification
+
+[[autodoc]] TFConvNextForImageClassification
+    - call

docs/source/en/model_doc/ctrl.mdx

@@ -38,10 +38,10 @@ Tips:
 - CTRL was trained with a causal language modeling (CLM) objective and is therefore powerful at predicting the next
   token in a sequence. Leveraging this feature allows CTRL to generate syntactically coherent text as it can be
   observed in the *run_generation.py* example script.
-- The PyTorch models can take the *past* as input, which is the previously computed key/value attention pairs. Using
-  this *past* value prevents the model from re-computing pre-computed values in the context of text generation. See
-  [reusing the past in generative models](../quickstart#using-the-past) for more information on the usage of
-  this argument.
+- The PyTorch models can take the `past_key_values` as input, which is the previously computed key/value attention pairs.
+  TensorFlow models accepts `past` as input. Using the `past_key_values` value prevents the model from re-computing
+  pre-computed values in the context of text generation. See the [`forward`](model_doc/ctrl#transformers.CTRLModel.forward)
+  method for more information on the usage of this argument.
 
 This model was contributed by [keskarnitishr](https://huggingface.co/keskarnitishr). The original code can be found
 [here](https://github.com/salesforce/ctrl).