[test_all] trigger

.github/ISSUE_TEMPLATE/bug-report.yml

@@ -46,7 +46,7 @@ body:
           - Big Model Inference: @SunMarc
           - quantization (bitsandbytes, autogpt): @SunMarc and @younesbelkada
         
-        Documentation: @stevhliu
+        Documentation: @stevhliu and @MKhalusova
         
         Model hub:
 

.github/actions/post-slack/action.yml

@@ -1,79 +0,0 @@
-name: Send message to slack
-
-description: 'Send results to slack'
-author: 'Hugging Face'
-inputs:
-  slack_channel:
-    required: true
-    type: string
-  title:
-    required: true
-    type: string
-  status:
-    required: true
-    type: string
-  slack_token:
-    required: true
-    type: string
-
-runs:
-  using: "composite"
-  steps:
-    - name: Create content to post
-      id: create-message
-      run: |
-        if [ "${{ inputs.status }}" == "success" ]; then
-          echo STATUS_MESSAGE='🟢 Tests are passing!' >> $GITHUB_ENV
-        else
-          echo STATUS_MESSAGE='🔴 Tests failed! Please check the GitHub action link below' >> $GITHUB_ENV
-        fi
-      shell: bash
-
-    - name: Post Canceled results Slack channel
-      id: post-slack
-      uses: slackapi/slack-github-action@6c661ce58804a1a20f6dc5fbee7f0381b469e001
-      with:
-        # Slack channel id, channel name, or user id to post message.
-        # See also: https://api.slack.com/methods/chat.postMessage#channels
-        channel-id: ${{ inputs.slack_channel }}
-        # For posting a rich message using Block Kit
-        payload: |
-          {
-            "text": "${{ inputs.title }}",
-            "blocks": [
-              {
-                "type": "header",
-                "text": {
-                    "type": "plain_text",
-                    "text": "${{ inputs.title }}"
-                }
-              },
-              {
-                "type": "section",
-                "text": {
-                  "type": "mrkdwn",
-                  "text": "${{ env.STATUS_MESSAGE }}"
-                }
-              },
-              {
-                "type": "section",
-                "text": {"type": "mrkdwn", "text": "*Click the button for more details about the commit*"},
-                "accessory": {
-                    "type": "button",
-                    "text": {"type": "plain_text", "text": "Check Commit results"},
-                    "url": "${{ github.event.pull_request.html_url || github.event.head_commit.url }}"
-                }
-              },
-              {
-                "type": "section",
-                "text": {"type": "mrkdwn", "text": "*Click here for more details about the action ran*"},
-                "accessory": {
-                    "type": "button",
-                    "text": {"type": "plain_text", "text": "Check Action results"},
-                    "url": "${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
-                }
-              }
-            ]
-          }
-      env:
-        SLACK_BOT_TOKEN: ${{ inputs.slack_token }}

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

@@ -16,7 +16,7 @@ jobs:
     name: "Add new model like template tests"
     runs-on: ubuntu-22.04
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v3
 
       - name: Install dependencies
         run: |
@@ -74,7 +74,7 @@ jobs:
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: run_all_tests_new_models_test_reports
           path: reports/tests_new_models

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

@@ -27,7 +27,7 @@ jobs:
         uses: docker/setup-buildx-action@v3
       -
         name: Check out code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
       -
         name: Login to DockerHub
         uses: docker/login-action@v3
@@ -76,7 +76,7 @@ jobs:
         uses: docker/setup-buildx-action@v3
       -
         name: Check out code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
       -
         name: Login to DockerHub
         uses: docker/login-action@v3
@@ -113,7 +113,7 @@ jobs:
         uses: docker/setup-buildx-action@v3
       -
         name: Check out code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
       -
         name: Login to DockerHub
         uses: docker/login-action@v3
@@ -145,7 +145,7 @@ jobs:
         uses: docker/setup-buildx-action@v3
       -
         name: Check out code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
       -
         name: Login to DockerHub
         uses: docker/login-action@v3
@@ -181,7 +181,7 @@ jobs:
         uses: docker/setup-buildx-action@v3
       -
         name: Check out code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
       -
         name: Login to DockerHub
         uses: docker/login-action@v3
@@ -207,7 +207,7 @@ jobs:
         uses: docker/setup-buildx-action@v3
       - 
         name: Check out code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
       - 
         name: Login to DockerHub
         uses: docker/login-action@v3
@@ -248,7 +248,7 @@ jobs:
         uses: docker/setup-buildx-action@v3
       -
         name: Check out code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
       -
         name: Login to DockerHub
         uses: docker/login-action@v3
@@ -274,7 +274,7 @@ jobs:
         uses: docker/setup-buildx-action@v3
       - 
         name: Check out code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
       - 
         name: Login to DockerHub
         uses: docker/login-action@v3
@@ -315,7 +315,7 @@ jobs:
         uses: docker/setup-buildx-action@v3
       -
         name: Check out code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
       -
         name: Login to DockerHub
         uses: docker/login-action@v3

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

@@ -30,7 +30,7 @@ jobs:
         uses: docker/setup-buildx-action@v2
       -
         name: Check out code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
       -
         name: Login to DockerHub
         uses: docker/login-action@v2
@@ -67,7 +67,7 @@ jobs:
         uses: docker/setup-buildx-action@v2
       -
         name: Check out code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
       -
         name: Login to DockerHub
         uses: docker/login-action@v2

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

@@ -23,7 +23,7 @@ jobs:
         uses: docker/setup-buildx-action@v2
       -
         name: Check out code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
       -
         id: get-base-image
         name: Get Base Image
@@ -67,7 +67,7 @@ jobs:
         uses: docker/setup-buildx-action@v2
       -
         name: Check out code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
       -
         id: get-base-image
         name: Get Base Image

.github/workflows/check_tiny_models.yml

@@ -17,11 +17,11 @@ jobs:
     runs-on: ubuntu-22.04
     steps:
       - name: Checkout transformers
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
         with:
           fetch-depth: 2
 
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v3
       - name: Set up Python 3.8
         uses: actions/setup-python@v4
         with:
@@ -44,7 +44,7 @@ jobs:
 
       - name: Local tiny model reports artifacts
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: tiny_local_model_creation_reports
           path: tiny_local_models/reports
@@ -56,7 +56,7 @@ jobs:
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: tiny_local_model_creation_reports
           path: reports/tests_pipelines
@@ -76,7 +76,7 @@ jobs:
 
       - name: New tiny model creation reports artifacts
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: tiny_model_creation_reports
           path: tiny_models/reports

.github/workflows/doctest_job.yml

@@ -1,81 +0,0 @@
-name: Doctest job
-
-on:
-  workflow_call:
-    inputs:
-      job_splits:
-        required: true
-        type: string
-      split_keys:
-        required: true
-        type: string
-
-env:
-  HF_HOME: /mnt/cache
-  TRANSFORMERS_IS_CI: yes
-  RUN_SLOW: yes
-  OMP_NUM_THREADS: 16
-  MKL_NUM_THREADS: 16
-  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
-  TF_FORCE_GPU_ALLOW_GROWTH: true
-
-jobs:
-  run_doctests:
-    name: " "
-    strategy:
-      fail-fast: false
-      matrix:
-        split_keys: ${{ fromJson(inputs.split_keys) }}
-    runs-on: [single-gpu, nvidia-gpu, t4, ci]
-    container:
-      image: huggingface/transformers-all-latest-gpu
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    steps:
-      - name: Update clone
-        working-directory: /transformers
-        run: git fetch && git checkout ${{ github.sha }}
-
-      - name: Reinstall transformers in edit mode (remove the one installed during docker image build)
-        working-directory: /transformers
-        run: python3 -m pip uninstall -y transformers && python3 -m pip install -e .[flax]
-
-      - name: GPU visibility
-        working-directory: /transformers
-        run: |
-          python3 utils/print_env.py
-
-      - name: Show installed libraries and their versions
-        run: pip freeze
-
-      - name: Get doctest files
-        working-directory: /transformers
-        run: |
-          echo "${{ toJson(fromJson(inputs.job_splits)[matrix.split_keys]) }}" > doc_tests.txt
-          cat doc_tests.txt
-
-      - name: Set `split_keys`
-        shell: bash
-        run: |
-          echo "${{ matrix.split_keys }}"
-          split_keys=${{ matrix.split_keys }}
-          split_keys=${split_keys//'/'/'_'}
-          echo "split_keys"
-          echo "split_keys=$split_keys" >> $GITHUB_ENV
-
-      - name: Run doctests
-        working-directory: /transformers
-        run: |
-          cat doc_tests.txt
-          python3 -m pytest -v --make-reports doc_tests_gpu_${{ env.split_keys }} --doctest-modules $(cat doc_tests.txt) -sv --doctest-continue-on-failure --doctest-glob="*.md"
-
-      - name: Failure short reports
-        if: ${{ failure() }}
-        continue-on-error: true
-        run: cat /transformers/reports/doc_tests_gpu_${{ env.split_keys }}/failures_short.txt
-
-      - name: "Test suite reports artifacts: doc_tests_gpu_test_reports_${{ env.split_keys }}"
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v4
-        with:
-          name: doc_tests_gpu_test_reports_${{ env.split_keys }}
-          path: /transformers/reports/doc_tests_gpu_${{ env.split_keys }}

.github/workflows/doctests.yml

@@ -3,85 +3,81 @@ name: Doctests
 on:
   push:
     branches:
-      - run_doctest*
+      - doctest*
   repository_dispatch:
   schedule:
     - cron: "17 2 * * *"
 
+
 env:
-  NUM_SLICES: 3
+  HF_HOME: /mnt/cache
+  TRANSFORMERS_IS_CI: yes
+  RUN_SLOW: yes
+  OMP_NUM_THREADS: 16
+  MKL_NUM_THREADS: 16
+  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
+  TF_FORCE_GPU_ALLOW_GROWTH: true
 
 jobs:
-  setup:
-    name: Setup
+  run_doctests:
     runs-on: [single-gpu, nvidia-gpu, t4, ci]
     container:
       image: huggingface/transformers-all-latest-gpu
       options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    outputs:
-      job_splits: ${{ steps.set-matrix.outputs.job_splits }}
-      split_keys: ${{ steps.set-matrix.outputs.split_keys }}
     steps:
-      - name: Update clone
-        working-directory: /transformers
-        run: |
-          git fetch && git checkout ${{ github.sha }}
+      - name: uninstall transformers (installed during docker image build)
+        run: python3 -m pip uninstall -y transformers
 
-      - name: Reinstall transformers in edit mode (remove the one installed during docker image build)
-        working-directory: /transformers
-        run: python3 -m pip uninstall -y transformers && python3 -m pip install -e .
+      - uses: actions/checkout@v3
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
+
+      - name: Install transformers in edit mode
+        run: python3 -m pip install -e .[flax]
+
+      - name: GPU visibility
+        run: |
+          python3 utils/print_env.py
 
       - name: Show installed libraries and their versions
-        working-directory: /transformers
         run: pip freeze
 
-      - name: Check values for matrix
-        working-directory: /transformers
+      - name: Get doctest files
         run: |
-          python3 utils/split_doctest_jobs.py
-          python3 utils/split_doctest_jobs.py --only_return_keys --num_splits ${{ env.NUM_SLICES }}
+          $(python3 -c 'from utils.tests_fetcher import get_all_doctest_files; to_test = get_all_doctest_files(); to_test = " ".join(to_test); fp = open("doc_tests.txt", "w"); fp.write(to_test); fp.close()')
 
-      - id: set-matrix
-        working-directory: /transformers
-        name: Set values for matrix
+      - name: Run doctests
         run: |
-          echo "job_splits=$(python3 utils/split_doctest_jobs.py)" >> $GITHUB_OUTPUT
-          echo "split_keys=$(python3 utils/split_doctest_jobs.py --only_return_keys --num_splits ${{ env.NUM_SLICES }})" >> $GITHUB_OUTPUT
+          python3 -m pytest -v --make-reports doc_tests_gpu --doctest-modules $(cat doc_tests.txt) -sv --doctest-continue-on-failure --doctest-glob="*.md"
+
+      - 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@v3
+        with:
+          name: doc_tests_gpu_test_reports
+          path: reports/doc_tests_gpu
 
-  call_doctest_job:
-    name: "Call doctest jobs"
-    needs: setup
-    strategy:
-      fail-fast: false
-      matrix:
-        split_keys: ${{ fromJson(needs.setup.outputs.split_keys) }}
-    uses: ./.github/workflows/doctest_job.yml
-    with:
-      job_splits: ${{ needs.setup.outputs.job_splits }}
-      split_keys: ${{ toJson(matrix.split_keys) }}
-    secrets: inherit
 
   send_results:
     name: Send results to webhook
     runs-on: ubuntu-22.04
     if: always()
-    needs: [call_doctest_job]
+    needs: [run_doctests]
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions/download-artifact@v4
+      - uses: actions/checkout@v3
+      - uses: actions/download-artifact@v3
       - name: Send message to Slack
         env:
           CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
-          ACCESS_REPO_INFO_TOKEN: ${{ secrets.ACCESS_REPO_INFO_TOKEN }}
-          # Use `CI_SLACK_CHANNEL_DUMMY_TESTS` when doing experimentation
-          SLACK_REPORT_CHANNEL: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY_DOCS }}
+          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
-
-      - name: "Upload results"
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v4
-        with:
-          name: doc_test_results
-          path: doc_test_results

.github/workflows/model-templates.yml

@@ -10,7 +10,7 @@ jobs:
     runs-on: ubuntu-22.04
     steps:
       - name: Checkout repository
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
 
       - name: Install dependencies
         run: |
@@ -75,7 +75,7 @@ jobs:
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: run_all_tests_templates_test_reports
           path: reports/tests_templates

.github/workflows/model_jobs.yml

@@ -96,7 +96,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ inputs.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ inputs.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
           path: /transformers/reports/${{ inputs.machine_type }}_tests_gpu_${{ matrix.folders }}

.github/workflows/push-important-models.yml

@@ -1,137 +0,0 @@
-name: Slow tests on important models (on Push - A10)
-
-on:
-  push:
-    branches: [ main ]
-
-env:
-  IS_GITHUB_CI: "1"
-  OUTPUT_SLACK_CHANNEL_ID: "C06L2SGMEEA"
-  HF_HUB_READ_TOKEN: ${{ secrets.HF_HUB_READ_TOKEN }}
-  HF_HOME: /mnt/cache 
-  TRANSFORMERS_IS_CI: yes 
-  OMP_NUM_THREADS: 8 
-  MKL_NUM_THREADS: 8 
-  RUN_SLOW: yes # For gated repositories, we still need to agree to share information on the Hub repo. page in order to get access. # This token is created under the bot `hf-transformers-bot`. 
-  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }} 
-  TF_FORCE_GPU_ALLOW_GROWTH: true 
-  RUN_PT_TF_CROSS_TESTS: 1
-
-jobs:
-  get_modified_models:
-    name: "Get all modified files"
-    runs-on: ubuntu-latest
-    outputs:
-      matrix: ${{ steps.set-matrix.outputs.matrix }}
-    steps:
-      - name: Check out code
-        uses: actions/checkout@v4
-      
-      - name: Get changed files
-        id: changed-files
-        uses: tj-actions/changed-files@3f54ebb830831fc121d3263c1857cfbdc310cdb9 #v42
-        with:
-          files: src/transformers/models/**
-      
-      - name: Run step if only the files listed above change
-        if: steps.changed-files.outputs.any_changed == 'true'
-        id: set-matrix
-        env:
-          ALL_CHANGED_FILES: ${{ steps.changed-files.outputs.all_changed_files }}
-        run: |
-            model_arrays=()
-            for file in $ALL_CHANGED_FILES; do
-                model_path="${file#*models/}"
-                model_path="models/${model_path%%/*}"
-                if grep -qFx "$model_path" utils/important_models.txt; then
-                    # Append the file to the matrix string
-                    model_arrays+=("$model_path")
-                fi
-            done
-            matrix_string=$(printf '"%s", ' "${model_arrays[@]}" | sed 's/, $//')
-            echo "matrix=[$matrix_string]" >> $GITHUB_OUTPUT
-  test_modified_files:
-    needs: get_modified_models
-    name: Slow & FA2 tests
-    runs-on: ubuntu-latest
-    runs-on: [single-gpu, nvidia-gpu, a10, ci]
-    container:
-      image: huggingface/transformers-all-latest-gpu
-      options: --gpus all --privileged --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    if: ${{ needs.get_modified_models.outputs.matrix != '[]' && needs.get_modified_models.outputs.matrix != '' && fromJson(needs.get_modified_models.outputs.matrix)[0] != null }}
-      strategy:
-        fail-fast: false
-        matrix: 
-          model-name: ${{ fromJson(needs.get_modified_models.outputs.matrix) }}
-
-    steps:
-      - name: Check out code
-        uses: actions/checkout@v4
-      
-      - name: Install locally transformers & other libs
-        run: |
-          apt install sudo
-          sudo -H pip install --upgrade pip
-          sudo -H pip uninstall -y transformers 
-          sudo -H pip install -U -e ".[testing]" 
-          MAX_JOBS=4 pip install flash-attn --no-build-isolation
-          pip install bitsandbytes
-      
-      - name: NVIDIA-SMI
-        run: |
-          nvidia-smi
-      
-      - name: Show installed libraries and their versions
-        run: pip freeze
-      
-      - name: Run FA2 tests
-        id: run_fa2_tests
-        run:
-          pytest -m "flash_attn_test" --make-reports=${{ matrix.model-name }}_fa2_tests/ tests/${{ matrix.model-name }}/test_modeling_*
-      
-      - name: "Test suite reports artifacts: ${{ matrix.model-name }}_fa2_tests"
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v4
-        with:
-          name: ${{ matrix.model-name }}_fa2_tests
-          path: /transformers/reports/${{ matrix.model-name }}_fa2_tests
-      
-      - name: Post to Slack
-        if: always()
-        uses: ./.github/actions/post-slack
-        with:
-          slack_channel: ${{ env.OUTPUT_SLACK_CHANNEL_ID }}
-          title: 🤗 Results of the FA2 tests - ${{ matrix.model-name }}
-          status: ${{ steps.run_fa2_tests.conclusion}}
-          slack_token: ${{ secrets.CI_SLACK_BOT_TOKEN }}
-      
-      - name: Run integration tests
-        id: run_integration_tests
-        if: always()
-        run:
-          pytest -k "IntegrationTest"  --make-reports=tests_integration_${{ matrix.model-name }} tests/${{ matrix.model-name }}/test_modeling_*
-      
-      - name: "Test suite reports artifacts: tests_integration_${{ matrix.model-name }}"
-        if: ${{ always() }}
-        uses: actions/upload-artifact@v4
-        with:
-          name: tests_integration_${{ matrix.model-name }}
-          path: /transformers/reports/tests_integration_${{ matrix.model-name }}
-
-      - name: Post to Slack
-        if: always()
-        uses: ./.github/actions/post-slack
-        with:
-          slack_channel: ${{ env.OUTPUT_SLACK_CHANNEL_ID }}
-          title: 🤗 Results of the Integration tests - ${{ matrix.model-name }}
-          status: ${{ steps.run_integration_tests.conclusion}}
-          slack_token: ${{ secrets.CI_SLACK_BOT_TOKEN }}
-
-      - name: Tailscale # In order to be able to SSH when a test fails
-        if: ${{ failure() || runner.debug == '1'}}
-        uses: huggingface/tailscale-action@v1
-        with:
-          authkey: ${{ secrets.TAILSCALE_SSH_AUTHKEY }}
-          slackChannel: ${{ secrets.SLACK_CIFEEDBACK_CHANNEL }}
-          slackToken: ${{ secrets.SLACK_CIFEEDBACK_BOT_TOKEN }}
-          waitForSSH: true

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

@@ -117,7 +117,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports_postfix_nightly"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports_postfix_nightly
           path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
@@ -178,7 +178,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports_postfix_nightly"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports_postfix_nightly
           path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
@@ -240,7 +240,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports_postfix_nightly"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports_postfix_nightly
           path: /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu
@@ -262,8 +262,8 @@ jobs:
         run: |
           echo "Setup status: ${{ needs.setup.result }}"
 
-      - uses: actions/checkout@v4
-      - uses: actions/download-artifact@v4
+      - uses: actions/checkout@v3
+      - uses: actions/download-artifact@v3
       - name: Send message to Slack
         env:
           CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}

.github/workflows/self-past.yml

@@ -143,7 +143,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports_postfix_${{ inputs.framework }}-${{ inputs.version }}"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports_postfix_${{ inputs.framework }}-${{ inputs.version }}
           path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
@@ -223,7 +223,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports_postfix_${{ inputs.framework }}-${{ inputs.version }}"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports_postfix_${{ inputs.framework }}-${{ inputs.version }}
           path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
@@ -295,7 +295,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports_postfix_${{ inputs.framework }}-${{ inputs.version }}"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports_postfix_${{ inputs.framework }}-${{ inputs.version }}
           path: /transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu
@@ -317,8 +317,8 @@ jobs:
         run: |
           echo "Setup status: ${{ needs.setup.result }}"
 
-      - uses: actions/checkout@v4
-      - uses: actions/download-artifact@v4
+      - uses: actions/checkout@v3
+      - uses: actions/download-artifact@v3
 
       # Create a directory to store test failure tables in the next step
       - name: Create directory
@@ -344,7 +344,7 @@ jobs:
       # Upload complete failure tables, as they might be big and only truncated versions could be sent to Slack.
       - name: Failure table artifacts
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: test_failure_tables_${{ inputs.framework }}-${{ inputs.version }}
           path: test_failure_tables

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

@@ -23,7 +23,7 @@ jobs:
     runs-on: ubuntu-22.04
     steps:
       - name: Checkout transformers
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
         with:
           fetch-depth: 2
 
@@ -121,7 +121,7 @@ jobs:
           python3 utils/tests_fetcher.py --diff_with_last_commit | tee test_preparation.txt
 
       - name: Report fetched tests
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: test_fetched
           path: /transformers/test_preparation.txt
@@ -239,7 +239,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
           path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
@@ -288,7 +288,7 @@ jobs:
           echo "env.CI_BRANCH = ${{ env.CI_BRANCH }}"
           echo "env.CI_SHA = ${{ env.CI_SHA }}"
 
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v3
         # To avoid failure when multiple commits are merged into `main` in a short period of time.
         # Checking out to an old commit beyond the fetch depth will get an error `fatal: reference is not a tree: ...
         # (Only required for `workflow_run` event, where we get the latest HEAD on `main` instead of the event commit)
@@ -303,7 +303,7 @@ jobs:
           git checkout ${{ env.CI_SHA }}
           echo "log = $(git log -n 1)"
 
-      - uses: actions/download-artifact@v4
+      - uses: actions/download-artifact@v3
       - name: Send message to Slack
         env:
           CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}

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

@@ -19,7 +19,7 @@ jobs:
       outputs:
         changed: ${{ steps.was_changed.outputs.changed }}
       steps:
-        - uses: actions/checkout@v4
+        - uses: actions/checkout@v3
           with: 
             fetch-depth: "2"
         

.github/workflows/self-push.yml

@@ -97,7 +97,7 @@ jobs:
           python3 utils/tests_fetcher.py --diff_with_last_commit | tee test_preparation.txt
 
       - name: Report fetched tests
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: test_fetched
           path: /transformers/test_preparation.txt
@@ -209,7 +209,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
           path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
@@ -304,7 +304,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
           path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
@@ -394,7 +394,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports
           path: /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu
@@ -484,7 +484,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports
           path: /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu
@@ -530,7 +530,7 @@ jobs:
           echo "env.CI_BRANCH = ${{ env.CI_BRANCH }}"
           echo "env.CI_SHA = ${{ env.CI_SHA }}"
 
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v3
         # To avoid failure when multiple commits are merged into `main` in a short period of time.
         # Checking out to an old commit beyond the fetch depth will get an error `fatal: reference is not a tree: ...
         # (Only required for `workflow_run` event, where we get the latest HEAD on `main` instead of the event commit)
@@ -545,7 +545,7 @@ jobs:
           git checkout ${{ env.CI_SHA }}
           echo "log = $(git log -n 1)"
 
-      - uses: actions/download-artifact@v4
+      - uses: actions/download-artifact@v3
       - name: Send message to Slack
         env:
           CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}

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

@@ -29,7 +29,7 @@ jobs:
     runs-on: ubuntu-22.04
     steps:
       - name: Checkout transformers
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
         with:
           fetch-depth: 2
 
@@ -171,7 +171,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
           path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
@@ -239,7 +239,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
           path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
@@ -296,7 +296,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_examples_gpu"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_examples_gpu
           path: /transformers/reports/${{ matrix.machine_type }}_examples_gpu
@@ -352,7 +352,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_tests_torch_pipeline_gpu"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_tests_torch_pipeline_gpu
           path: /transformers/reports/${{ matrix.machine_type }}_tests_torch_pipeline_gpu
@@ -409,7 +409,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_tests_torch_deepspeed_gpu_test_reports"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_tests_torch_deepspeed_gpu_test_reports
           path: /transformers/reports/${{ matrix.machine_type }}_tests_torch_deepspeed_gpu
@@ -430,7 +430,7 @@ jobs:
     ]
     steps:
       - name: Checkout transformers
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
         with:
           fetch-depth: 2
 
@@ -443,7 +443,7 @@ jobs:
       - name: Create output directory
         run: mkdir warnings_in_ci
 
-      - uses: actions/download-artifact@v4
+      - uses: actions/download-artifact@v3
         with:
           path: warnings_in_ci
 
@@ -458,7 +458,7 @@ jobs:
 
       - name: Upload artifact
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: warnings_in_ci
           path: warnings_in_ci/selected_warnings.json
@@ -487,8 +487,8 @@ jobs:
           echo "Runner status: ${{ needs.check_runners.result }}"
           echo "Setup status: ${{ needs.setup.result }}"
 
-      - uses: actions/checkout@v4
-      - uses: actions/download-artifact@v4
+      - uses: actions/checkout@v3
+      - uses: actions/download-artifact@v3
       - name: Send message to Slack
         env:
           CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
@@ -513,7 +513,7 @@ jobs:
       # Upload complete failure tables, as they might be big and only truncated versions could be sent to Slack.
       - name: Failure table artifacts
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: test_failure_tables
           path: test_failure_tables

.github/workflows/self-scheduled.yml

@@ -33,7 +33,7 @@ env:
 
 jobs:
   setup:
-    if: contains(fromJSON('["run_tests_gpu", "run_tests_quantization_torch_gpu"]'), inputs.job)
+    if: ${{ inputs.job == 'run_tests_gpu' }}
     name: Setup
     strategy:
       matrix:
@@ -45,7 +45,6 @@ jobs:
     outputs:
       folder_slices: ${{ steps.set-matrix.outputs.folder_slices }}
       slice_ids: ${{ steps.set-matrix.outputs.slice_ids }}
-      quantization_matrix: ${{ steps.set-matrix-quantization.outputs.quantization_matrix }}
     steps:
       - name: Update clone
         working-directory: /transformers
@@ -64,19 +63,11 @@ jobs:
         run: pip freeze
 
       - id: set-matrix
-        if: ${{ inputs.job == 'run_tests_gpu' }}
         name: Identify models to test
         working-directory: /transformers/tests
         run: |
           echo "folder_slices=$(python3 ../utils/split_model_tests.py --num_splits ${{ env.NUM_SLICES }})" >> $GITHUB_OUTPUT
           echo "slice_ids=$(python3 -c 'd = list(range(${{ env.NUM_SLICES }})); print(d)')" >> $GITHUB_OUTPUT
-      
-      - id: set-matrix-quantization
-        if: ${{ inputs.job == 'run_tests_quantization_torch_gpu' }}
-        name: Identify quantization method to test
-        working-directory: /transformers/tests
-        run: |
-          echo "quantization_matrix=$(python3 -c 'import os; tests = os.getcwd(); quantization_tests = os.listdir(os.path.join(tests, "quantization")); d = sorted(list(filter(os.path.isdir, [f"quantization/{x}" for x in quantization_tests]))) ;  print(d)')" >> $GITHUB_OUTPUT
 
       - name: NVIDIA-SMI
         run: |
@@ -143,7 +134,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_tests_torch_pipeline_gpu"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_tests_torch_pipeline_gpu
           path: /transformers/reports/${{ matrix.machine_type }}_tests_torch_pipeline_gpu
@@ -194,7 +185,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_tests_tf_pipeline_gpu"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_tests_tf_pipeline_gpu
           path: /transformers/reports/${{ matrix.machine_type }}_tests_tf_pipeline_gpu
@@ -245,7 +236,7 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_examples_gpu"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_examples_gpu
           path: /transformers/reports/${{ matrix.machine_type }}_examples_gpu
@@ -305,34 +296,23 @@ jobs:
 
       - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports
           path: /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu
 
   run_tests_quantization_torch_gpu:
     if: ${{ inputs.job == 'run_tests_quantization_torch_gpu' }}
-    name: " "
-    needs: setup
+    name: Quantization tests
     strategy:
       fail-fast: false
       matrix:
-        folders: ${{ fromJson(needs.setup.outputs.quantization_matrix) }}
         machine_type: [single-gpu, multi-gpu]
     runs-on: ['${{ matrix.machine_type }}', nvidia-gpu, t4, daily-ci]
     container:
       image: huggingface/transformers-quantization-latest-gpu
       options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
     steps:
-      - name: Echo folder ${{ matrix.folders }}
-        shell: bash
-        run: |
-          echo "${{ matrix.folders }}"
-          matrix_folders=${{ matrix.folders }}
-          matrix_folders=${matrix_folders/'quantization/'/'quantization_'}
-          echo "$matrix_folders"
-          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
-
       - name: Update clone
         working-directory: /transformers
         run: git fetch && git checkout ${{ github.sha }}
@@ -357,19 +337,19 @@ jobs:
       - name: Run quantization tests on GPU
         working-directory: /transformers
         run: |
-          python3 -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_quantization_torch_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
+          python3 -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_quantization_torch_gpu tests/quantization
 
       - name: Failure short reports
         if: ${{ failure() }}
         continue-on-error: true
-        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_quantization_torch_gpu_${{ matrix.folders }}/failures_short.txt
+        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_quantization_torch_gpu/failures_short.txt
 
-      - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_tests_quantization_torch_gpu_${{ env.matrix_folders }}"
+      - name: "Test suite reports artifacts: ${{ matrix.machine_type }}_run_tests_quantization_torch_gpu"
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
-          name: ${{ matrix.machine_type }}_run_tests_quantization_torch_gpu_${{ env.matrix_folders }}
-          path: /transformers/reports/${{ matrix.machine_type }}_tests_quantization_torch_gpu_${{ matrix.folders }}
+          name: ${{ matrix.machine_type }}_run_tests_quantization_torch_gpu
+          path: /transformers/reports/${{ matrix.machine_type }}_tests_quantization_torch_gpu
 
   run_extract_warnings:
     # Let's only do this for the job `run_tests_gpu` to simplify the (already complex) logic.
@@ -379,7 +359,7 @@ jobs:
     needs: [setup, run_tests_gpu]
     steps:
       - name: Checkout transformers
-        uses: actions/checkout@v4
+        uses: actions/checkout@v3
         with:
           fetch-depth: 2
 
@@ -392,7 +372,7 @@ jobs:
       - name: Create output directory
         run: mkdir warnings_in_ci
 
-      - uses: actions/download-artifact@v4
+      - uses: actions/download-artifact@v3
         with:
           path: warnings_in_ci
 
@@ -407,7 +387,7 @@ jobs:
 
       - name: Upload artifact
         if: ${{ always() }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: warnings_in_ci
           path: warnings_in_ci/selected_warnings.json
@@ -433,6 +413,4 @@ jobs:
       slack_report_channel: ${{ inputs.slack_report_channel }}
       # This would be an empty string if `setup` is skipped.
       folder_slices: ${{ needs.setup.outputs.folder_slices }}
-      quantization_matrix: ${{ needs.setup.outputs.quantization_matrix }}
-      
-    secrets: inherit
+    secrets: inherit

.github/workflows/slack-report.yml

@@ -15,9 +15,6 @@ on:
       folder_slices:
         required: true
         type: string
-      quantization_matrix:
-        required: true
-        type: string
 
 
 jobs:
@@ -32,10 +29,9 @@ jobs:
         run: |
           echo "Setup status: ${{ inputs.setup_status }}"
 
-      - uses: actions/checkout@v4
-      - uses: actions/download-artifact@v4
+      - uses: actions/checkout@v3
+      - uses: actions/download-artifact@v3
       - name: Send message to Slack
-        if: ${{ inputs.job != 'run_tests_quantization_torch_gpu' }}
         env:
           CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
           CI_SLACK_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID }}
@@ -57,31 +53,12 @@ jobs:
           pip install slack_sdk
           pip show slack_sdk
           python utils/notification_service.py "${{ inputs.folder_slices }}"
-      
-      - uses: actions/checkout@v4
-      - uses: actions/download-artifact@v4
-      - name: Send message to Slack for quantization workflow
-        if: ${{ inputs.job == 'run_tests_quantization_torch_gpu' }}
-        env:
-          CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
-          ACCESS_REPO_INFO_TOKEN: ${{ secrets.ACCESS_REPO_INFO_TOKEN }}
-          SLACK_REPORT_CHANNEL: ${{ inputs.slack_report_channel }}
-          CI_EVENT: scheduled
-          CI_SHA: ${{ github.sha }}
-          SETUP_STATUS: ${{ inputs.setup_status }}
-        # We pass `needs.setup.outputs.quantization_matrix` as the argument. A processing in `notification_service_quantization.py` to change
-        # `quantization/bnb` to `quantization_bnb` is required, as the artifact names use `_` instead of `/`.
-        run: |
-          sudo apt-get install -y curl
-          pip install slack_sdk
-          pip show slack_sdk
-          python utils/notification_service_quantization.py "${{ inputs.quantization_matrix }}" 
-  
+
       # Upload complete failure tables, as they might be big and only truncated versions could be sent to Slack.
       - name: Failure table artifacts
         # Only the model testing job is concerned for this step
         if: ${{ inputs.job == 'run_tests_gpu' }}
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v3
         with:
           name: prev_ci_results
           path: prev_ci_results

.github/workflows/stale.yml

@@ -12,7 +12,7 @@ jobs:
     env:
       GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
     steps:
-    - uses: actions/checkout@v4
+    - uses: actions/checkout@v3
 
     - name: Setup Python
       uses: actions/setup-python@v4

.github/workflows/update_metdata.yml

@@ -14,7 +14,7 @@ jobs:
         shell: bash -l {0}
 
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v3
 
       - name: Setup environment
         run: |

README.md

@@ -331,7 +331,7 @@ Current number of checkpoints: ![](https://img.shields.io/endpoint?url=https://h
 1. **[CLVP](https://huggingface.co/docs/transformers/model_doc/clvp)** released with the paper [Better speech synthesis through scaling](https://arxiv.org/abs/2305.07243) by James Betker.
 1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
 1. **[CodeLlama](https://huggingface.co/docs/transformers/model_doc/llama_code)** (from MetaAI) released with the paper [Code Llama: Open Foundation Models for Code](https://ai.meta.com/research/publications/code-llama-open-foundation-models-for-code/) by Baptiste Rozière, Jonas Gehring, Fabian Gloeckle, Sten Sootla, Itai Gat, Xiaoqing Ellen Tan, Yossi Adi, Jingyu Liu, Tal Remez, Jérémy Rapin, Artyom Kozhevnikov, Ivan Evtimov, Joanna Bitton, Manish Bhatt, Cristian Canton Ferrer, Aaron Grattafiori, Wenhan Xiong, Alexandre Défossez, Jade Copet, Faisal Azhar, Hugo Touvron, Louis Martin, Nicolas Usunier, Thomas Scialom, Gabriel Synnaeve.
-1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (from Cohere) released with the paper [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere.
+1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (from Cohere) released with the paper [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere. 
 1. **[Conditional DETR](https://huggingface.co/docs/transformers/model_doc/conditional_detr)** (from Microsoft Research Asia) released with the paper [Conditional DETR for Fast Training Convergence](https://arxiv.org/abs/2108.06152) by Depu Meng, Xiaokang Chen, Zejia Fan, Gang Zeng, Houqiang Li, Yuhui Yuan, Lei Sun, Jingdong Wang.
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
 1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
@@ -389,13 +389,11 @@ Current number of checkpoints: ![](https://img.shields.io/endpoint?url=https://h
 1. **[GPTBigCode](https://huggingface.co/docs/transformers/model_doc/gpt_bigcode)** (from BigCode) released with the paper [SantaCoder: don't reach for the stars!](https://arxiv.org/abs/2301.03988) by Loubna Ben Allal, Raymond Li, Denis Kocetkov, Chenghao Mou, Christopher Akiki, Carlos Munoz Ferrandis, Niklas Muennighoff, Mayank Mishra, Alex Gu, Manan Dey, Logesh Kumar Umapathi, Carolyn Jane Anderson, Yangtian Zi, Joel Lamy Poirier, Hailey Schoelkopf, Sergey Troshin, Dmitry Abulkhanov, Manuel Romero, Michael Lappert, Francesco De Toni, Bernardo García del Río, Qian Liu, Shamik Bose, Urvashi Bhattacharyya, Terry Yue Zhuo, Ian Yu, Paulo Villegas, Marco Zocca, Sourab Mangrulkar, David Lansky, Huu Nguyen, Danish Contractor, Luis Villa, Jia Li, Dzmitry Bahdanau, Yacine Jernite, Sean Hughes, Daniel Fried, Arjun Guha, Harm de Vries, Leandro von Werra.
 1. **[GPTSAN-japanese](https://huggingface.co/docs/transformers/model_doc/gptsan-japanese)** released in the repository [tanreinama/GPTSAN](https://github.com/tanreinama/GPTSAN/blob/main/report/model.md) by Toshiyuki Sakamoto(tanreinama).
 1. **[Graphormer](https://huggingface.co/docs/transformers/model_doc/graphormer)** (from Microsoft) released with the paper [Do Transformers Really Perform Bad for Graph Representation?](https://arxiv.org/abs/2106.05234) by Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu.
-1. **[Grounding DINO](https://huggingface.co/docs/transformers/main/model_doc/grounding-dino)** (from Institute for AI, Tsinghua-Bosch Joint Center for ML, Tsinghua University, IDEA Research and others) released with the paper [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://arxiv.org/abs/2303.05499) by Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang.
 1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
 1. **[HerBERT](https://huggingface.co/docs/transformers/model_doc/herbert)** (from Allegro.pl, AGH University of Science and Technology) released with the paper [KLEJ: Comprehensive Benchmark for Polish Language Understanding](https://www.aclweb.org/anthology/2020.acl-main.111.pdf) by Piotr Rybak, Robert Mroczkowski, Janusz Tracz, Ireneusz Gawlik.
 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. **[IDEFICS](https://huggingface.co/docs/transformers/model_doc/idefics)** (from HuggingFace) released with the paper [OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documents](https://huggingface.co/papers/2306.16527) by Hugo Laurençon, Lucile Saulnier, Léo Tronchon, Stas Bekman, Amanpreet Singh, Anton Lozhkov, Thomas Wang, Siddharth Karamcheti, Alexander M. Rush, Douwe Kiela, Matthieu Cord, Victor Sanh.
-1. **[Idefics2](https://huggingface.co/docs/transformers/main/model_doc/idefics2)** (from Hugging Face) released with the blog [IDEFICS2](https://huggingface.co/blog/idefics2) by Léo Tronchon, Hugo Laurencon, Victor Sanh.
 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. **[Informer](https://huggingface.co/docs/transformers/model_doc/informer)** (from Beihang University, UC Berkeley, Rutgers University, SEDD Company) released with the paper [Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting](https://arxiv.org/abs/2012.07436) by Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang.
 1. **[InstructBLIP](https://huggingface.co/docs/transformers/model_doc/instructblip)** (from Salesforce) released with the paper [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://arxiv.org/abs/2305.06500) by Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, Steven Hoi.
@@ -478,7 +476,6 @@ Current number of checkpoints: ![](https://img.shields.io/endpoint?url=https://h
 1. **[Qwen2MoE](https://huggingface.co/docs/transformers/main/model_doc/qwen2_moe)** (from the Qwen team, Alibaba Group) released with [blog post](https://qwenlm.github.io/blog/qwen-moe/) by Bo Zheng, Dayiheng Liu, Rui Men, Junyang Lin, Zhou San, Bowen Yu, An Yang, Mingfeng Xue, Fei Huang, Binyuan Hui, Mei Li, Tianyu Liu, Xingzhang Ren, Xuancheng Ren, Kexin Yang, Chang Zhou, Jingren Zhou.
 1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
-1. **[RecurrentGemma](https://huggingface.co/docs/transformers/main/model_doc/recurrent-gemma)** (from Google) released with the paper [RecurrentGemma: Moving Past Transformers for Efficient Open Language Models](https://storage.googleapis.com/deepmind-media/gemma/recurrentgemma-report.pdf) by the Griffin, RLHF and Gemma Teams.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
 1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Platforms) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.

README_de.md

@@ -327,7 +327,7 @@ Aktuelle Anzahl der Checkpoints: ![](https://img.shields.io/endpoint?url=https:/
 1. **[CLVP](https://huggingface.co/docs/transformers/model_doc/clvp)** released with the paper [Better speech synthesis through scaling](https://arxiv.org/abs/2305.07243) by James Betker.
 1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
 1. **[CodeLlama](https://huggingface.co/docs/transformers/model_doc/llama_code)** (from MetaAI) released with the paper [Code Llama: Open Foundation Models for Code](https://ai.meta.com/research/publications/code-llama-open-foundation-models-for-code/) by Baptiste Rozière, Jonas Gehring, Fabian Gloeckle, Sten Sootla, Itai Gat, Xiaoqing Ellen Tan, Yossi Adi, Jingyu Liu, Tal Remez, Jérémy Rapin, Artyom Kozhevnikov, Ivan Evtimov, Joanna Bitton, Manish Bhatt, Cristian Canton Ferrer, Aaron Grattafiori, Wenhan Xiong, Alexandre Défossez, Jade Copet, Faisal Azhar, Hugo Touvron, Louis Martin, Nicolas Usunier, Thomas Scialom, Gabriel Synnaeve.
-1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (from Cohere) released with the paper [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere.
+1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (from Cohere) released with the paper [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere. 
 1. **[Conditional DETR](https://huggingface.co/docs/transformers/model_doc/conditional_detr)** (from Microsoft Research Asia) released with the paper [Conditional DETR for Fast Training Convergence](https://arxiv.org/abs/2108.06152) by Depu Meng, Xiaokang Chen, Zejia Fan, Gang Zeng, Houqiang Li, Yuhui Yuan, Lei Sun, Jingdong Wang.
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
 1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
@@ -385,13 +385,11 @@ Aktuelle Anzahl der Checkpoints: ![](https://img.shields.io/endpoint?url=https:/
 1. **[GPTBigCode](https://huggingface.co/docs/transformers/model_doc/gpt_bigcode)** (from BigCode) released with the paper [SantaCoder: don't reach for the stars!](https://arxiv.org/abs/2301.03988) by Loubna Ben Allal, Raymond Li, Denis Kocetkov, Chenghao Mou, Christopher Akiki, Carlos Munoz Ferrandis, Niklas Muennighoff, Mayank Mishra, Alex Gu, Manan Dey, Logesh Kumar Umapathi, Carolyn Jane Anderson, Yangtian Zi, Joel Lamy Poirier, Hailey Schoelkopf, Sergey Troshin, Dmitry Abulkhanov, Manuel Romero, Michael Lappert, Francesco De Toni, Bernardo García del Río, Qian Liu, Shamik Bose, Urvashi Bhattacharyya, Terry Yue Zhuo, Ian Yu, Paulo Villegas, Marco Zocca, Sourab Mangrulkar, David Lansky, Huu Nguyen, Danish Contractor, Luis Villa, Jia Li, Dzmitry Bahdanau, Yacine Jernite, Sean Hughes, Daniel Fried, Arjun Guha, Harm de Vries, Leandro von Werra.
 1. **[GPTSAN-japanese](https://huggingface.co/docs/transformers/model_doc/gptsan-japanese)** released in the repository [tanreinama/GPTSAN](https://github.com/tanreinama/GPTSAN/blob/main/report/model.md) by Toshiyuki Sakamoto(tanreinama).
 1. **[Graphormer](https://huggingface.co/docs/transformers/model_doc/graphormer)** (from Microsoft) released with the paper [Do Transformers Really Perform Bad for Graph Representation?](https://arxiv.org/abs/2106.05234) by Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu.
-1. **[Grounding DINO](https://huggingface.co/docs/transformers/main/model_doc/grounding-dino)** (from Institute for AI, Tsinghua-Bosch Joint Center for ML, Tsinghua University, IDEA Research and others) released with the paper [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://arxiv.org/abs/2303.05499) by Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang.
 1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
 1. **[HerBERT](https://huggingface.co/docs/transformers/model_doc/herbert)** (from Allegro.pl, AGH University of Science and Technology) released with the paper [KLEJ: Comprehensive Benchmark for Polish Language Understanding](https://www.aclweb.org/anthology/2020.acl-main.111.pdf) by Piotr Rybak, Robert Mroczkowski, Janusz Tracz, Ireneusz Gawlik.
 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. **[IDEFICS](https://huggingface.co/docs/transformers/model_doc/idefics)** (from HuggingFace) released with the paper [OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documents](https://huggingface.co/papers/2306.16527) by Hugo Laurençon, Lucile Saulnier, Léo Tronchon, Stas Bekman, Amanpreet Singh, Anton Lozhkov, Thomas Wang, Siddharth Karamcheti, Alexander M. Rush, Douwe Kiela, Matthieu Cord, Victor Sanh.
-1. **[Idefics2](https://huggingface.co/docs/transformers/main/model_doc/idefics2)** (from Hugging Face) released with the paper [IDEFICS2](https://huggingface.co/blog/idefics2) by Léo Tronchon, Hugo Laurencon, Victor Sanh.
 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. **[Informer](https://huggingface.co/docs/transformers/model_doc/informer)** (from Beihang University, UC Berkeley, Rutgers University, SEDD Company) released with the paper [Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting](https://arxiv.org/abs/2012.07436) by Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang.
 1. **[InstructBLIP](https://huggingface.co/docs/transformers/model_doc/instructblip)** (from Salesforce) released with the paper [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://arxiv.org/abs/2305.06500) by Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, Steven Hoi.
@@ -474,7 +472,6 @@ Aktuelle Anzahl der Checkpoints: ![](https://img.shields.io/endpoint?url=https:/
 1. **[Qwen2MoE](https://huggingface.co/docs/transformers/main/model_doc/qwen2_moe)** (from the Qwen team, Alibaba Group) released with the paper [blog post](https://qwenlm.github.io/blog/qwen-moe/) by Bo Zheng, Dayiheng Liu, Rui Men, Junyang Lin, Zhou San, Bowen Yu, An Yang, Mingfeng Xue, Fei Huang, Binyuan Hui, Mei Li, Tianyu Liu, Xingzhang Ren, Xuancheng Ren, Kexin Yang, Chang Zhou, Jingren Zhou.
 1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
-1. **[RecurrentGemma](https://huggingface.co/docs/transformers/main/model_doc/recurrent-gemma)** (from Google) released with the paper [RecurrentGemma: Moving Past Transformers for Efficient Open Language Models](https://storage.googleapis.com/deepmind-media/gemma/recurrentgemma-report.pdf) by the Griffin, RLHF and Gemma Teams.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
 1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Platforms) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.

README_es.md

@@ -304,7 +304,7 @@ Número actual de puntos de control: ![](https://img.shields.io/endpoint?url=htt
 1. **[CLVP](https://huggingface.co/docs/transformers/model_doc/clvp)** released with the paper [Better speech synthesis through scaling](https://arxiv.org/abs/2305.07243) by James Betker.
 1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
 1. **[CodeLlama](https://huggingface.co/docs/transformers/model_doc/llama_code)** (from MetaAI) released with the paper [Code Llama: Open Foundation Models for Code](https://ai.meta.com/research/publications/code-llama-open-foundation-models-for-code/) by Baptiste Rozière, Jonas Gehring, Fabian Gloeckle, Sten Sootla, Itai Gat, Xiaoqing Ellen Tan, Yossi Adi, Jingyu Liu, Tal Remez, Jérémy Rapin, Artyom Kozhevnikov, Ivan Evtimov, Joanna Bitton, Manish Bhatt, Cristian Canton Ferrer, Aaron Grattafiori, Wenhan Xiong, Alexandre Défossez, Jade Copet, Faisal Azhar, Hugo Touvron, Louis Martin, Nicolas Usunier, Thomas Scialom, Gabriel Synnaeve.
-1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (from Cohere) released with the paper [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere.
+1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (from Cohere) released with the paper [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere. 
 1. **[Conditional DETR](https://huggingface.co/docs/transformers/model_doc/conditional_detr)** (from Microsoft Research Asia) released with the paper [Conditional DETR for Fast Training Convergence](https://arxiv.org/abs/2108.06152) by Depu Meng, Xiaokang Chen, Zejia Fan, Gang Zeng, Houqiang Li, Yuhui Yuan, Lei Sun, Jingdong Wang.
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
 1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
@@ -362,13 +362,11 @@ Número actual de puntos de control: ![](https://img.shields.io/endpoint?url=htt
 1. **[GPTBigCode](https://huggingface.co/docs/transformers/model_doc/gpt_bigcode)** (from BigCode) released with the paper [SantaCoder: don't reach for the stars!](https://arxiv.org/abs/2301.03988) by Loubna Ben Allal, Raymond Li, Denis Kocetkov, Chenghao Mou, Christopher Akiki, Carlos Munoz Ferrandis, Niklas Muennighoff, Mayank Mishra, Alex Gu, Manan Dey, Logesh Kumar Umapathi, Carolyn Jane Anderson, Yangtian Zi, Joel Lamy Poirier, Hailey Schoelkopf, Sergey Troshin, Dmitry Abulkhanov, Manuel Romero, Michael Lappert, Francesco De Toni, Bernardo García del Río, Qian Liu, Shamik Bose, Urvashi Bhattacharyya, Terry Yue Zhuo, Ian Yu, Paulo Villegas, Marco Zocca, Sourab Mangrulkar, David Lansky, Huu Nguyen, Danish Contractor, Luis Villa, Jia Li, Dzmitry Bahdanau, Yacine Jernite, Sean Hughes, Daniel Fried, Arjun Guha, Harm de Vries, Leandro von Werra.
 1. **[GPTSAN-japanese](https://huggingface.co/docs/transformers/model_doc/gptsan-japanese)** released in the repository [tanreinama/GPTSAN](https://github.com/tanreinama/GPTSAN/blob/main/report/model.md) by Toshiyuki Sakamoto(tanreinama).
 1. **[Graphormer](https://huggingface.co/docs/transformers/model_doc/graphormer)** (from Microsoft) released with the paper [Do Transformers Really Perform Bad for Graph Representation?](https://arxiv.org/abs/2106.05234) by Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu.
-1. **[Grounding DINO](https://huggingface.co/docs/transformers/main/model_doc/grounding-dino)** (from Institute for AI, Tsinghua-Bosch Joint Center for ML, Tsinghua University, IDEA Research and others) released with the paper [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://arxiv.org/abs/2303.05499) by Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang.
 1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
 1. **[HerBERT](https://huggingface.co/docs/transformers/model_doc/herbert)** (from Allegro.pl, AGH University of Science and Technology) released with the paper [KLEJ: Comprehensive Benchmark for Polish Language Understanding](https://www.aclweb.org/anthology/2020.acl-main.111.pdf) by Piotr Rybak, Robert Mroczkowski, Janusz Tracz, Ireneusz Gawlik.
 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. **[IDEFICS](https://huggingface.co/docs/transformers/model_doc/idefics)** (from HuggingFace) released with the paper [OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documents](https://huggingface.co/papers/2306.16527) by Hugo Laurençon, Lucile Saulnier, Léo Tronchon, Stas Bekman, Amanpreet Singh, Anton Lozhkov, Thomas Wang, Siddharth Karamcheti, Alexander M. Rush, Douwe Kiela, Matthieu Cord, Victor Sanh.
-1. **[Idefics2](https://huggingface.co/docs/transformers/main/model_doc/idefics2)** (from Hugging Face) released with the paper [IDEFICS2](https://huggingface.co/blog/idefics2) by Léo Tronchon, Hugo Laurencon, Victor Sanh.
 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. **[Informer](https://huggingface.co/docs/transformers/model_doc/informer)** (from Beihang University, UC Berkeley, Rutgers University, SEDD Company) released with the paper [Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting](https://arxiv.org/abs/2012.07436) by Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang.
 1. **[InstructBLIP](https://huggingface.co/docs/transformers/model_doc/instructblip)** (from Salesforce) released with the paper [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://arxiv.org/abs/2305.06500) by Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, Steven Hoi.
@@ -445,13 +443,12 @@ Número actual de puntos de control: ![](https://img.shields.io/endpoint?url=htt
 1. **[Pop2Piano](https://huggingface.co/docs/transformers/model_doc/pop2piano)** released with the paper [Pop2Piano : Pop Audio-based Piano Cover Generation](https://arxiv.org/abs/2211.00895) by Jongho Choi, Kyogu Lee.
 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. **[PVT](https://huggingface.co/docs/transformers/model_doc/pvt)** (from Nanjing University, The University of Hong Kong etc.) released with the paper [Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions](https://arxiv.org/pdf/2102.12122.pdf) by Wenhai Wang, Enze Xie, Xiang Li, Deng-Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, Ling Shao.
-1. **[PVTv2](https://huggingface.co/docs/transformers/model_doc/pvt_v2)** (from Shanghai AI Laboratory, Nanjing University, The University of Hong Kong etc.) released with the paper [PVT v2: Improved Baselines with Pyramid Vision Transformer](https://arxiv.org/abs/2106.13797) by Wenhai Wang, Enze Xie, Xiang Li, Deng-Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, Ling Shao.
+1. **[PVTv2](https://huggingface.co/docs/transformers/model_doc/pvt_v2)** (from Shanghai AI Laboratory, Nanjing University, The University of Hong Kong etc.) released with the paper [PVT v2: Improved Baselines with Pyramid Vision Transformer](https://arxiv.org/abs/2106.13797) by Wenhai Wang, Enze Xie, Xiang Li, Deng-Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, Ling Shao. 
 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. **[Qwen2](https://huggingface.co/docs/transformers/model_doc/qwen2)** (from the Qwen team, Alibaba Group) released with the paper [Qwen Technical Report](https://arxiv.org/abs/2309.16609) by Jinze Bai, Shuai Bai, Yunfei Chu, Zeyu Cui, Kai Dang, Xiaodong Deng, Yang Fan, Wenbin Ge, Yu Han, Fei Huang, Binyuan Hui, Luo Ji, Mei Li, Junyang Lin, Runji Lin, Dayiheng Liu, Gao Liu, Chengqiang Lu, Keming Lu, Jianxin Ma, Rui Men, Xingzhang Ren, Xuancheng Ren, Chuanqi Tan, Sinan Tan, Jianhong Tu, Peng Wang, Shijie Wang, Wei Wang, Shengguang Wu, Benfeng Xu, Jin Xu, An Yang, Hao Yang, Jian Yang, Shusheng Yang, Yang Yao, Bowen Yu, Hongyi Yuan, Zheng Yuan, Jianwei Zhang, Xingxuan Zhang, Yichang Zhang, Zhenru Zhang, Chang Zhou, Jingren Zhou, Xiaohuan Zhou and Tianhang Zhu.
 1. **[Qwen2MoE](https://huggingface.co/docs/transformers/main/model_doc/qwen2_moe)** (from the Qwen team, Alibaba Group) released with the paper [blog post](https://qwenlm.github.io/blog/qwen-moe/) by Bo Zheng, Dayiheng Liu, Rui Men, Junyang Lin, Zhou San, Bowen Yu, An Yang, Mingfeng Xue, Fei Huang, Binyuan Hui, Mei Li, Tianyu Liu, Xingzhang Ren, Xuancheng Ren, Kexin Yang, Chang Zhou, Jingren Zhou.
 1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
-1. **[RecurrentGemma](https://huggingface.co/docs/transformers/main/model_doc/recurrent-gemma)** (from Google) released with the paper [RecurrentGemma: Moving Past Transformers for Efficient Open Language Models](https://storage.googleapis.com/deepmind-media/gemma/recurrentgemma-report.pdf) by the Griffin, RLHF and Gemma Teams.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
 1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Platforms) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
@@ -474,9 +471,9 @@ Número actual de puntos de control: ![](https://img.shields.io/endpoint?url=htt
 1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (from Facebook), released together with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
 1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (from Tel Aviv University), released together with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
 1. **[SqueezeBERT](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
-1. **[StableLm](https://huggingface.co/docs/transformers/model_doc/stablelm)** (from Stability AI) released with the paper [StableLM 3B 4E1T (Technical Report)](https://stability.wandb.io/stability-llm/stable-lm/reports/StableLM-3B-4E1T--VmlldzoyMjU4?accessToken=u3zujipenkx5g7rtcj9qojjgxpconyjktjkli2po09nffrffdhhchq045vp0wyfo) by  Jonathan Tow, Marco Bellagente, Dakota Mahan, Carlos Riquelme Ruiz, Duy Phung, Maksym Zhuravinskyi, Nathan Cooper, Nikhil Pinnaparaju, Reshinth Adithyan, and James Baicoianu.
-1. **[Starcoder2](https://huggingface.co/docs/transformers/model_doc/starcoder2)** (from BigCode team) released with a coming soon paper.
-1. **[SuperPoint](https://huggingface.co/docs/transformers/model_doc/superpoint)** (from MagicLeap) released with the paper [SuperPoint: Self-Supervised Interest Point Detection and Description](https://arxiv.org/abs/1712.07629) by Daniel DeTone, Tomasz Malisiewicz and Andrew Rabinovich.
+1. **[StableLm](https://huggingface.co/docs/transformers/model_doc/stablelm)** (from Stability AI) released with the paper [StableLM 3B 4E1T (Technical Report)](https://stability.wandb.io/stability-llm/stable-lm/reports/StableLM-3B-4E1T--VmlldzoyMjU4?accessToken=u3zujipenkx5g7rtcj9qojjgxpconyjktjkli2po09nffrffdhhchq045vp0wyfo) by  Jonathan Tow, Marco Bellagente, Dakota Mahan, Carlos Riquelme Ruiz, Duy Phung, Maksym Zhuravinskyi, Nathan Cooper, Nikhil Pinnaparaju, Reshinth Adithyan, and James Baicoianu. 
+1. **[Starcoder2](https://huggingface.co/docs/transformers/model_doc/starcoder2)** (from BigCode team) released with the paper [StarCoder 2 and The Stack v2: The Next Generation](https://arxiv.org/abs/2402.19173) by Anton Lozhkov, Raymond Li, Loubna Ben Allal, Federico Cassano, Joel Lamy-Poirier, Nouamane Tazi, Ao Tang, Dmytro Pykhtar, Jiawei Liu, Yuxiang Wei, Tianyang Liu, Max Tian, Denis Kocetkov, Arthur Zucker, Younes Belkada, Zijian Wang, Qian Liu, Dmitry Abulkhanov, Indraneil Paul, Zhuang Li, Wen-Ding Li, Megan Risdal, Jia Li, Jian Zhu, Terry Yue Zhuo, Evgenii Zheltonozhskii, Nii Osae Osae Dade, Wenhao Yu, Lucas Krauß, Naman Jain, Yixuan Su, Xuanli He, Manan Dey, Edoardo Abati, Yekun Chai, Niklas Muennighoff, Xiangru Tang, Muhtasham Oblokulov, Christopher Akiki, Marc Marone, Chenghao Mou, Mayank Mishra, Alex Gu, Binyuan Hui, Tri Dao, Armel Zebaze, Olivier Dehaene, Nicolas Patry, Canwen Xu, Julian McAuley, Han Hu, Torsten Scholak, Sebastien Paquet, Jennifer Robinson, Carolyn Jane Anderson, Nicolas Chapados, Mostofa Patwary, Nima Tajbakhsh, Yacine Jernite, Carlos Muñoz Ferrandis, Lingming Zhang, Sean Hughes, Thomas Wolf, Arjun Guha, Leandro von Werra, and Harm de Vries.
+1. **[SuperPoint](https://huggingface.co/docs/transformers/model_doc/superpoint)** (from MagicLeap) released with the paper [SuperPoint: Self-Supervised Interest Point Detection and Description](https://arxiv.org/abs/1712.07629) by Daniel DeTone, Tomasz Malisiewicz and Andrew Rabinovich. 
 1. **[SwiftFormer](https://huggingface.co/docs/transformers/model_doc/swiftformer)** (from MBZUAI) released with the paper [SwiftFormer: Efficient Additive Attention for Transformer-based Real-time Mobile Vision Applications](https://arxiv.org/abs/2303.15446) by Abdelrahman Shaker, Muhammad Maaz, Hanoona Rasheed, Salman Khan, Ming-Hsuan Yang, Fahad Shahbaz Khan.
 1. **[Swin Transformer](https://huggingface.co/docs/transformers/model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
 1. **[Swin Transformer V2](https://huggingface.co/docs/transformers/model_doc/swinv2)** (from Microsoft) released with the paper [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) by Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo.

README_fr.md

@@ -383,13 +383,11 @@ Nombre actuel de points de contrôle : ![](https://img.shields.io/endpoint?url=h
 1. **[GPTBigCode](https://huggingface.co/docs/transformers/model_doc/gpt_bigcode)** (de BigCode) a été publié dans l'article [SantaCoder: don't reach for the stars!](https://arxiv.org/abs/2301.03988) par Loubna Ben Allal, Raymond Li, Denis Kocetkov, Chenghao Mou, Christopher Akiki, Carlos Munoz Ferrandis, Niklas Muennighoff, Mayank Mishra, Alex Gu, Manan Dey, Logesh Kumar Umapathi, Carolyn Jane Anderson, Yangtian Zi, Joel Lamy Poirier, Hailey Schoelkopf, Sergey Troshin, Dmitry Abulkhanov, Manuel Romero, Michael Lappert, Francesco De Toni, Bernardo García del Río, Qian Liu, Shamik Bose, Urvashi Bhattacharyya, Terry Yue Zhuo, Ian Yu, Paulo Villegas, Marco Zocca, Sourab Mangrulkar, David Lansky, Huu Nguyen, Danish Contractor, Luis Villa, Jia Li, Dzmitry Bahdanau, Yacine Jernite, Sean Hughes, Daniel Fried, Arjun Guha, Harm de Vries, Leandro von Werra.
 1. **[GPTSAN-japanese](https://huggingface.co/docs/transformers/model_doc/gptsan-japanese)** a été publié dans le dépôt [tanreinama/GPTSAN](https://github.com/tanreinama/GPTSAN/blob/main/report/model.md) par Toshiyuki Sakamoto (tanreinama).
 1. **[Graphormer](https://huggingface.co/docs/transformers/model_doc/graphormer)** (de Microsoft) a été publié dans l'article [Do Transformers Really Perform Bad for Graph Representation?](https://arxiv.org/abs/2106.05234) par Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu.
-1. **[Grounding DINO](https://huggingface.co/docs/transformers/main/model_doc/grounding-dino)** (de Institute for AI, Tsinghua-Bosch Joint Center for ML, Tsinghua University, IDEA Research and others) publié dans l'article [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://arxiv.org/abs/2303.05499) parShilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang.
 1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (de l'UCSD, NVIDIA) a été publié dans l'article [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) par Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
 1. **[HerBERT](https://huggingface.co/docs/transformers/model_doc/herbert)** (d'Allegro.pl, AGH University of Science and Technology) a été publié dans l'article [KLEJ: Comprehensive Benchmark for Polish Language Understanding](https://www.aclweb.org/anthology/2020.acl-main.111.pdf) par Piotr Rybak, Robert Mroczkowski, Janusz Tracz, Ireneusz Gawlik.
 1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (de Facebook) a été publié dans l'article [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) par 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)** (de Berkeley) a été publié dans l'article [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) par Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
 1. **[IDEFICS](https://huggingface.co/docs/transformers/model_doc/idefics)** (de HuggingFace) a été publié dans l'article [OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documents](https://huggingface.co/papers/2306.16527) par Hugo Laurençon, Lucile Saulnier, Léo Tronchon, Stas Bekman, Amanpreet Singh, Anton Lozhkov, Thomas Wang, Siddharth Karamcheti, Alexander M. Rush, Douwe Kiela, Matthieu Cord, Victor Sanh.
-1. **[Idefics2](https://huggingface.co/docs/transformers/main/model_doc/idefics2)** (de Hugging Face) publié dans l'article [IDEFICS2](https://huggingface.co/blog/idefics2) parLéo Tronchon, Hugo Laurencon, Victor Sanh.
 1. **[ImageGPT](https://huggingface.co/docs/transformers/model_doc/imagegpt)** (d'OpenAI) a été publié dans l'article [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) par Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
 1. **[Informer](https://huggingface.co/docs/transformers/model_doc/informer)** (de l'Université de Beihang, UC Berkeley, Rutgers University, SEDD Company) a été publié dans l'article [Informer : Au-delà du Transformer efficace pour la prévision de séries temporel
 1. **[InstructBLIP](https://huggingface.co/docs/transformers/model_doc/instructblip)** (de Salesforce) a été publié dans l'article [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://arxiv.org/abs/2305.06500) de Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, Steven Hoi.
@@ -472,7 +470,6 @@ Nombre actuel de points de contrôle : ![](https://img.shields.io/endpoint?url=h
 1. **[Qwen2MoE](https://huggingface.co/docs/transformers/main/model_doc/qwen2_moe)** (de l'équipe Qwen, Alibaba Group) a été publié avec le rapport technique [blog post](https://qwenlm.github.io/blog/qwen-moe/) par Bo Zheng, Dayiheng Liu, Rui Men, Junyang Lin, Zhou San, Bowen Yu, An Yang, Mingfeng Xue, Fei Huang, Binyuan Hui, Mei Li, Tianyu Liu, Xingzhang Ren, Xuancheng Ren, Kexin Yang, Chang Zhou, Jingren Zhou.
 1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (de Facebook) a été publié dans l'article [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) par Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (de Google Research) a été publié dans l'article [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) par Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat et Ming-Wei Chang.
-1. **[RecurrentGemma](https://huggingface.co/docs/transformers/main/model_doc/recurrent-gemma)** (de Google) publié dans l'article [RecurrentGemma: Moving Past Transformers for Efficient Open Language Models](https://storage.googleapis.com/deepmind-media/gemma/recurrentgemma-report.pdf) parthe Griffin, RLHF and Gemma Teams.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (de Google Research) a été publié dans l'article [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) par Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
 1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (de META Platforms) a été publié dans l'article [Designing Network Design Space](https://arxiv.org/abs/2003.13678) par Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (de Google Research) a été publié dans l'article [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) par Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.

README_hd.md

@@ -336,13 +336,11 @@ conda install conda-forge::transformers
 1. **[GPTBigCode](https://huggingface.co/docs/transformers/model_doc/gpt_bigcode)** (BigCode से) Loubna Ben Allal, Raymond Li, Denis Kocetkov, Chenghao Mou, Christopher Akiki, Carlos Munoz Ferrandis, Niklas Muennighoff, Mayank Mishra, Alex Gu, Manan Dey, Logesh Kumar Umapathi, Carolyn Jane Anderson, Yangtian Zi, Joel Lamy Poirier, Hailey Schoelkopf, Sergey Troshin, Dmitry Abulkhanov, Manuel Romero, Michael Lappert, Francesco De Toni, Bernardo García del Río, Qian Liu, Shamik Bose, Urvashi Bhattacharyya, Terry Yue Zhuo, Ian Yu, Paulo Villegas, Marco Zocca, Sourab Mangrulkar, David Lansky, Huu Nguyen, Danish Contractor, Luis Villa, Jia Li, Dzmitry Bahdanau, Yacine Jernite, Sean Hughes, Daniel Fried, Arjun Guha, Harm de Vries, Leandro von Werra. द्वाराअनुसंधान पत्र [SantaCoder: don't reach for the stars!](https://arxiv.org/abs/2301.03988) के साथ जारी किया गया
 1. **[GPTSAN-japanese](https://huggingface.co/docs/transformers/model_doc/gptsan-japanese)** released in the repository [tanreinama/GPTSAN](https://github.com/tanreinama/GPTSAN/blob/main/report/model.md) by Toshiyuki Sakamoto(tanreinama).
 1. **[Graphormer](https://huggingface.co/docs/transformers/model_doc/graphormer)** (from Microsoft) released with the paper [Do Transformers Really Perform Bad for Graph Representation?](https://arxiv.org/abs/2106.05234) by Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu.
-1. **[Grounding DINO](https://huggingface.co/docs/transformers/main/model_doc/grounding-dino)** (Institute for AI, Tsinghua-Bosch Joint Center for ML, Tsinghua University, IDEA Research and others से) Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang. द्वाराअनुसंधान पत्र [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://arxiv.org/abs/2303.05499) के साथ जारी किया गया
 1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (UCSD, NVIDIA से) साथ में कागज [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) जियारुई जू, शालिनी डी मेलो, सिफ़ी लियू, वोनमिन बायन, थॉमस ब्रेउएल, जान कौट्ज़, ज़ियाओलोंग वांग द्वारा।
 1. **[HerBERT](https://huggingface.co/docs/transformers/model_doc/herbert)** (Allegro.pl, AGH University of Science and Technology से) Piotr Rybak, Robert Mroczkowski, Janusz Tracz, Ireneusz Gawlik. द्वाराअनुसंधान पत्र [KLEJ: Comprehensive Benchmark for Polish Language Understanding](https://www.aclweb.org/anthology/2020.acl-main.111.pdf) के साथ जारी किया गया
 1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (फेसबुक से) साथ में पेपर [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) वेई-निंग सू, बेंजामिन बोल्टे, याओ-हंग ह्यूबर्ट त्साई, कुशाल लखोटिया, रुस्लान सालाखुतदीनोव, अब्देलरहमान मोहम्मद द्वारा।
 1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (बर्कले से) साथ में कागज [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) सेहून किम, अमीर घोलमी, ज़ेवेई याओ, माइकल डब्ल्यू महोनी, कर्ट केटज़र द्वारा।
 1. **[IDEFICS](https://huggingface.co/docs/transformers/model_doc/idefics)** (from HuggingFace) released with the paper [OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documents](https://huggingface.co/papers/2306.16527) by Hugo Laurençon, Lucile Saulnier, Léo Tronchon, Stas Bekman, Amanpreet Singh, Anton Lozhkov, Thomas Wang, Siddharth Karamcheti, Alexander M. Rush, Douwe Kiela, Matthieu Cord, Victor Sanh.
-1. **[Idefics2](https://huggingface.co/docs/transformers/main/model_doc/idefics2)** (Hugging Face से) Léo Tronchon, Hugo Laurencon, Victor Sanh. द्वाराअनुसंधान पत्र [IDEFICS2](https://huggingface.co/blog/idefics2) के साथ जारी किया गया
 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. **[Informer](https://huggingface.co/docs/transformers/model_doc/informer)** (from Beihang University, UC Berkeley, Rutgers University, SEDD Company) released with the paper [Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting](https://arxiv.org/abs/2012.07436) by Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang.
 1. **[InstructBLIP](https://huggingface.co/docs/transformers/model_doc/instructblip)** (Salesforce से) Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, Steven Hoi. द्वाराअनुसंधान पत्र [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://arxiv.org/abs/2305.06500) के साथ जारी किया गया
@@ -425,7 +423,6 @@ conda install conda-forge::transformers
 1. **[Qwen2MoE](https://huggingface.co/docs/transformers/main/model_doc/qwen2_moe)** (the Qwen team, Alibaba Group से) Bo Zheng, Dayiheng Liu, Rui Men, Junyang Lin, Zhou San, Bowen Yu, An Yang, Mingfeng Xue, Fei Huang, Binyuan Hui, Mei Li, Tianyu Liu, Xingzhang Ren, Xuancheng Ren, Kexin Yang, Chang Zhou, Jingren Zhou. द्वाराअनुसंधान पत्र [blog post](https://qwenlm.github.io/blog/qwen-moe/) के साथ जारी किया गया
 1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (फेसबुक से) साथ में कागज [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) पैट्रिक लुईस, एथन पेरेज़, अलेक्जेंड्रा पिक्टस, फैबियो पेट्रोनी, व्लादिमीर कारपुखिन, नमन गोयल, हेनरिक कुटलर, माइक लुईस, वेन-ताउ यिह, टिम रॉकटाशेल, सेबस्टियन रिडेल, डौवे कीला द्वारा।
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (Google अनुसंधान से) केल्विन गु, केंटन ली, ज़ोरा तुंग, पानुपोंग पसुपत और मिंग-वेई चांग द्वारा साथ में दिया गया पेपर [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909)।
-1. **[RecurrentGemma](https://huggingface.co/docs/transformers/main/model_doc/recurrent-gemma)** (Google से) the Griffin, RLHF and Gemma Teams. द्वाराअनुसंधान पत्र [RecurrentGemma: Moving Past Transformers for Efficient Open Language Models](https://storage.googleapis.com/deepmind-media/gemma/recurrentgemma-report.pdf) के साथ जारी किया गया
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
 1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (META रिसर्च से) [Designing Network Design Space](https://arxiv.org/abs/2003.13678) पेपर के साथ जारी किया गया एब्स/2003.13678) इलिजा राडोसावोविक, राज प्रतीक कोसाराजू, रॉस गिर्शिक, कैमिंग ही, पिओटर डॉलर द्वारा।
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (गूगल रिसर्च से) साथ वाला पेपर [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) ह्युंग वोन चुंग, थिबॉल्ट फ़ेवरी, हेनरी त्साई, एम. जॉनसन, सेबेस्टियन रुडर द्वारा।

README_ja.md

@@ -396,13 +396,11 @@ Flax、PyTorch、TensorFlowをcondaでインストールする方法は、それ
 1. **[GPTBigCode](https://huggingface.co/docs/transformers/model_doc/gpt_bigcode)** (BigCode から) Loubna Ben Allal, Raymond Li, Denis Kocetkov, Chenghao Mou, Christopher Akiki, Carlos Munoz Ferrandis, Niklas Muennighoff, Mayank Mishra, Alex Gu, Manan Dey, Logesh Kumar Umapathi, Carolyn Jane Anderson, Yangtian Zi, Joel Lamy Poirier, Hailey Schoelkopf, Sergey Troshin, Dmitry Abulkhanov, Manuel Romero, Michael Lappert, Francesco De Toni, Bernardo García del Río, Qian Liu, Shamik Bose, Urvashi Bhattacharyya, Terry Yue Zhuo, Ian Yu, Paulo Villegas, Marco Zocca, Sourab Mangrulkar, David Lansky, Huu Nguyen, Danish Contractor, Luis Villa, Jia Li, Dzmitry Bahdanau, Yacine Jernite, Sean Hughes, Daniel Fried, Arjun Guha, Harm de Vries, Leandro von Werra. から公開された研究論文 [SantaCoder: don't reach for the stars!](https://arxiv.org/abs/2301.03988)
 1. **[GPTSAN-japanese](https://huggingface.co/docs/transformers/model_doc/gptsan-japanese)** [tanreinama/GPTSAN](https://github.com/tanreinama/GPTSAN/blob/main/report/model.md) 坂本俊之(tanreinama)からリリースされました.
 1. **[Graphormer](https://huggingface.co/docs/transformers/model_doc/graphormer)** (Microsoft から) Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu から公開された研究論文: [Do Transformers Really Perform Bad for Graph Representation?](https://arxiv.org/abs/2106.05234).
-1. **[Grounding DINO](https://huggingface.co/docs/transformers/main/model_doc/grounding-dino)** (Institute for AI, Tsinghua-Bosch Joint Center for ML, Tsinghua University, IDEA Research and others から) Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang. から公開された研究論文 [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://arxiv.org/abs/2303.05499)
 1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (UCSD, NVIDIA から) Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang から公開された研究論文: [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094)
 1. **[HerBERT](https://huggingface.co/docs/transformers/model_doc/herbert)** (Allegro.pl, AGH University of Science and Technology から) Piotr Rybak, Robert Mroczkowski, Janusz Tracz, Ireneusz Gawlik. から公開された研究論文 [KLEJ: Comprehensive Benchmark for Polish Language Understanding](https://www.aclweb.org/anthology/2020.acl-main.111.pdf)
 1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (Facebook から) Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed から公開された研究論文: [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447)
 1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (Berkeley から) Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer から公開された研究論文: [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321)
 1. **[IDEFICS](https://huggingface.co/docs/transformers/model_doc/idefics)** (from HuggingFace) released with the paper [OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documents](https://huggingface.co/papers/2306.16527) by Hugo Laurençon, Lucile Saulnier, Léo Tronchon, Stas Bekman, Amanpreet Singh, Anton Lozhkov, Thomas Wang, Siddharth Karamcheti, Alexander M. Rush, Douwe Kiela, Matthieu Cord, Victor Sanh.
-1. **[Idefics2](https://huggingface.co/docs/transformers/main/model_doc/idefics2)** (Hugging Face から) Léo Tronchon, Hugo Laurencon, Victor Sanh. から公開された研究論文 [IDEFICS2](https://huggingface.co/blog/idefics2)
 1. **[ImageGPT](https://huggingface.co/docs/transformers/model_doc/imagegpt)** (OpenAI から) Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever から公開された研究論文: [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/)
 1. **[Informer](https://huggingface.co/docs/transformers/model_doc/informer)** (from Beihang University, UC Berkeley, Rutgers University, SEDD Company) released with the paper [Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting](https://arxiv.org/abs/2012.07436) by Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang.
 1. **[InstructBLIP](https://huggingface.co/docs/transformers/model_doc/instructblip)** (Salesforce から) Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, Steven Hoi. から公開された研究論文 [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://arxiv.org/abs/2305.06500)
@@ -485,7 +483,6 @@ Flax、PyTorch、TensorFlowをcondaでインストールする方法は、それ
 1. **[Qwen2MoE](https://huggingface.co/docs/transformers/main/model_doc/qwen2_moe)** (the Qwen team, Alibaba Group から) Bo Zheng, Dayiheng Liu, Rui Men, Junyang Lin, Zhou San, Bowen Yu, An Yang, Mingfeng Xue, Fei Huang, Binyuan Hui, Mei Li, Tianyu Liu, Xingzhang Ren, Xuancheng Ren, Kexin Yang, Chang Zhou, Jingren Zhou. から公開された研究論文 [blog post](https://qwenlm.github.io/blog/qwen-moe/)
 1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (Facebook から) Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela から公開された研究論文: [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401)
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (Google Research から) Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang から公開された研究論文: [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909)
-1. **[RecurrentGemma](https://huggingface.co/docs/transformers/main/model_doc/recurrent-gemma)** (Google から) the Griffin, RLHF and Gemma Teams. から公開された研究論文 [RecurrentGemma: Moving Past Transformers for Efficient Open Language Models](https://storage.googleapis.com/deepmind-media/gemma/recurrentgemma-report.pdf)
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (Google Research から) Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya から公開された研究論文: [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451)
 1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (META Platforms から) Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár から公開された研究論文: [Designing Network Design Space](https://arxiv.org/abs/2003.13678)
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (Google Research から) Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder から公開された研究論文: [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821)

README_ko.md

@@ -311,13 +311,11 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[GPTBigCode](https://huggingface.co/docs/transformers/model_doc/gpt_bigcode)** (BigCode 에서 제공)은 Loubna Ben Allal, Raymond Li, Denis Kocetkov, Chenghao Mou, Christopher Akiki, Carlos Munoz Ferrandis, Niklas Muennighoff, Mayank Mishra, Alex Gu, Manan Dey, Logesh Kumar Umapathi, Carolyn Jane Anderson, Yangtian Zi, Joel Lamy Poirier, Hailey Schoelkopf, Sergey Troshin, Dmitry Abulkhanov, Manuel Romero, Michael Lappert, Francesco De Toni, Bernardo García del Río, Qian Liu, Shamik Bose, Urvashi Bhattacharyya, Terry Yue Zhuo, Ian Yu, Paulo Villegas, Marco Zocca, Sourab Mangrulkar, David Lansky, Huu Nguyen, Danish Contractor, Luis Villa, Jia Li, Dzmitry Bahdanau, Yacine Jernite, Sean Hughes, Daniel Fried, Arjun Guha, Harm de Vries, Leandro von Werra.의 [SantaCoder: don't reach for the stars!](https://arxiv.org/abs/2301.03988)논문과 함께 발표했습니다.
 1. **[GPTSAN-japanese](https://huggingface.co/docs/transformers/model_doc/gptsan-japanese)** released in the repository [tanreinama/GPTSAN](https://github.com/tanreinama/GPTSAN/blob/main/report/model.md) by Toshiyuki Sakamoto(tanreinama).
 1. **[Graphormer](https://huggingface.co/docs/transformers/model_doc/graphormer)** (from Microsoft) Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu  의 [Do Transformers Really Perform Bad for Graph Representation?](https://arxiv.org/abs/2106.05234)  논문과 함께 발표했습니다.
-1. **[Grounding DINO](https://huggingface.co/docs/transformers/main/model_doc/grounding-dino)** (Institute for AI, Tsinghua-Bosch Joint Center for ML, Tsinghua University, IDEA Research and others 에서 제공)은 Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang.의 [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://arxiv.org/abs/2303.05499)논문과 함께 발표했습니다.
 1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (UCSD, NVIDIA 에서) Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang 의 [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) 논문과 함께 발표했습니다.
 1. **[HerBERT](https://huggingface.co/docs/transformers/model_doc/herbert)** (Allegro.pl, AGH University of Science and Technology 에서 제공)은 Piotr Rybak, Robert Mroczkowski, Janusz Tracz, Ireneusz Gawlik.의 [KLEJ: Comprehensive Benchmark for Polish Language Understanding](https://www.aclweb.org/anthology/2020.acl-main.111.pdf)논문과 함께 발표했습니다.
 1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (Facebook 에서) Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed 의 [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) 논문과 함께 발표했습니다.
 1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (Berkeley 에서) Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer 의 [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) 논문과 함께 발표했습니다.
 1. **[IDEFICS](https://huggingface.co/docs/transformers/model_doc/idefics)** (from HuggingFace) released with the paper [OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documents](https://huggingface.co/papers/2306.16527) by Hugo Laurençon, Lucile Saulnier, Léo Tronchon, Stas Bekman, Amanpreet Singh, Anton Lozhkov, Thomas Wang, Siddharth Karamcheti, Alexander M. Rush, Douwe Kiela, Matthieu Cord, Victor Sanh.
-1. **[Idefics2](https://huggingface.co/docs/transformers/main/model_doc/idefics2)** (Hugging Face 에서 제공)은 Léo Tronchon, Hugo Laurencon, Victor Sanh.의 [IDEFICS2](https://huggingface.co/blog/idefics2)논문과 함께 발표했습니다.
 1. **[ImageGPT](https://huggingface.co/docs/transformers/model_doc/imagegpt)** (OpenAI 에서) Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever 의 [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) 논문과 함께 발표했습니다.
 1. **[Informer](https://huggingface.co/docs/transformers/model_doc/informer)** (from Beihang University, UC Berkeley, Rutgers University, SEDD Company) released with the paper [Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting](https://arxiv.org/abs/2012.07436) by Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang.
 1. **[InstructBLIP](https://huggingface.co/docs/transformers/model_doc/instructblip)** (Salesforce 에서 제공)은 Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, Steven Hoi.의 [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://arxiv.org/abs/2305.06500)논문과 함께 발표했습니다.
@@ -400,7 +398,6 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[Qwen2MoE](https://huggingface.co/docs/transformers/main/model_doc/qwen2_moe)** (the Qwen team, Alibaba Group 에서 제공)은 Bo Zheng, Dayiheng Liu, Rui Men, Junyang Lin, Zhou San, Bowen Yu, An Yang, Mingfeng Xue, Fei Huang, Binyuan Hui, Mei Li, Tianyu Liu, Xingzhang Ren, Xuancheng Ren, Kexin Yang, Chang Zhou, Jingren Zhou.의 [blog post](https://qwenlm.github.io/blog/qwen-moe/)논문과 함께 발표했습니다.
 1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (Facebook 에서) Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela 의 [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) 논문과 함께 발표했습니다.
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (Google Research 에서) Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang 의 [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) 논문과 함께 발표했습니다.
-1. **[RecurrentGemma](https://huggingface.co/docs/transformers/main/model_doc/recurrent-gemma)** (Google 에서 제공)은 the Griffin, RLHF and Gemma Teams.의 [RecurrentGemma: Moving Past Transformers for Efficient Open Language Models](https://storage.googleapis.com/deepmind-media/gemma/recurrentgemma-report.pdf)논문과 함께 발표했습니다.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (Google Research 에서) Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya 의 [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) 논문과 함께 발표했습니다.
 1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (META Research 에서) Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár 의 [Designing Network Design Space](https://arxiv.org/abs/2003.13678) 논문과 함께 발표했습니다.
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (Google Research 에서) Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder 의 [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) 논문과 함께 발표했습니다.

README_pt-br.md

@@ -333,10 +333,10 @@ Número atual de pontos de verificação: ![](https://img.shields.io/endpoint?ur
 1. **[CLAP](https://huggingface.co/docs/transformers/model_doc/clap)** (from LAION-AI) released with the paper [Large-scale Contrastive Language-Audio Pretraining with Feature Fusion and Keyword-to-Caption Augmentation](https://arxiv.org/abs/2211.06687) by Yusong Wu, Ke Chen, Tianyu Zhang, Yuchen Hui, Taylor Berg-Kirkpatrick, Shlomo Dubnov.
 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. **[CLIPSeg](https://huggingface.co/docs/transformers/model_doc/clipseg)** (from University of Göttingen) released with the paper [Image Segmentation Using Text and Image Prompts](https://arxiv.org/abs/2112.10003) by Timo Lüddecke and Alexander Ecker.
-1. **[CLVP](https://huggingface.co/docs/transformers/model_doc/clvp)** released with the paper [Better speech synthesis through scaling](https://arxiv.org/abs/2305.07243) by James Betker.
+1. **[CLVP](https://huggingface.co/docs/transformers/model_doc/clvp)** released with the paper [Better speech synthesis through scaling](https://arxiv.org/abs/2305.07243) by James Betker. 
 1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
 1. **[CodeLlama](https://huggingface.co/docs/transformers/model_doc/llama_code)** (from MetaAI) released with the paper [Code Llama: Open Foundation Models for Code](https://ai.meta.com/research/publications/code-llama-open-foundation-models-for-code/) by Baptiste Rozière, Jonas Gehring, Fabian Gloeckle, Sten Sootla, Itai Gat, Xiaoqing Ellen Tan, Yossi Adi, Jingyu Liu, Tal Remez, Jérémy Rapin, Artyom Kozhevnikov, Ivan Evtimov, Joanna Bitton, Manish Bhatt, Cristian Canton Ferrer, Aaron Grattafiori, Wenhan Xiong, Alexandre Défossez, Jade Copet, Faisal Azhar, Hugo Touvron, Louis Martin, Nicolas Usunier, Thomas Scialom, Gabriel Synnaeve.
-1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (from Cohere) released with the paper [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere.
+1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (from Cohere) released with the paper [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere. 
 1. **[Conditional DETR](https://huggingface.co/docs/transformers/model_doc/conditional_detr)** (from Microsoft Research Asia) released with the paper [Conditional DETR for Fast Training Convergence](https://arxiv.org/abs/2108.06152) by Depu Meng, Xiaokang Chen, Zejia Fan, Gang Zeng, Houqiang Li, Yuhui Yuan, Lei Sun, Jingdong Wang.
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
 1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
@@ -380,7 +380,7 @@ Número atual de pontos de verificação: ![](https://img.shields.io/endpoint?ur
 1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
 1. **[FocalNet](https://huggingface.co/docs/transformers/model_doc/focalnet)** (from Microsoft Research) released with the paper [Focal Modulation Networks](https://arxiv.org/abs/2203.11926) by Jianwei Yang, Chunyuan Li, Xiyang Dai, Lu Yuan, Jianfeng Gao.
 1. **[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. **[Fuyu](https://huggingface.co/docs/transformers/model_doc/fuyu)** (from ADEPT) Rohan Bavishi, Erich Elsen, Curtis Hawthorne, Maxwell Nye, Augustus Odena, Arushi Somani, Sağnak Taşırlar. Released with the paper [blog post](https://www.adept.ai/blog/fuyu-8b)
+1. **[Fuyu](https://huggingface.co/docs/transformers/model_doc/fuyu)** (from ADEPT) Rohan Bavishi, Erich Elsen, Curtis Hawthorne, Maxwell Nye, Augustus Odena, Arushi Somani, Sağnak Taşırlar. Released with the paper [blog post](https://www.adept.ai/blog/fuyu-8b) 
 1. **[Gemma](https://huggingface.co/docs/transformers/model_doc/gemma)** (from Google) released with the paper [Gemma: Open Models Based on Gemini Technology and Research](https://blog.google/technology/developers/gemma-open-models/) by the Gemma Google team.
 1. **[GIT](https://huggingface.co/docs/transformers/model_doc/git)** (from Microsoft Research) released with the paper [GIT: A Generative Image-to-text Transformer for Vision and Language](https://arxiv.org/abs/2205.14100) by Jianfeng Wang, Zhengyuan Yang, Xiaowei Hu, Linjie Li, Kevin Lin, Zhe Gan, Zicheng Liu, Ce Liu, Lijuan Wang.
 1. **[GLPN](https://huggingface.co/docs/transformers/model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
@@ -394,13 +394,11 @@ Número atual de pontos de verificação: ![](https://img.shields.io/endpoint?ur
 1. **[GPTBigCode](https://huggingface.co/docs/transformers/model_doc/gpt_bigcode)** (from BigCode) released with the paper [SantaCoder: don't reach for the stars!](https://arxiv.org/abs/2301.03988) by Loubna Ben Allal, Raymond Li, Denis Kocetkov, Chenghao Mou, Christopher Akiki, Carlos Munoz Ferrandis, Niklas Muennighoff, Mayank Mishra, Alex Gu, Manan Dey, Logesh Kumar Umapathi, Carolyn Jane Anderson, Yangtian Zi, Joel Lamy Poirier, Hailey Schoelkopf, Sergey Troshin, Dmitry Abulkhanov, Manuel Romero, Michael Lappert, Francesco De Toni, Bernardo García del Río, Qian Liu, Shamik Bose, Urvashi Bhattacharyya, Terry Yue Zhuo, Ian Yu, Paulo Villegas, Marco Zocca, Sourab Mangrulkar, David Lansky, Huu Nguyen, Danish Contractor, Luis Villa, Jia Li, Dzmitry Bahdanau, Yacine Jernite, Sean Hughes, Daniel Fried, Arjun Guha, Harm de Vries, Leandro von Werra.
 1. **[GPTSAN-japanese](https://huggingface.co/docs/transformers/model_doc/gptsan-japanese)** released in the repository [tanreinama/GPTSAN](https://github.com/tanreinama/GPTSAN/blob/main/report/model.md) by Toshiyuki Sakamoto(tanreinama).
 1. **[Graphormer](https://huggingface.co/docs/transformers/model_doc/graphormer)** (from Microsoft) released with the paper [Do Transformers Really Perform Bad for Graph Representation?](https://arxiv.org/abs/2106.05234) by Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu.
-1. **[Grounding DINO](https://huggingface.co/docs/transformers/main/model_doc/grounding-dino)** (from Institute for AI, Tsinghua-Bosch Joint Center for ML, Tsinghua University, IDEA Research and others) released with the paper [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://arxiv.org/abs/2303.05499) by Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang.
 1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
 1. **[HerBERT](https://huggingface.co/docs/transformers/model_doc/herbert)** (from Allegro.pl, AGH University of Science and Technology) released with the paper [KLEJ: Comprehensive Benchmark for Polish Language Understanding](https://www.aclweb.org/anthology/2020.acl-main.111.pdf) by Piotr Rybak, Robert Mroczkowski, Janusz Tracz, Ireneusz Gawlik.
 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. **[IDEFICS](https://huggingface.co/docs/transformers/model_doc/idefics)** (from HuggingFace) released with the paper [OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documents](https://huggingface.co/papers/2306.16527) by Hugo Laurençon, Lucile Saulnier, Léo Tronchon, Stas Bekman, Amanpreet Singh, Anton Lozhkov, Thomas Wang, Siddharth Karamcheti, Alexander M. Rush, Douwe Kiela, Matthieu Cord, Victor Sanh.
-1. **[Idefics2](https://huggingface.co/docs/transformers/main/model_doc/idefics2)** (from Hugging Face) released with the paper [IDEFICS2](https://huggingface.co/blog/idefics2) by Léo Tronchon, Hugo Laurencon, Victor Sanh.
 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. **[Informer](https://huggingface.co/docs/transformers/model_doc/informer)** (from Beihang University, UC Berkeley, Rutgers University, SEDD Company) released with the paper [Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting](https://arxiv.org/abs/2012.07436) by Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang.
 1. **[InstructBLIP](https://huggingface.co/docs/transformers/model_doc/instructblip)** (from Salesforce) released with the paper [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://arxiv.org/abs/2305.06500) by Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, Steven Hoi.
@@ -437,7 +435,7 @@ Número atual de pontos de verificação: ![](https://img.shields.io/endpoint?ur
 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. **[MGP-STR](https://huggingface.co/docs/transformers/model_doc/mgp-str)** (from Alibaba Research) released with the paper [Multi-Granularity Prediction for Scene Text Recognition](https://arxiv.org/abs/2209.03592) by Peng Wang, Cheng Da, and Cong Yao.
 1. **[Mistral](https://huggingface.co/docs/transformers/model_doc/mistral)** (from Mistral AI) by The [Mistral AI](https://mistral.ai) team: Albert Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lélio Renard Lavaud, Lucile Saulnier, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix, William El Sayed.
-1. **[Mixtral](https://huggingface.co/docs/transformers/model_doc/mixtral)** (from Mistral AI) by The [Mistral AI](https://mistral.ai) team: Albert Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lélio Renard Lavaud, Lucile Saulnier, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix, William El Sayed.
+1. **[Mixtral](https://huggingface.co/docs/transformers/model_doc/mixtral)** (from Mistral AI) by The [Mistral AI](https://mistral.ai) team: Albert Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lélio Renard Lavaud, Lucile Saulnier, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix, William El Sayed. 
 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. **[MMS](https://huggingface.co/docs/transformers/model_doc/mms)** (from Facebook) released with the paper [Scaling Speech Technology to 1,000+ Languages](https://arxiv.org/abs/2305.13516) by Vineel Pratap, Andros Tjandra, Bowen Shi, Paden Tomasello, Arun Babu, Sayani Kundu, Ali Elkahky, Zhaoheng Ni, Apoorv Vyas, Maryam Fazel-Zarandi, Alexei Baevski, Yossi Adi, Xiaohui Zhang, Wei-Ning Hsu, Alexis Conneau, Michael Auli.
 1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
@@ -483,7 +481,6 @@ Número atual de pontos de verificação: ![](https://img.shields.io/endpoint?ur
 1. **[Qwen2MoE](https://huggingface.co/docs/transformers/main/model_doc/qwen2_moe)** (from the Qwen team, Alibaba Group) released with the paper [blog post](https://qwenlm.github.io/blog/qwen-moe/) by Bo Zheng, Dayiheng Liu, Rui Men, Junyang Lin, Zhou San, Bowen Yu, An Yang, Mingfeng Xue, Fei Huang, Binyuan Hui, Mei Li, Tianyu Liu, Xingzhang Ren, Xuancheng Ren, Kexin Yang, Chang Zhou, Jingren Zhou.
 1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
-1. **[RecurrentGemma](https://huggingface.co/docs/transformers/main/model_doc/recurrent-gemma)** (from Google) released with the paper [RecurrentGemma: Moving Past Transformers for Efficient Open Language Models](https://storage.googleapis.com/deepmind-media/gemma/recurrentgemma-report.pdf) by the Griffin, RLHF and Gemma Teams.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
 1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Platforms) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.

README_ru.md

@@ -323,10 +323,10 @@ conda install conda-forge::transformers
 1. **[CLAP](https://huggingface.co/docs/transformers/model_doc/clap)** (from LAION-AI) released with the paper [Large-scale Contrastive Language-Audio Pretraining with Feature Fusion and Keyword-to-Caption Augmentation](https://arxiv.org/abs/2211.06687) by Yusong Wu, Ke Chen, Tianyu Zhang, Yuchen Hui, Taylor Berg-Kirkpatrick, Shlomo Dubnov.
 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. **[CLIPSeg](https://huggingface.co/docs/transformers/model_doc/clipseg)** (from University of Göttingen) released with the paper [Image Segmentation Using Text and Image Prompts](https://arxiv.org/abs/2112.10003) by Timo Lüddecke and Alexander Ecker.
-1. **[CLVP](https://huggingface.co/docs/transformers/model_doc/clvp)** released with the paper [Better speech synthesis through scaling](https://arxiv.org/abs/2305.07243) by James Betker.
+1. **[CLVP](https://huggingface.co/docs/transformers/model_doc/clvp)** released with the paper [Better speech synthesis through scaling](https://arxiv.org/abs/2305.07243) by James Betker. 
 1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
 1. **[CodeLlama](https://huggingface.co/docs/transformers/model_doc/llama_code)** (from MetaAI) released with the paper [Code Llama: Open Foundation Models for Code](https://ai.meta.com/research/publications/code-llama-open-foundation-models-for-code/) by Baptiste Rozière, Jonas Gehring, Fabian Gloeckle, Sten Sootla, Itai Gat, Xiaoqing Ellen Tan, Yossi Adi, Jingyu Liu, Tal Remez, Jérémy Rapin, Artyom Kozhevnikov, Ivan Evtimov, Joanna Bitton, Manish Bhatt, Cristian Canton Ferrer, Aaron Grattafiori, Wenhan Xiong, Alexandre Défossez, Jade Copet, Faisal Azhar, Hugo Touvron, Louis Martin, Nicolas Usunier, Thomas Scialom, Gabriel Synnaeve.
-1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (from Cohere) released with the paper [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere.
+1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (from Cohere) released with the paper [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere. 
 1. **[Conditional DETR](https://huggingface.co/docs/transformers/model_doc/conditional_detr)** (from Microsoft Research Asia) released with the paper [Conditional DETR for Fast Training Convergence](https://arxiv.org/abs/2108.06152) by Depu Meng, Xiaokang Chen, Zejia Fan, Gang Zeng, Houqiang Li, Yuhui Yuan, Lei Sun, Jingdong Wang.
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
 1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
@@ -384,13 +384,11 @@ conda install conda-forge::transformers
 1. **[GPTBigCode](https://huggingface.co/docs/transformers/model_doc/gpt_bigcode)** (from BigCode) released with the paper [SantaCoder: don't reach for the stars!](https://arxiv.org/abs/2301.03988) by Loubna Ben Allal, Raymond Li, Denis Kocetkov, Chenghao Mou, Christopher Akiki, Carlos Munoz Ferrandis, Niklas Muennighoff, Mayank Mishra, Alex Gu, Manan Dey, Logesh Kumar Umapathi, Carolyn Jane Anderson, Yangtian Zi, Joel Lamy Poirier, Hailey Schoelkopf, Sergey Troshin, Dmitry Abulkhanov, Manuel Romero, Michael Lappert, Francesco De Toni, Bernardo García del Río, Qian Liu, Shamik Bose, Urvashi Bhattacharyya, Terry Yue Zhuo, Ian Yu, Paulo Villegas, Marco Zocca, Sourab Mangrulkar, David Lansky, Huu Nguyen, Danish Contractor, Luis Villa, Jia Li, Dzmitry Bahdanau, Yacine Jernite, Sean Hughes, Daniel Fried, Arjun Guha, Harm de Vries, Leandro von Werra.
 1. **[GPTSAN-japanese](https://huggingface.co/docs/transformers/model_doc/gptsan-japanese)** released in the repository [tanreinama/GPTSAN](https://github.com/tanreinama/GPTSAN/blob/main/report/model.md) by Toshiyuki Sakamoto(tanreinama).
 1. **[Graphormer](https://huggingface.co/docs/transformers/model_doc/graphormer)** (from Microsoft) released with the paper [Do Transformers Really Perform Bad for Graph Representation?](https://arxiv.org/abs/2106.05234) by Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu.
-1. **[Grounding DINO](https://huggingface.co/docs/transformers/main/model_doc/grounding-dino)** (from Institute for AI, Tsinghua-Bosch Joint Center for ML, Tsinghua University, IDEA Research and others) released with the paper [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://arxiv.org/abs/2303.05499) by Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang.
 1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
 1. **[HerBERT](https://huggingface.co/docs/transformers/model_doc/herbert)** (from Allegro.pl, AGH University of Science and Technology) released with the paper [KLEJ: Comprehensive Benchmark for Polish Language Understanding](https://www.aclweb.org/anthology/2020.acl-main.111.pdf) by Piotr Rybak, Robert Mroczkowski, Janusz Tracz, Ireneusz Gawlik.
 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. **[IDEFICS](https://huggingface.co/docs/transformers/model_doc/idefics)** (from HuggingFace) released with the paper [OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documents](https://huggingface.co/papers/2306.16527) by Hugo Laurençon, Lucile Saulnier, Léo Tronchon, Stas Bekman, Amanpreet Singh, Anton Lozhkov, Thomas Wang, Siddharth Karamcheti, Alexander M. Rush, Douwe Kiela, Matthieu Cord, Victor Sanh.
-1. **[Idefics2](https://huggingface.co/docs/transformers/main/model_doc/idefics2)** (from Hugging Face) released with the paper [IDEFICS2](https://huggingface.co/blog/idefics2) by Léo Tronchon, Hugo Laurencon, Victor Sanh.
 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. **[Informer](https://huggingface.co/docs/transformers/model_doc/informer)** (from Beihang University, UC Berkeley, Rutgers University, SEDD Company) released with the paper [Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting](https://arxiv.org/abs/2012.07436) by Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang.
 1. **[InstructBLIP](https://huggingface.co/docs/transformers/model_doc/instructblip)** (from Salesforce) released with the paper [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://arxiv.org/abs/2305.06500) by Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, Steven Hoi.
@@ -426,8 +424,8 @@ conda install conda-forge::transformers
 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. **[MGP-STR](https://huggingface.co/docs/transformers/model_doc/mgp-str)** (from Alibaba Research) released with the paper [Multi-Granularity Prediction for Scene Text Recognition](https://arxiv.org/abs/2209.03592) by Peng Wang, Cheng Da, and Cong Yao.
-1. **[Mistral](https://huggingface.co/docs/transformers/model_doc/mistral)** (from Mistral AI) by The [Mistral AI](https://mistral.ai) team: Albert Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lélio Renard Lavaud, Lucile Saulnier, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix, William El Sayed.
-1. **[Mixtral](https://huggingface.co/docs/transformers/model_doc/mixtral)** (from Mistral AI) by The [Mistral AI](https://mistral.ai) team: Albert Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lélio Renard Lavaud, Lucile Saulnier, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix, William El Sayed.
+1. **[Mistral](https://huggingface.co/docs/transformers/model_doc/mistral)** (from Mistral AI) by The [Mistral AI](https://mistral.ai) team: Albert Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lélio Renard Lavaud, Lucile Saulnier, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix, William El Sayed. 
+1. **[Mixtral](https://huggingface.co/docs/transformers/model_doc/mixtral)** (from Mistral AI) by The [Mistral AI](https://mistral.ai) team: Albert Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lélio Renard Lavaud, Lucile Saulnier, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix, William El Sayed. 
 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. **[MMS](https://huggingface.co/docs/transformers/model_doc/mms)** (from Facebook) released with the paper [Scaling Speech Technology to 1,000+ Languages](https://arxiv.org/abs/2305.13516) by Vineel Pratap, Andros Tjandra, Bowen Shi, Paden Tomasello, Arun Babu, Sayani Kundu, Ali Elkahky, Zhaoheng Ni, Apoorv Vyas, Maryam Fazel-Zarandi, Alexei Baevski, Yossi Adi, Xiaohui Zhang, Wei-Ning Hsu, Alexis Conneau, Michael Auli.
 1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
@@ -473,7 +471,6 @@ conda install conda-forge::transformers
 1. **[Qwen2MoE](https://huggingface.co/docs/transformers/main/model_doc/qwen2_moe)** (from the Qwen team, Alibaba Group) released with the paper [blog post](https://qwenlm.github.io/blog/qwen-moe/) by Bo Zheng, Dayiheng Liu, Rui Men, Junyang Lin, Zhou San, Bowen Yu, An Yang, Mingfeng Xue, Fei Huang, Binyuan Hui, Mei Li, Tianyu Liu, Xingzhang Ren, Xuancheng Ren, Kexin Yang, Chang Zhou, Jingren Zhou.
 1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
-1. **[RecurrentGemma](https://huggingface.co/docs/transformers/main/model_doc/recurrent-gemma)** (from Google) released with the paper [RecurrentGemma: Moving Past Transformers for Efficient Open Language Models](https://storage.googleapis.com/deepmind-media/gemma/recurrentgemma-report.pdf) by the Griffin, RLHF and Gemma Teams.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
 1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Platforms) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.

README_te.md

@@ -325,10 +325,10 @@ Flax, PyTorch లేదా TensorFlow యొక్క ఇన్‌స్టా
 1. **[CLAP](https://huggingface.co/docs/transformers/model_doc/clap)** (from LAION-AI) released with the paper [Large-scale Contrastive Language-Audio Pretraining with Feature Fusion and Keyword-to-Caption Augmentation](https://arxiv.org/abs/2211.06687) by Yusong Wu, Ke Chen, Tianyu Zhang, Yuchen Hui, Taylor Berg-Kirkpatrick, Shlomo Dubnov.
 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. **[CLIPSeg](https://huggingface.co/docs/transformers/model_doc/clipseg)** (from University of Göttingen) released with the paper [Image Segmentation Using Text and Image Prompts](https://arxiv.org/abs/2112.10003) by Timo Lüddecke and Alexander Ecker.
-1. **[CLVP](https://huggingface.co/docs/transformers/model_doc/clvp)** released with the paper [Better speech synthesis through scaling](https://arxiv.org/abs/2305.07243) by James Betker.
+1. **[CLVP](https://huggingface.co/docs/transformers/model_doc/clvp)** released with the paper [Better speech synthesis through scaling](https://arxiv.org/abs/2305.07243) by James Betker. 
 1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
 1. **[CodeLlama](https://huggingface.co/docs/transformers/model_doc/llama_code)** (from MetaAI) released with the paper [Code Llama: Open Foundation Models for Code](https://ai.meta.com/research/publications/code-llama-open-foundation-models-for-code/) by Baptiste Rozière, Jonas Gehring, Fabian Gloeckle, Sten Sootla, Itai Gat, Xiaoqing Ellen Tan, Yossi Adi, Jingyu Liu, Tal Remez, Jérémy Rapin, Artyom Kozhevnikov, Ivan Evtimov, Joanna Bitton, Manish Bhatt, Cristian Canton Ferrer, Aaron Grattafiori, Wenhan Xiong, Alexandre Défossez, Jade Copet, Faisal Azhar, Hugo Touvron, Louis Martin, Nicolas Usunier, Thomas Scialom, Gabriel Synnaeve.
-1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (from Cohere) released with the paper [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere.
+1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (from Cohere) released with the paper [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere. 
 1. **[Conditional DETR](https://huggingface.co/docs/transformers/model_doc/conditional_detr)** (from Microsoft Research Asia) released with the paper [Conditional DETR for Fast Training Convergence](https://arxiv.org/abs/2108.06152) by Depu Meng, Xiaokang Chen, Zejia Fan, Gang Zeng, Houqiang Li, Yuhui Yuan, Lei Sun, Jingdong Wang.
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
 1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
@@ -386,13 +386,11 @@ Flax, PyTorch లేదా TensorFlow యొక్క ఇన్‌స్టా
 1. **[GPTBigCode](https://huggingface.co/docs/transformers/model_doc/gpt_bigcode)** (from BigCode) released with the paper [SantaCoder: don't reach for the stars!](https://arxiv.org/abs/2301.03988) by Loubna Ben Allal, Raymond Li, Denis Kocetkov, Chenghao Mou, Christopher Akiki, Carlos Munoz Ferrandis, Niklas Muennighoff, Mayank Mishra, Alex Gu, Manan Dey, Logesh Kumar Umapathi, Carolyn Jane Anderson, Yangtian Zi, Joel Lamy Poirier, Hailey Schoelkopf, Sergey Troshin, Dmitry Abulkhanov, Manuel Romero, Michael Lappert, Francesco De Toni, Bernardo García del Río, Qian Liu, Shamik Bose, Urvashi Bhattacharyya, Terry Yue Zhuo, Ian Yu, Paulo Villegas, Marco Zocca, Sourab Mangrulkar, David Lansky, Huu Nguyen, Danish Contractor, Luis Villa, Jia Li, Dzmitry Bahdanau, Yacine Jernite, Sean Hughes, Daniel Fried, Arjun Guha, Harm de Vries, Leandro von Werra.
 1. **[GPTSAN-japanese](https://huggingface.co/docs/transformers/model_doc/gptsan-japanese)** released in the repository [tanreinama/GPTSAN](https://github.com/tanreinama/GPTSAN/blob/main/report/model.md) by Toshiyuki Sakamoto(tanreinama).
 1. **[Graphormer](https://huggingface.co/docs/transformers/model_doc/graphormer)** (from Microsoft) released with the paper [Do Transformers Really Perform Bad for Graph Representation?](https://arxiv.org/abs/2106.05234) by Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu.
-1. **[Grounding DINO](https://huggingface.co/docs/transformers/main/model_doc/grounding-dino)** (from Institute for AI, Tsinghua-Bosch Joint Center for ML, Tsinghua University, IDEA Research and others) released with the paper [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://arxiv.org/abs/2303.05499) by Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang.
 1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
 1. **[HerBERT](https://huggingface.co/docs/transformers/model_doc/herbert)** (from Allegro.pl, AGH University of Science and Technology) released with the paper [KLEJ: Comprehensive Benchmark for Polish Language Understanding](https://www.aclweb.org/anthology/2020.acl-main.111.pdf) by Piotr Rybak, Robert Mroczkowski, Janusz Tracz, Ireneusz Gawlik.
 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. **[IDEFICS](https://huggingface.co/docs/transformers/model_doc/idefics)** (from HuggingFace) released with the paper [OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documents](https://huggingface.co/papers/2306.16527) by Hugo Laurençon, Lucile Saulnier, Léo Tronchon, Stas Bekman, Amanpreet Singh, Anton Lozhkov, Thomas Wang, Siddharth Karamcheti, Alexander M. Rush, Douwe Kiela, Matthieu Cord, Victor Sanh.
-1. **[Idefics2](https://huggingface.co/docs/transformers/main/model_doc/idefics2)** (from Hugging Face) released with the paper [IDEFICS2](https://huggingface.co/blog/idefics2) by Léo Tronchon, Hugo Laurencon, Victor Sanh.
 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. **[Informer](https://huggingface.co/docs/transformers/model_doc/informer)** (from Beihang University, UC Berkeley, Rutgers University, SEDD Company) released with the paper [Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting](https://arxiv.org/abs/2012.07436) by Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang.
 1. **[InstructBLIP](https://huggingface.co/docs/transformers/model_doc/instructblip)** (from Salesforce) released with the paper [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://arxiv.org/abs/2305.06500) by Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, Steven Hoi.
@@ -429,7 +427,7 @@ Flax, PyTorch లేదా TensorFlow యొక్క ఇన్‌స్టా
 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. **[MGP-STR](https://huggingface.co/docs/transformers/model_doc/mgp-str)** (from Alibaba Research) released with the paper [Multi-Granularity Prediction for Scene Text Recognition](https://arxiv.org/abs/2209.03592) by Peng Wang, Cheng Da, and Cong Yao.
 1. **[Mistral](https://huggingface.co/docs/transformers/model_doc/mistral)** (from Mistral AI) by The [Mistral AI](https://mistral.ai) team: Albert Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lélio Renard Lavaud, Lucile Saulnier, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix, William El Sayed.
-1. **[Mixtral](https://huggingface.co/docs/transformers/model_doc/mixtral)** (from Mistral AI) by The [Mistral AI](https://mistral.ai) team: Albert Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lélio Renard Lavaud, Lucile Saulnier, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix, William El Sayed.
+1. **[Mixtral](https://huggingface.co/docs/transformers/model_doc/mixtral)** (from Mistral AI) by The [Mistral AI](https://mistral.ai) team: Albert Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lélio Renard Lavaud, Lucile Saulnier, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix, William El Sayed. 
 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. **[MMS](https://huggingface.co/docs/transformers/model_doc/mms)** (from Facebook) released with the paper [Scaling Speech Technology to 1,000+ Languages](https://arxiv.org/abs/2305.13516) by Vineel Pratap, Andros Tjandra, Bowen Shi, Paden Tomasello, Arun Babu, Sayani Kundu, Ali Elkahky, Zhaoheng Ni, Apoorv Vyas, Maryam Fazel-Zarandi, Alexei Baevski, Yossi Adi, Xiaohui Zhang, Wei-Ning Hsu, Alexis Conneau, Michael Auli.
 1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
@@ -475,7 +473,6 @@ Flax, PyTorch లేదా TensorFlow యొక్క ఇన్‌స్టా
 1. **[Qwen2MoE](https://huggingface.co/docs/transformers/main/model_doc/qwen2_moe)** (from the Qwen team, Alibaba Group) released with the paper [blog post](https://qwenlm.github.io/blog/qwen-moe/) by Bo Zheng, Dayiheng Liu, Rui Men, Junyang Lin, Zhou San, Bowen Yu, An Yang, Mingfeng Xue, Fei Huang, Binyuan Hui, Mei Li, Tianyu Liu, Xingzhang Ren, Xuancheng Ren, Kexin Yang, Chang Zhou, Jingren Zhou.
 1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
-1. **[RecurrentGemma](https://huggingface.co/docs/transformers/main/model_doc/recurrent-gemma)** (from Google) released with the paper [RecurrentGemma: Moving Past Transformers for Efficient Open Language Models](https://storage.googleapis.com/deepmind-media/gemma/recurrentgemma-report.pdf) by the Griffin, RLHF and Gemma Teams.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
 1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Platforms) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.

README_vi.md

@@ -327,7 +327,7 @@ Số lượng điểm kiểm tra hiện tại: ![](https://img.shields.io/endpoi
 1. **[CLVP](https://huggingface.co/docs/transformers/model_doc/clvp)** được phát hành với bài báo [Better speech synthesis through scaling](https://arxiv.org/abs/2305.07243) by James Betker.
 1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (từ Salesforce) được phát hành với bài báo [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
 1. **[CodeLlama](https://huggingface.co/docs/transformers/model_doc/llama_code)** (từ MetaAI) được phát hành với bài báo [Code Llama: Open Foundation Models for Code](https://ai.meta.com/research/publications/code-llama-open-foundation-models-for-code/) by Baptiste Rozière, Jonas Gehring, Fabian Gloeckle, Sten Sootla, Itai Gat, Xiaoqing Ellen Tan, Yossi Adi, Jingyu Liu, Tal Remez, Jérémy Rapin, Artyom Kozhevnikov, Ivan Evtimov, Joanna Bitton, Manish Bhatt, Cristian Canton Ferrer, Aaron Grattafiori, Wenhan Xiong, Alexandre Défossez, Jade Copet, Faisal Azhar, Hugo Touvron, Louis Martin, Nicolas Usunier, Thomas Scialom, Gabriel Synnaeve.
-1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (từ Cohere) được phát hành với bài báo [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere.
+1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (từ Cohere) được phát hành với bài báo [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere. 
 1. **[Conditional DETR](https://huggingface.co/docs/transformers/model_doc/conditional_detr)** (từ Microsoft Research Asia) được phát hành với bài báo [Conditional DETR for Fast Training Convergence](https://arxiv.org/abs/2108.06152) by Depu Meng, Xiaokang Chen, Zejia Fan, Gang Zeng, Houqiang Li, Yuhui Yuan, Lei Sun, Jingdong Wang.
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (từ YituTech) được phát hành với bài báo [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
 1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (từ Facebook AI) được phát hành với bài báo [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.
@@ -385,13 +385,11 @@ Số lượng điểm kiểm tra hiện tại: ![](https://img.shields.io/endpoi
 1. **[GPTBigCode](https://huggingface.co/docs/transformers/model_doc/gpt_bigcode)** (từ BigCode) được phát hành với bài báo [SantaCoder: don't reach for the stars!](https://arxiv.org/abs/2301.03988) by Loubna Ben Allal, Raymond Li, Denis Kocetkov, Chenghao Mou, Christopher Akiki, Carlos Munoz Ferrandis, Niklas Muennighoff, Mayank Mishra, Alex Gu, Manan Dey, Logesh Kumar Umapathi, Carolyn Jane Anderson, Yangtian Zi, Joel Lamy Poirier, Hailey Schoelkopf, Sergey Troshin, Dmitry Abulkhanov, Manuel Romero, Michael Lappert, Francesco De Toni, Bernardo García del Río, Qian Liu, Shamik Bose, Urvashi Bhattacharyya, Terry Yue Zhuo, Ian Yu, Paulo Villegas, Marco Zocca, Sourab Mangrulkar, David Lansky, Huu Nguyen, Danish Contractor, Luis Villa, Jia Li, Dzmitry Bahdanau, Yacine Jernite, Sean Hughes, Daniel Fried, Arjun Guha, Harm de Vries, Leandro von Werra.
 1. **[GPTSAN-japanese](https://huggingface.co/docs/transformers/model_doc/gptsan-japanese)** released in the repository [tanreinama/GPTSAN](https://github.com/tanreinama/GPTSAN/blob/main/report/model.md) by Toshiyuki Sakamoto(tanreinama).
 1. **[Graphormer](https://huggingface.co/docs/transformers/model_doc/graphormer)** (từ Microsoft) được phát hành với bài báo [Do Transformers Really Perform Bad for Graph Representation?](https://arxiv.org/abs/2106.05234) by Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu.
-1. **[Grounding DINO](https://huggingface.co/docs/transformers/main/model_doc/grounding-dino)** (từ Institute for AI, Tsinghua-Bosch Joint Center for ML, Tsinghua University, IDEA Research and others) được phát hành với bài báo [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://arxiv.org/abs/2303.05499) by Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang.
 1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (từ UCSD, NVIDIA) được phát hành với bài báo [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
 1. **[HerBERT](https://huggingface.co/docs/transformers/model_doc/herbert)** (từ Allegro.pl, AGH University of Science and Technology) được phát hành với bài báo [KLEJ: Comprehensive Benchmark for Polish Language Understanding](https://www.aclweb.org/anthology/2020.acl-main.111.pdf) by Piotr Rybak, Robert Mroczkowski, Janusz Tracz, Ireneusz Gawlik.
 1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (từ Facebook) được phát hành với bài báo [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)** (từ Berkeley) được phát hành với bài báo [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. **[IDEFICS](https://huggingface.co/docs/transformers/model_doc/idefics)** (từ HuggingFace) được phát hành với bài báo [OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documents](https://huggingface.co/papers/2306.16527) by Hugo Laurençon, Lucile Saulnier, Léo Tronchon, Stas Bekman, Amanpreet Singh, Anton Lozhkov, Thomas Wang, Siddharth Karamcheti, Alexander M. Rush, Douwe Kiela, Matthieu Cord, Victor Sanh.
-1. **[Idefics2](https://huggingface.co/docs/transformers/main/model_doc/idefics2)** (từ Hugging Face) được phát hành với bài báo [IDEFICS2](https://huggingface.co/blog/idefics2) by Léo Tronchon, Hugo Laurencon, Victor Sanh.
 1. **[ImageGPT](https://huggingface.co/docs/transformers/model_doc/imagegpt)** (từ OpenAI) được phát hành với bài báo [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. **[Informer](https://huggingface.co/docs/transformers/model_doc/informer)** (từ Beihang University, UC Berkeley, Rutgers University, SEDD Company) được phát hành với bài báo [Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting](https://arxiv.org/abs/2012.07436) by Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang.
 1. **[InstructBLIP](https://huggingface.co/docs/transformers/model_doc/instructblip)** (từ Salesforce) được phát hành với bài báo [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://arxiv.org/abs/2305.06500) by Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, Steven Hoi.
@@ -474,7 +472,6 @@ Số lượng điểm kiểm tra hiện tại: ![](https://img.shields.io/endpoi
 1. **[Qwen2MoE](https://huggingface.co/docs/transformers/main/model_doc/qwen2_moe)** (từ the Qwen team, Alibaba Group) được phát hành với bài báo [blog post](https://qwenlm.github.io/blog/qwen-moe/) by Bo Zheng, Dayiheng Liu, Rui Men, Junyang Lin, Zhou San, Bowen Yu, An Yang, Mingfeng Xue, Fei Huang, Binyuan Hui, Mei Li, Tianyu Liu, Xingzhang Ren, Xuancheng Ren, Kexin Yang, Chang Zhou, Jingren Zhou.
 1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (từ Facebook) được phát hành với bài báo [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (từ Google Research) được phát hành với bài báo [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. **[RecurrentGemma](https://huggingface.co/docs/transformers/main/model_doc/recurrent-gemma)** (từ Google) được phát hành với bài báo [RecurrentGemma: Moving Past Transformers for Efficient Open Language Models](https://storage.googleapis.com/deepmind-media/gemma/recurrentgemma-report.pdf) by the Griffin, RLHF and Gemma Teams.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (từ Google Research) được phát hành với bài báo [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
 1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (từ META Platforms) được phát hành với bài báo [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (từ Google Research) được phát hành với bài báo [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.

README_zh-hans.md

@@ -335,13 +335,11 @@ conda install conda-forge::transformers
 1. **[GPTBigCode](https://huggingface.co/docs/transformers/model_doc/gpt_bigcode)** (来自 BigCode) 伴随论文 [SantaCoder: don't reach for the stars!](https://arxiv.org/abs/2301.03988) 由 Loubna Ben Allal, Raymond Li, Denis Kocetkov, Chenghao Mou, Christopher Akiki, Carlos Munoz Ferrandis, Niklas Muennighoff, Mayank Mishra, Alex Gu, Manan Dey, Logesh Kumar Umapathi, Carolyn Jane Anderson, Yangtian Zi, Joel Lamy Poirier, Hailey Schoelkopf, Sergey Troshin, Dmitry Abulkhanov, Manuel Romero, Michael Lappert, Francesco De Toni, Bernardo García del Río, Qian Liu, Shamik Bose, Urvashi Bhattacharyya, Terry Yue Zhuo, Ian Yu, Paulo Villegas, Marco Zocca, Sourab Mangrulkar, David Lansky, Huu Nguyen, Danish Contractor, Luis Villa, Jia Li, Dzmitry Bahdanau, Yacine Jernite, Sean Hughes, Daniel Fried, Arjun Guha, Harm de Vries, Leandro von Werra 发布。
 1. **[GPTSAN-japanese](https://huggingface.co/docs/transformers/model_doc/gptsan-japanese)** released in the repository [tanreinama/GPTSAN](https://github.com/tanreinama/GPTSAN/blob/main/report/model.md) by 坂本俊之(tanreinama).
 1. **[Graphormer](https://huggingface.co/docs/transformers/model_doc/graphormer)** (from Microsoft) released with the paper [Do Transformers Really Perform Bad for Graph Representation?](https://arxiv.org/abs/2106.05234) by Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu.
-1. **[Grounding DINO](https://huggingface.co/docs/transformers/main/model_doc/grounding-dino)** (来自 Institute for AI, Tsinghua-Bosch Joint Center for ML, Tsinghua University, IDEA Research and others) 伴随论文 [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://arxiv.org/abs/2303.05499) 由 Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang 发布。
 1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (来自 UCSD, NVIDIA) 伴随论文 [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) 由 Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang 发布。
 1. **[HerBERT](https://huggingface.co/docs/transformers/model_doc/herbert)** (来自 Allegro.pl, AGH University of Science and Technology) 伴随论文 [KLEJ: Comprehensive Benchmark for Polish Language Understanding](https://www.aclweb.org/anthology/2020.acl-main.111.pdf) 由 Piotr Rybak, Robert Mroczkowski, Janusz Tracz, Ireneusz Gawlik 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. **[IDEFICS](https://huggingface.co/docs/transformers/model_doc/idefics)** (from HuggingFace) released with the paper [OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documents](https://huggingface.co/papers/2306.16527) by Hugo Laurençon, Lucile Saulnier, Léo Tronchon, Stas Bekman, Amanpreet Singh, Anton Lozhkov, Thomas Wang, Siddharth Karamcheti, Alexander M. Rush, Douwe Kiela, Matthieu Cord, Victor Sanh.
-1. **[Idefics2](https://huggingface.co/docs/transformers/main/model_doc/idefics2)** (来自 Hugging Face) 伴随论文 [IDEFICS2](https://huggingface.co/blog/idefics2) 由 Léo Tronchon, Hugo Laurencon, Victor Sanh 发布。
 1. **[ImageGPT](https://huggingface.co/docs/transformers/model_doc/imagegpt)** (来自 OpenAI) 伴随论文 [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) 由 Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever 发布。
 1. **[Informer](https://huggingface.co/docs/transformers/model_doc/informer)** (from Beihang University, UC Berkeley, Rutgers University, SEDD Company) released with the paper [Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting](https://arxiv.org/abs/2012.07436) by Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang.
 1. **[InstructBLIP](https://huggingface.co/docs/transformers/model_doc/instructblip)** (来自 Salesforce) 伴随论文 [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://arxiv.org/abs/2305.06500) 由 Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, Steven Hoi 发布。
@@ -424,7 +422,6 @@ conda install conda-forge::transformers
 1. **[Qwen2MoE](https://huggingface.co/docs/transformers/main/model_doc/qwen2_moe)** (来自 the Qwen team, Alibaba Group) 伴随论文 [blog post](https://qwenlm.github.io/blog/qwen-moe/) by Bo Zheng, Dayiheng Liu, Rui Men, Junyang Lin, Zhou San, Bowen Yu, An Yang, Mingfeng Xue, Fei Huang, Binyuan Hui, Mei Li, Tianyu Liu, Xingzhang Ren, Xuancheng Ren, Kexin Yang, Chang Zhou, Jingren Zhou 发布.
 1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (来自 Facebook) 伴随论文 [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) 由 Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela 发布。
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (来自 Google Research) 伴随论文 [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) 由 Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang 发布。
-1. **[RecurrentGemma](https://huggingface.co/docs/transformers/main/model_doc/recurrent-gemma)** (来自 Google) 伴随论文 [RecurrentGemma: Moving Past Transformers for Efficient Open Language Models](https://storage.googleapis.com/deepmind-media/gemma/recurrentgemma-report.pdf) 由 the Griffin, RLHF and Gemma Teams 发布。
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (来自 Google Research) 伴随论文 [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) 由 Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya 发布。
 1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Research) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (来自 Google Research) 伴随论文 [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) 由 Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder 发布。

README_zh-hant.md

@@ -289,7 +289,7 @@ conda install conda-forge::transformers
 1. **[CLVP](https://huggingface.co/docs/transformers/model_doc/clvp)** released with the paper [Better speech synthesis through scaling](https://arxiv.org/abs/2305.07243) by James Betker.
 1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
 1. **[CodeLlama](https://huggingface.co/docs/transformers/model_doc/llama_code)** (from MetaAI) released with the paper [Code Llama: Open Foundation Models for Code](https://ai.meta.com/research/publications/code-llama-open-foundation-models-for-code/) by Baptiste Rozière, Jonas Gehring, Fabian Gloeckle, Sten Sootla, Itai Gat, Xiaoqing Ellen Tan, Yossi Adi, Jingyu Liu, Tal Remez, Jérémy Rapin, Artyom Kozhevnikov, Ivan Evtimov, Joanna Bitton, Manish Bhatt, Cristian Canton Ferrer, Aaron Grattafiori, Wenhan Xiong, Alexandre Défossez, Jade Copet, Faisal Azhar, Hugo Touvron, Louis Martin, Nicolas Usunier, Thomas Scialom, Gabriel Synnaeve.
-1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (from Cohere) released with the paper [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere.
+1. **[Cohere](https://huggingface.co/docs/transformers/model_doc/cohere)** (from Cohere) released with the paper [Command-R: Retrieval Augmented Generation at Production Scale](<https://txt.cohere.com/command-r/>) by Cohere. 
 1. **[Conditional DETR](https://huggingface.co/docs/transformers/model_doc/conditional_detr)** (from Microsoft Research Asia) released with the paper [Conditional DETR for Fast Training Convergence](https://arxiv.org/abs/2108.06152) by Depu Meng, Xiaokang Chen, Zejia Fan, Gang Zeng, Houqiang Li, Yuhui Yuan, Lei Sun, Jingdong Wang.
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
 1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
@@ -347,13 +347,11 @@ conda install conda-forge::transformers
 1. **[GPTBigCode](https://huggingface.co/docs/transformers/model_doc/gpt_bigcode)** (from BigCode) released with the paper [SantaCoder: don't reach for the stars!](https://arxiv.org/abs/2301.03988) by Loubna Ben Allal, Raymond Li, Denis Kocetkov, Chenghao Mou, Christopher Akiki, Carlos Munoz Ferrandis, Niklas Muennighoff, Mayank Mishra, Alex Gu, Manan Dey, Logesh Kumar Umapathi, Carolyn Jane Anderson, Yangtian Zi, Joel Lamy Poirier, Hailey Schoelkopf, Sergey Troshin, Dmitry Abulkhanov, Manuel Romero, Michael Lappert, Francesco De Toni, Bernardo García del Río, Qian Liu, Shamik Bose, Urvashi Bhattacharyya, Terry Yue Zhuo, Ian Yu, Paulo Villegas, Marco Zocca, Sourab Mangrulkar, David Lansky, Huu Nguyen, Danish Contractor, Luis Villa, Jia Li, Dzmitry Bahdanau, Yacine Jernite, Sean Hughes, Daniel Fried, Arjun Guha, Harm de Vries, Leandro von Werra.
 1. **[GPTSAN-japanese](https://huggingface.co/docs/transformers/model_doc/gptsan-japanese)** released in the repository [tanreinama/GPTSAN](https://github.com/tanreinama/GPTSAN/blob/main/report/model.md) by 坂本俊之(tanreinama).
 1. **[Graphormer](https://huggingface.co/docs/transformers/model_doc/graphormer)** (from Microsoft) released with the paper [Do Transformers Really Perform Bad for Graph Representation?](https://arxiv.org/abs/2106.05234) by Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu.
-1. **[Grounding DINO](https://huggingface.co/docs/transformers/main/model_doc/grounding-dino)** (from Institute for AI, Tsinghua-Bosch Joint Center for ML, Tsinghua University, IDEA Research and others) released with the paper [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://arxiv.org/abs/2303.05499) by Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang.
 1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
 1. **[HerBERT](https://huggingface.co/docs/transformers/model_doc/herbert)** (from Allegro.pl, AGH University of Science and Technology) released with the paper [KLEJ: Comprehensive Benchmark for Polish Language Understanding](https://www.aclweb.org/anthology/2020.acl-main.111.pdf) by Piotr Rybak, Robert Mroczkowski, Janusz Tracz, Ireneusz Gawlik.
 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. **[IDEFICS](https://huggingface.co/docs/transformers/model_doc/idefics)** (from HuggingFace) released with the paper [OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documents](https://huggingface.co/papers/2306.16527) by Hugo Laurençon, Lucile Saulnier, Léo Tronchon, Stas Bekman, Amanpreet Singh, Anton Lozhkov, Thomas Wang, Siddharth Karamcheti, Alexander M. Rush, Douwe Kiela, Matthieu Cord, Victor Sanh.
-1. **[Idefics2](https://huggingface.co/docs/transformers/main/model_doc/idefics2)** (from Hugging Face) released with the paper [IDEFICS2](https://huggingface.co/blog/idefics2) by Léo Tronchon, Hugo Laurencon, Victor Sanh.
 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. **[Informer](https://huggingface.co/docs/transformers/model_doc/informer)** (from Beihang University, UC Berkeley, Rutgers University, SEDD Company) released with the paper [Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting](https://arxiv.org/abs/2012.07436) by Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang.
 1. **[InstructBLIP](https://huggingface.co/docs/transformers/model_doc/instructblip)** (from Salesforce) released with the paper [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://arxiv.org/abs/2305.06500) by Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, Steven Hoi.
@@ -430,13 +428,12 @@ conda install conda-forge::transformers
 1. **[Pop2Piano](https://huggingface.co/docs/transformers/model_doc/pop2piano)** released with the paper [Pop2Piano : Pop Audio-based Piano Cover Generation](https://arxiv.org/abs/2211.00895) by Jongho Choi, Kyogu Lee.
 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. **[PVT](https://huggingface.co/docs/transformers/model_doc/pvt)** (from Nanjing University, The University of Hong Kong etc.) released with the paper [Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions](https://arxiv.org/pdf/2102.12122.pdf) by Wenhai Wang, Enze Xie, Xiang Li, Deng-Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, Ling Shao.
-1. **[PVTv2](https://huggingface.co/docs/transformers/model_doc/pvt_v2)** (from Shanghai AI Laboratory, Nanjing University, The University of Hong Kong etc.) released with the paper [PVT v2: Improved Baselines with Pyramid Vision Transformer](https://arxiv.org/abs/2106.13797) by Wenhai Wang, Enze Xie, Xiang Li, Deng-Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, Ling Shao.
+1. **[PVTv2](https://huggingface.co/docs/transformers/model_doc/pvt_v2)** (from Shanghai AI Laboratory, Nanjing University, The University of Hong Kong etc.) released with the paper [PVT v2: Improved Baselines with Pyramid Vision Transformer](https://arxiv.org/abs/2106.13797) by Wenhai Wang, Enze Xie, Xiang Li, Deng-Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, Ling Shao. 
 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. **[Qwen2](https://huggingface.co/docs/transformers/model_doc/qwen2)** (from the Qwen team, Alibaba Group) released with the paper [Qwen Technical Report](https://arxiv.org/abs/2309.16609) by Jinze Bai, Shuai Bai, Yunfei Chu, Zeyu Cui, Kai Dang, Xiaodong Deng, Yang Fan, Wenbin Ge, Yu Han, Fei Huang, Binyuan Hui, Luo Ji, Mei Li, Junyang Lin, Runji Lin, Dayiheng Liu, Gao Liu, Chengqiang Lu, Keming Lu, Jianxin Ma, Rui Men, Xingzhang Ren, Xuancheng Ren, Chuanqi Tan, Sinan Tan, Jianhong Tu, Peng Wang, Shijie Wang, Wei Wang, Shengguang Wu, Benfeng Xu, Jin Xu, An Yang, Hao Yang, Jian Yang, Shusheng Yang, Yang Yao, Bowen Yu, Hongyi Yuan, Zheng Yuan, Jianwei Zhang, Xingxuan Zhang, Yichang Zhang, Zhenru Zhang, Chang Zhou, Jingren Zhou, Xiaohuan Zhou and Tianhang Zhu.
 1. **[Qwen2MoE](https://huggingface.co/docs/transformers/main/model_doc/qwen2_moe)** (from the Qwen team, Alibaba Group) released with the paper [blog post](https://qwenlm.github.io/blog/qwen-moe/) by Bo Zheng, Dayiheng Liu, Rui Men, Junyang Lin, Zhou San, Bowen Yu, An Yang, Mingfeng Xue, Fei Huang, Binyuan Hui, Mei Li, Tianyu Liu, Xingzhang Ren, Xuancheng Ren, Kexin Yang, Chang Zhou, Jingren Zhou.
 1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
-1. **[RecurrentGemma](https://huggingface.co/docs/transformers/main/model_doc/recurrent-gemma)** (from Google) released with the paper [RecurrentGemma: Moving Past Transformers for Efficient Open Language Models](https://storage.googleapis.com/deepmind-media/gemma/recurrentgemma-report.pdf) by the Griffin, RLHF and Gemma Teams.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
 1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Research) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
@@ -459,7 +456,7 @@ conda install conda-forge::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. **[StableLm](https://huggingface.co/docs/transformers/model_doc/stablelm)**  released with the paper [StableLM 3B 4E1T (Technical Report)](https://stability.wandb.io/stability-llm/stable-lm/reports/StableLM-3B-4E1T--VmlldzoyMjU4?accessToken=u3zujipenkx5g7rtcj9qojjgxpconyjktjkli2po09nffrffdhhchq045vp0wyfo) by  Jonathan Tow, Marco Bellagente, Dakota Mahan, Carlos Riquelme Ruiz, Duy Phung, Maksym Zhuravinskyi, Nathan Cooper, Nikhil Pinnaparaju, Reshinth Adithyan, and James Baicoianu.
+1. **[StableLm](https://huggingface.co/docs/transformers/model_doc/stablelm)**  released with the paper [StableLM 3B 4E1T (Technical Report)](https://stability.wandb.io/stability-llm/stable-lm/reports/StableLM-3B-4E1T--VmlldzoyMjU4?accessToken=u3zujipenkx5g7rtcj9qojjgxpconyjktjkli2po09nffrffdhhchq045vp0wyfo) by  Jonathan Tow, Marco Bellagente, Dakota Mahan, Carlos Riquelme Ruiz, Duy Phung, Maksym Zhuravinskyi, Nathan Cooper, Nikhil Pinnaparaju, Reshinth Adithyan, and James Baicoianu. 
 1. **[Starcoder2](https://huggingface.co/docs/transformers/model_doc/starcoder2)** (from BigCode team) released with the paper [StarCoder 2 and The Stack v2: The Next Generation](https://arxiv.org/abs/2402.19173) by Anton Lozhkov, Raymond Li, Loubna Ben Allal, Federico Cassano, Joel Lamy-Poirier, Nouamane Tazi, Ao Tang, Dmytro Pykhtar, Jiawei Liu, Yuxiang Wei, Tianyang Liu, Max Tian, Denis Kocetkov, Arthur Zucker, Younes Belkada, Zijian Wang, Qian Liu, Dmitry Abulkhanov, Indraneil Paul, Zhuang Li, Wen-Ding Li, Megan Risdal, Jia Li, Jian Zhu, Terry Yue Zhuo, Evgenii Zheltonozhskii, Nii Osae Osae Dade, Wenhao Yu, Lucas Krauß, Naman Jain, Yixuan Su, Xuanli He, Manan Dey, Edoardo Abati, Yekun Chai, Niklas Muennighoff, Xiangru Tang, Muhtasham Oblokulov, Christopher Akiki, Marc Marone, Chenghao Mou, Mayank Mishra, Alex Gu, Binyuan Hui, Tri Dao, Armel Zebaze, Olivier Dehaene, Nicolas Patry, Canwen Xu, Julian McAuley, Han Hu, Torsten Scholak, Sebastien Paquet, Jennifer Robinson, Carolyn Jane Anderson, Nicolas Chapados, Mostofa Patwary, Nima Tajbakhsh, Yacine Jernite, Carlos Muñoz Ferrandis, Lingming Zhang, Sean Hughes, Thomas Wolf, Arjun Guha, Leandro von Werra, and Harm de Vries.
 1. **[SuperPoint](https://huggingface.co/docs/transformers/model_doc/superpoint)** (from MagicLeap) released with the paper [SuperPoint: Self-Supervised Interest Point Detection and Description](https://arxiv.org/abs/1712.07629) by Daniel DeTone, Tomasz Malisiewicz and Andrew Rabinovich.
 1. **[SwiftFormer](https://huggingface.co/docs/transformers/model_doc/swiftformer)** (from MBZUAI) released with the paper [SwiftFormer: Efficient Additive Attention for Transformer-based Real-time Mobile Vision Applications](https://arxiv.org/abs/2303.15446) by Abdelrahman Shaker, Muhammad Maaz, Hanoona Rasheed, Salman Khan, Ming-Hsuan Yang, Fahad Shahbaz Khan.

docker/transformers-all-latest-gpu/Dockerfile

@@ -46,15 +46,11 @@ RUN python3 -m pip install --no-cache-dir git+https://github.com/huggingface/opt
 RUN python3 -m pip install --no-cache-dir decord av==9.2.0
 
 # For `dinat` model
-# The `XXX` part in `torchXXX` needs to match `PYTORCH` (to some extent)
-RUN python3 -m pip install --no-cache-dir natten==0.15.1+torch220$CUDA -f https://shi-labs.com/natten/wheels
+RUN python3 -m pip install --no-cache-dir 'natten<0.15.0' -f https://shi-labs.com/natten/wheels/$CUDA/
 
 # For `nougat` tokenizer
 RUN python3 -m pip install --no-cache-dir python-Levenshtein
 
-# For `FastSpeech2ConformerTokenizer` tokenizer
-RUN python3 -m pip install --no-cache-dir g2p-en
-
 # 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-quantization-latest-gpu/Dockerfile

@@ -9,7 +9,7 @@ SHELL ["sh", "-lc"]
 # The following `ARG` are mainly used to specify the versions explicitly & directly in this docker file, and not meant
 # to be used as arguments for docker build (so far).
 
-ARG PYTORCH='2.2.1'
+ARG PYTORCH='2.2.0'
 # Example: `cu102`, `cu113`, etc.
 ARG CUDA='cu118'
 
@@ -30,9 +30,6 @@ RUN python3 -m pip install --no-cache-dir -e ./transformers[dev-torch]
 
 RUN python3 -m pip install --no-cache-dir git+https://github.com/huggingface/accelerate@main#egg=accelerate
 
-# needed in bnb and awq
-RUN python3 -m pip install --no-cache-dir einops
-
 # Add bitsandbytes for mixed int8 testing
 RUN python3 -m pip install --no-cache-dir bitsandbytes
 
@@ -46,8 +43,7 @@ RUN python3 -m pip install --no-cache-dir git+https://github.com/huggingface/opt
 RUN python3 -m pip install --no-cache-dir aqlm[gpu]==1.0.2
 
 # Add autoawq for quantization testing
-# >=v0.2.3 needed for compatibility with torch 2.2.1
-RUN python3 -m pip install --no-cache-dir https://github.com/casper-hansen/AutoAWQ/releases/download/v0.2.3/autoawq-0.2.3+cu118-cp38-cp38-linux_x86_64.whl
+RUN python3 -m pip install --no-cache-dir https://github.com/casper-hansen/AutoAWQ/releases/download/v0.2.0/autoawq-0.2.0+cu118-cp38-cp38-linux_x86_64.whl
 
 # Add quanto for quantization testing
 RUN python3 -m pip install --no-cache-dir quanto

docs/source/de/add_new_pipeline.md

@@ -208,10 +208,14 @@ from transformers import pipeline
 classifier = pipeline("pair-classification", model="sgugger/finetuned-bert-mrpc")
 ```
 
-Dann können wir sie auf dem Hub mit der Methode `push_to_hub` freigeben:
+Dann können wir sie auf dem Hub mit der Methode `save_pretrained` in einem `Repository` freigeben:
 
 ```py
-classifier.push_to_hub("test-dynamic-pipeline")
+from huggingface_hub import Repository
+
+repo = Repository("test-dynamic-pipeline", clone_from="{your_username}/test-dynamic-pipeline")
+classifier.save_pretrained("test-dynamic-pipeline")
+repo.push_to_hub()
 ```
 
 Dadurch wird die Datei, in der Sie `PairClassificationPipeline` definiert haben, in den Ordner `"test-dynamic-pipeline"` kopiert,

docs/source/en/_toctree.yml

@@ -468,8 +468,6 @@
         title: RAG
       - local: model_doc/realm
         title: REALM
-      - local: model_doc/recurrent_gemma
-        title: RecurrentGemma
       - local: model_doc/reformer
         title: Reformer
       - local: model_doc/rembert
@@ -732,14 +730,10 @@
         title: FLAVA
       - local: model_doc/git
         title: GIT
-      - local: model_doc/grounding-dino
-        title: Grounding DINO
       - local: model_doc/groupvit
         title: GroupViT
       - local: model_doc/idefics
         title: IDEFICS
-      - local: model_doc/idefics2
-        title: Idefics2
       - local: model_doc/instructblip
         title: InstructBLIP
       - local: model_doc/kosmos-2

docs/source/en/add_new_pipeline.md

@@ -208,10 +208,14 @@ from transformers import pipeline
 classifier = pipeline("pair-classification", model="sgugger/finetuned-bert-mrpc")
 ```
 
-Then we can share it on the Hub by using the `push_to_hub` method:
+Then we can share it on the Hub by using the `save_pretrained` method in a `Repository`:
 
 ```py
-classifier.push_to_hub("test-dynamic-pipeline")
+from huggingface_hub import Repository
+
+repo = Repository("test-dynamic-pipeline", clone_from="{your_username}/test-dynamic-pipeline")
+classifier.save_pretrained("test-dynamic-pipeline")
+repo.push_to_hub()
 ```
 
 This will copy the file where you defined `PairClassificationPipeline` inside the folder `"test-dynamic-pipeline"`,

docs/source/en/generation_strategies.md

@@ -57,10 +57,9 @@ When you load a model explicitly, you can inspect the generation configuration t
 >>> model = AutoModelForCausalLM.from_pretrained("distilbert/distilgpt2")
 >>> model.generation_config
 GenerationConfig {
-  "bos_token_id": 50256,
-  "eos_token_id": 50256
+    "bos_token_id": 50256,
+    "eos_token_id": 50256,
 }
-<BLANKLINE>
 ```
 
 Printing out the `model.generation_config` reveals only the values that are different from the default generation
@@ -245,7 +244,8 @@ To enable multinomial sampling set `do_sample=True` and `num_beams=1`.
 
 >>> outputs = model.generate(**inputs, do_sample=True, num_beams=1, max_new_tokens=100)
 >>> tokenizer.batch_decode(outputs, skip_special_tokens=True)
-["Today was an amazing day because we received these wonderful items by the way of a gift shop. The box arrived on a Thursday and I opened it on Monday afternoon to receive the gifts. Both bags featured pieces from all the previous years!\n\nThe box had lots of surprises in it, including some sweet little mini chocolate chips! I don't think I'd eat all of these. This was definitely one of the most expensive presents I have ever got, I actually got most of them for free!\n\nThe first package came"]
+['Today was an amazing day because when you go to the World Cup and you don\'t, or when you don\'t get invited,
+that\'s a terrible feeling."']
 ```
 
 ### Beam-search decoding
@@ -393,7 +393,7 @@ just like in multinomial sampling. However, in assisted decoding, reducing the t
 >>> assistant_model = AutoModelForCausalLM.from_pretrained(assistant_checkpoint)
 >>> outputs = model.generate(**inputs, assistant_model=assistant_model, do_sample=True, temperature=0.5)
 >>> tokenizer.batch_decode(outputs, skip_special_tokens=True)
-['Alice and Bob, a couple of friends of mine, who are both in the same office as']
+['Alice and Bob are going to the same party. It is a small party, in a small']
 ```
 
 Alternativelly, you can also set the `prompt_lookup_num_tokens` to trigger n-gram based assisted decoding, as opposed

docs/source/en/index.md

@@ -154,13 +154,11 @@ Flax), PyTorch, and/or TensorFlow.
 |                   [GPTBigCode](model_doc/gpt_bigcode)                    |       ✅        |         ❌         |      ❌      |
 |               [GPTSAN-japanese](model_doc/gptsan-japanese)               |       ✅        |         ❌         |      ❌      |
 |                    [Graphormer](model_doc/graphormer)                    |       ✅        |         ❌         |      ❌      |
-|                [Grounding DINO](model_doc/grounding-dino)                |       ✅        |         ❌         |      ❌      |
 |                      [GroupViT](model_doc/groupvit)                      |       ✅        |         ✅         |      ❌      |
 |                       [HerBERT](model_doc/herbert)                       |       ✅        |         ✅         |      ✅      |
 |                        [Hubert](model_doc/hubert)                        |       ✅        |         ✅         |      ❌      |
 |                        [I-BERT](model_doc/ibert)                         |       ✅        |         ❌         |      ❌      |
 |                       [IDEFICS](model_doc/idefics)                       |       ✅        |         ❌         |      ❌      |
-|                      [Idefics2](model_doc/idefics2)                      |       ✅        |         ❌         |      ❌      |
 |                      [ImageGPT](model_doc/imagegpt)                      |       ✅        |         ❌         |      ❌      |
 |                      [Informer](model_doc/informer)                      |       ✅        |         ❌         |      ❌      |
 |                  [InstructBLIP](model_doc/instructblip)                  |       ✅        |         ❌         |      ❌      |
@@ -245,7 +243,6 @@ Flax), PyTorch, and/or TensorFlow.
 |                     [Qwen2MoE](model_doc/qwen2_moe)                      |       ✅        |         ❌         |      ❌      |
 |                           [RAG](model_doc/rag)                           |       ✅        |         ✅         |      ❌      |
 |                         [REALM](model_doc/realm)                         |       ✅        |         ❌         |      ❌      |
-|               [RecurrentGemma](model_doc/recurrent_gemma)                |       ✅        |         ❌         |      ❌      |
 |                      [Reformer](model_doc/reformer)                      |       ✅        |         ❌         |      ❌      |
 |                        [RegNet](model_doc/regnet)                        |       ✅        |         ✅         |      ✅      |
 |                       [RemBERT](model_doc/rembert)                       |       ✅        |         ✅         |      ❌      |

docs/source/en/model_doc/code_llama.md

@@ -65,9 +65,9 @@ After conversion, the model and tokenizer can be loaded via:
 >>> tokenizer = CodeLlamaTokenizer.from_pretrained("codellama/CodeLlama-7b-hf")
 >>> model = LlamaForCausalLM.from_pretrained("codellama/CodeLlama-7b-hf")
 >>> PROMPT = '''def remove_non_ascii(s: str) -> str:
-...     """ <FILL_ME>
-...     return result
-... '''
+    """ <FILL_ME>
+    return result
+'''
 >>> input_ids = tokenizer(PROMPT, return_tensors="pt")["input_ids"]
 >>> generated_ids = model.generate(input_ids, max_new_tokens=128)
 
@@ -75,10 +75,10 @@ After conversion, the model and tokenizer can be loaded via:
 >>> print(PROMPT.replace("<FILL_ME>", filling))
 def remove_non_ascii(s: str) -> str:
     """ Remove non-ASCII characters from a string.
-<BLANKLINE>
+
     Args:
         s: The string to remove non-ASCII characters from.
-<BLANKLINE>
+
     Returns:
         The string with non-ASCII characters removed.
     """
@@ -87,7 +87,6 @@ def remove_non_ascii(s: str) -> str:
         if ord(c) < 128:
             result += c
     return result
-<BLANKLINE>
 ```
 
 If you only want the infilled part:
@@ -96,8 +95,7 @@ If you only want the infilled part:
 >>> import torch
 
 >>> generator = pipeline("text-generation",model="codellama/CodeLlama-7b-hf",torch_dtype=torch.float16, device_map="auto")
->>> generator('def remove_non_ascii(s: str) -> str:\n    """ <FILL_ME>\n    return result', max_new_tokens = 128)
-[{'generated_text': 'def remove_non_ascii(s: str) -> str:\n    """ <FILL_ME>\n    return resultRemove non-ASCII characters from a string. """\n    result = ""\n    for c in s:\n        if ord(c) < 128:\n            result += c'}]
+>>> generator('def remove_non_ascii(s: str) -> str:\n    """ <FILL_ME>\n    return result', max_new_tokens = 128, return_type = 1)
 ```
 
 Under the hood, the tokenizer [automatically splits by `<FILL_ME>`](https://huggingface.co/docs/transformers/main/model_doc/code_llama#transformers.CodeLlamaTokenizer.fill_token) to create a formatted input string that follows [the original training pattern](https://github.com/facebookresearch/codellama/blob/cb51c14ec761370ba2e2bc351374a79265d0465e/llama/generation.py#L402). This is more robust than preparing the pattern yourself: it avoids pitfalls, such as token glueing, that are very hard to debug.  To see how much CPU and GPU memory you need for this model or others, try [this calculator](https://huggingface.co/spaces/hf-accelerate/model-memory-usage) which can help determine that value.

docs/source/en/model_doc/grounding-dino.md

@@ -1,97 +0,0 @@
-<!--Copyright 2024 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
-
-⚠️ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
-rendered properly in your Markdown viewer.
-
--->
-
-# Grounding DINO
-
-## Overview
-
-The Grounding DINO model was proposed in [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://arxiv.org/abs/2303.05499) by Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang. Grounding DINO extends a closed-set object detection model with a text encoder, enabling open-set object detection. The model achieves remarkable results, such as 52.5 AP on COCO zero-shot.
-
-The abstract from the paper is the following:
-
-*In this paper, we present an open-set object detector, called Grounding DINO, by marrying Transformer-based detector DINO with grounded pre-training, which can detect arbitrary objects with human inputs such as category names or referring expressions. The key solution of open-set object detection is introducing language to a closed-set detector for open-set concept generalization. To effectively fuse language and vision modalities, we conceptually divide a closed-set detector into three phases and propose a tight fusion solution, which includes a feature enhancer, a language-guided query selection, and a cross-modality decoder for cross-modality fusion. While previous works mainly evaluate open-set object detection on novel categories, we propose to also perform evaluations on referring expression comprehension for objects specified with attributes. Grounding DINO performs remarkably well on all three settings, including benchmarks on COCO, LVIS, ODinW, and RefCOCO/+/g. Grounding DINO achieves a 52.5 AP on the COCO detection zero-shot transfer benchmark, i.e., without any training data from COCO. It sets a new record on the ODinW zero-shot benchmark with a mean 26.1 AP.*
-
-<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/model_doc/grouding_dino_architecture.png"
-alt="drawing" width="600"/>
-
-<small> Grounding DINO overview. Taken from the <a href="https://arxiv.org/abs/2303.05499">original paper</a>. </small>
-
-This model was contributed by [EduardoPacheco](https://huggingface.co/EduardoPacheco) and [nielsr](https://huggingface.co/nielsr).
-The original code can be found [here](https://github.com/IDEA-Research/GroundingDINO).
-
-## Usage tips
-
-- One can use [`GroundingDinoProcessor`] to prepare image-text pairs for the model.
-- To separate classes in the text use a period e.g. "a cat. a dog."
-- When using multiple classes (e.g. `"a cat. a dog."`), use `post_process_grounded_object_detection` from [`GroundingDinoProcessor`] to post process outputs. Since, the labels returned from `post_process_object_detection` represent the indices from the model dimension where prob > threshold.
-
-Here's how to use the model for zero-shot object detection:
-
-```python
-import requests
-
-import torch
-from PIL import Image
-from transformers import AutoProcessor, AutoModelForZeroShotObjectDetection, 
-
-model_id = "IDEA-Research/grounding-dino-tiny"
-
-processor = AutoProcessor.from_pretrained(model_id)
-model = AutoModelForZeroShotObjectDetection.from_pretrained(model_id).to(device)
-
-image_url = "http://images.cocodataset.org/val2017/000000039769.jpg"
-image = Image.open(requests.get(image_url, stream=True).raw)
-# Check for cats and remote controls
-text = "a cat. a remote control."
-
-inputs = processor(images=image, text=text, return_tensors="pt").to(device)
-with torch.no_grad():
-    outputs = model(**inputs)
-
-results = processor.post_process_grounded_object_detection(
-    outputs,
-    inputs.input_ids,
-    box_threshold=0.4,
-    text_threshold=0.3,
-    target_sizes=[image.size[::-1]]
-)
-```
-
-
-## GroundingDinoImageProcessor
-
-[[autodoc]] GroundingDinoImageProcessor
-    - preprocess
-    - post_process_object_detection
-
-## GroundingDinoProcessor
-
-[[autodoc]] GroundingDinoProcessor
-    - post_process_grounded_object_detection
-
-## GroundingDinoConfig
-
-[[autodoc]] GroundingDinoConfig
-
-## GroundingDinoModel
-
-[[autodoc]] GroundingDinoModel
-    - forward
-
-## GroundingDinoForObjectDetection
-
-[[autodoc]] GroundingDinoForObjectDetection
-    - forward

docs/source/en/model_doc/idefics2.md

@@ -1,98 +0,0 @@
-<!--Copyright 2024 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
-
-⚠️ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
-rendered properly in your Markdown viewer.
-
--->
-
-# Idefics2
-
-## Overview
-
-The Idefics2 model was created by the [Hugging Face M4](https://huggingface.co/HuggingFaceM4) team and authored by Léo Tronchon, Hugo Laurencon, Victor Sanh.
-The accompanying blog post can be found [here](https://huggingface.co/blog/idefics2).
-
-Idefics2 is an open multimodal model that accepts arbitrary sequences of image and text inputs and produces text
-outputs. The model can answer questions about images, describe visual content, create stories grounded on multiple
-images, or simply behave as a pure language model without visual inputs. It improves upon IDEFICS-1, notably on
-document understanding, OCR, or visual reasoning. Idefics2 is lightweight (8 billion parameters) and treats
-images in their native aspect ratio and resolution, which allows for varying inference efficiency.
-
-Tips:
-- Each sample can contain multiple images, and the number of images can vary between samples. The processor will pad the inputs to the maximum number of images in a batch for input to the model.
-- The processor has a `do_image_splitting` option. If `True`, each input image will be split into 4 sub-images, and concatenated with the original to form 5 images. This is useful for increasing model performance. Make sure `processor.image_processor.do_image_splitting` is set to `False` if the model was not trained with this option.
-- `text` passed to the processor should have the `<image>` tokens where the images should be inserted. And `<end_of_utterance>` at the end of each utterance if the text is a chat message.
-- The processor has its own `apply_chat_template` method to convert chat messages to text that can then be passed as `text` to the processor.
-
-Example of how to use the processor on chat messages:
-```python
-import requests
-from PIL import Image
-from transformers import Idefics2Processor, Idefics2ForConditionalGeneration
-
-url_1 = "http://images.cocodataset.org/val2017/000000039769.jpg"
-url_2 = "http://images.cocodataset.org/val2017/000000219578.jpg"
-
-image_1 = Image.open(requests.get(url_1, stream=True).raw)
-image_2 = Image.open(requests.get(url_2, stream=True).raw)
-images = [image_1, image_2]
-
-messages = [{
-    "role": "user",
-    "content": [
-        {"type": "text", "text": "What’s the difference between these two images?"},
-        {"type": "image"},
-        {"type": "image"},
-    ],
-}]
-
-processor = Idefics2Processor.from_pretrained("HuggingFaceM4/idefics2-8b")
-model = Idefics2ForConditionalGeneration.from_pretrained("HuggingFaceM4/idefics2-8b")
-
-text = processor.apply_chat_template(messages)
-# "User: What’s the difference between these two images?<image><image><end_of_utterance>\n"
-print(text)
-
-inputs = processor(images=images, text=text)
-
-generated_text = model.generate(**inputs)
-```
-
-This model was contributed by [amyeroberts](https://huggingface.co/amyeroberts).
-The original code can be found [here](https://huggingface.co/HuggingFaceM4/idefics2).
-
-
-## Idefics2Config
-
-[[autodoc]] Idefics2Config
-
-
-## Idefics2Model
-
-[[autodoc]] Idefics2Model
-    - forward
-
-
-## Idefics2ForConditionalGeneration
-
-[[autodoc]] Idefics2ForConditionalGeneration
-    - forward
-
-
-## Idefics2ImageProcessor
-[[autodoc]] Idefics2ImageProcessor
-    - preprocess
-
-
-## Idefics2Processor
-[[autodoc]] Idefics2Processor
-    - __call__

docs/source/en/model_doc/llava.md

@@ -43,13 +43,13 @@ The original code can be found [here](https://github.com/haotian-liu/LLaVA/tree/
 - For better results, we recommend users to prompt the model with the correct prompt format: 
 
 ```bash
-"USER: <image>\n<prompt> ASSISTANT:"
+"USER: <image>\n<prompt>ASSISTANT:"
 ```
 
 For multiple turns conversation:
 
 ```bash
-"USER: <image>\n<prompt1> ASSISTANT: <answer1></s>USER: <prompt2> ASSISTANT: <answer2></s>USER: <prompt3> ASSISTANT:"
+"USER: <image>\n<prompt1>ASSISTANT: <answer1>USER: <prompt2>ASSISTANT: <answer2>USER: <prompt3>ASSISTANT:"
 ```
 
 ### Using Flash Attention 2

docs/source/en/model_doc/phi.md

@@ -92,9 +92,7 @@ Phi-2 has been integrated in the development version (4.37.0.dev) of `transforme
 >>> outputs = model.generate(**inputs, max_length=30)
 >>> text = tokenizer.batch_decode(outputs)[0]
 >>> print(text)
-Can you help me write a formal email to a potential business partner proposing a joint venture?
-Input: Company A: ABC Inc.
-Company B
+'Can you help me write a formal email to a potential business partner proposing a joint venture?\nInput: Company A: ABC Inc.\nCompany B: XYZ Ltd.\nJoint Venture: A new online platform for e-commerce'
 ```
 
 ### Example :
@@ -136,7 +134,7 @@ To load and run a model using Flash Attention 2, refer to the snippet below:
 >>> from transformers import PhiForCausalLM, AutoTokenizer
 
 >>> # define the model and tokenizer and push the model and tokens to the GPU.
->>> model = PhiForCausalLM.from_pretrained("microsoft/phi-1_5", torch_dtype=torch.float16, attn_implementation="flash_attention_2").to("cuda")  # doctest: +SKIP
+>>> model = PhiForCausalLM.from_pretrained("microsoft/phi-1_5", torch_dtype=torch.float16, attn_implementation="flash_attention_2").to("cuda")
 >>> tokenizer = AutoTokenizer.from_pretrained("microsoft/phi-1_5")
 
 >>> # feel free to change the prompt to your liking.
@@ -146,9 +144,9 @@ To load and run a model using Flash Attention 2, refer to the snippet below:
 >>> tokens = tokenizer(prompt, return_tensors="pt").to("cuda")
 
 >>> # use the model to generate new tokens.
->>> generated_output = model.generate(**tokens, use_cache=True, max_new_tokens=10)  # doctest: +SKIP
+>>> generated_output = model.generate(**tokens, use_cache=True, max_new_tokens=10)
 
->>> tokenizer.batch_decode(generated_output)[0]  # doctest: +SKIP
+>>> tokenizer.batch_decode(generated_output)[0]
 'If I were an AI that had just achieved a breakthrough in machine learning, I would be thrilled'
 ```
 

docs/source/en/model_doc/recurrent_gemma.md

@@ -1,48 +0,0 @@
-<!--Copyright 2024 The HuggingFace Team. All rights reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
-the License. You may obtain a copy of the License at
-
-http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
-an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
-specific language governing permissions and limitations under the License.
-
-⚠️ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
-rendered properly in your Markdown viewer.
-
--->
-
-# RecurrentGemma
-
-## Overview
-
-The Recurrent Gemma model was proposed in [RecurrentGemma: Moving Past Transformers for Efficient Open Language Models](https://storage.googleapis.com/deepmind-media/gemma/recurrentgemma-report.pdf) by the Griffin, RLHF and Gemma Teams of Google.
-
-The abstract from the paper is the following:
-
-*We introduce RecurrentGemma, an open language model which uses Google’s novel Griffin architecture. Griffin combines linear recurrences with local attention to achieve excellent performance on language. It has a fixed-sized state, which reduces memory use and enables efficient inference on long sequences. We provide a pre-trained model with 2B non-embedding parameters, and an instruction tuned variant. Both models achieve comparable performance to Gemma-2B despite being trained on fewer tokens.*
-
-Tips:
-
-- The original checkpoints can be converted using the conversion script [`src/transformers/models/recurrent_gemma/convert_recurrent_gemma_weights_to_hf.py`](https://github.com/huggingface/transformers/blob/main/src/transformers/models/recurrent_gemma/convert_recurrent_gemma_to_hf.py). 
-
-This model was contributed by [Arthur Zucker](https://huggingface.co/ArthurZ). The original code can be found [here](https://github.com/google-deepmind/recurrentgemma).
-
-
-## RecurrentGemmaConfig
-
-[[autodoc]] RecurrentGemmaConfig
-
-
-## RecurrentGemmaModel
-
-[[autodoc]] RecurrentGemmaModel
-    - forward
-
-## RecurrentGemmaForCausalLM
-
-[[autodoc]] RecurrentGemmaForCausalLM
-    - forward
-

docs/source/en/model_doc/stablelm.md

@@ -37,21 +37,19 @@ We also provide `StableLM Zephyr 3B`, an instruction fine-tuned version of the m
 The following code snippet demonstrates how to use `StableLM 3B 4E1T` for inference:
 
 ```python
->>> from transformers import AutoModelForCausalLM, AutoTokenizer, set_seed
+>>> from transformers import AutoModelForCausalLM, AutoTokenizer
 >>> device = "cuda" # the device to load the model onto
 
->>> set_seed(0)
-
 >>> tokenizer = AutoTokenizer.from_pretrained("stabilityai/stablelm-3b-4e1t")
 >>> model = AutoModelForCausalLM.from_pretrained("stabilityai/stablelm-3b-4e1t")
->>> model.to(device)  # doctest: +IGNORE_RESULT
+>>> model.to(device)
 
 >>> model_inputs = tokenizer("The weather is always wonderful in", return_tensors="pt").to(model.device)
 
 >>> generated_ids = model.generate(**model_inputs, max_length=32, do_sample=True)
 >>> responses = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
 >>> responses
-['The weather is always wonderful in Costa Rica, which makes it a prime destination for retirees. That’s where the Pensionado program comes in, offering']
+['The weather is always wonderful in Santa Barbara and, for visitors hoping to make the move to our beautiful seaside city, this town offers plenty of great places to...']
 ```
 
 ## Combining StableLM and Flash Attention 2
@@ -68,21 +66,19 @@ Now, to run the model with Flash Attention 2, refer to the snippet below:
 
 ```python
 >>> import torch
->>> from transformers import AutoModelForCausalLM, AutoTokenizer, set_seed
+>>> from transformers import AutoModelForCausalLM, AutoTokenizer
 >>> device = "cuda" # the device to load the model onto
 
->>> set_seed(0)
-
 >>> tokenizer = AutoTokenizer.from_pretrained("stabilityai/stablelm-3b-4e1t")
->>> model = AutoModelForCausalLM.from_pretrained("stabilityai/stablelm-3b-4e1t", torch_dtype=torch.bfloat16, attn_implementation="flash_attention_2")  # doctest: +SKIP
->>> model.to(device)  # doctest: +SKIP
+>>> model = AutoModelForCausalLM.from_pretrained("stabilityai/stablelm-3b-4e1t", torch_dtype=torch.bfloat16, attn_implementation="flash_attention_2")
+>>> model.to(device)
 
 >>> model_inputs = tokenizer("The weather is always wonderful in", return_tensors="pt").to(model.device)
 
->>> generated_ids = model.generate(**model_inputs, max_length=32, do_sample=True)  # doctest: +SKIP
->>> responses = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)  # doctest: +SKIP
->>> responses  # doctest: +SKIP
-['The weather is always wonderful in Costa Rica, which makes it a prime destination for retirees. That’s where the Pensionado program comes in, offering']
+>>> generated_ids = model.generate(**model_inputs, max_length=32, do_sample=True)
+>>> responses = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
+>>> responses
+['The weather is always wonderful in Santa Barbara and, for visitors hoping to make the move to our beautiful seaside city, this town offers plenty of great places to...']
 ```
 
 

docs/source/en/model_doc/starcoder2.md

@@ -42,10 +42,11 @@ These ready-to-use checkpoints can be downloaded and used via the HuggingFace Hu
 >>> prompt = "def print_hello_world():"
 
 >>> model_inputs = tokenizer([prompt], return_tensors="pt").to("cuda")
+>>> model.to(device)
 
 >>> generated_ids = model.generate(**model_inputs, max_new_tokens=10, do_sample=False)
 >>> tokenizer.batch_decode(generated_ids)[0]
-'def print_hello_world():\n    print("Hello World!")\n\ndef print'
+"def print_hello_world():\n\treturn 'Hello World!'"
 ```
 
 ## Starcoder2Config

docs/source/en/model_doc/t5.md

@@ -309,7 +309,7 @@ The predicted tokens will then be placed between the sentinel tokens.
 >>> sequence_ids = model.generate(input_ids)
 >>> sequences = tokenizer.batch_decode(sequence_ids)
 >>> sequences
-['<pad> <extra_id_0> park offers <extra_id_1> the <extra_id_2> park.</s>']
+['<pad><extra_id_0> park offers<extra_id_1> the<extra_id_2> park.</s>']
 ```
 
 ## Performance

docs/source/en/model_doc/udop.md

@@ -56,25 +56,14 @@ image = Image.open(name_of_your_document).convert("RGB")
 width, height = image.size
 ```
 
-One can use [`UdopProcessor`] to prepare images and text for the model, which takes care of all of this. By default, this class uses the Tesseract engine to extract a list of words and boxes (coordinates) from a given document. Its functionality is equivalent to that of [`LayoutLMv3Processor`], hence it supports passing either `apply_ocr=False` in case you prefer to use your own OCR engine or `apply_ocr=True` in case you want the default OCR engine to be used. Refer to the [usage guide of LayoutLMv2](layoutlmv2#usage-layoutlmv2processor) regarding all possible use cases (the functionality of `UdopProcessor` is identical).
-
-- If using an own OCR engine of choice, one recommendation is Azure's [Read API](https://learn.microsoft.com/en-us/azure/ai-services/computer-vision/how-to/call-read-api), which supports so-called line segments. Use of segment position embeddings typically results in better performance.
 - At inference time, it's recommended to use the `generate` method to autoregressively generate text given a document image.
-- The model has been pre-trained on both self-supervised and supervised objectives. One can use the various task prefixes (prompts) used during pre-training to test out the out-of-the-box capabilities. For instance, the model can be prompted with "Question answering. What is the date?", as "Question answering." is the task prefix used during pre-training for DocVQA. Refer to the [paper](https://arxiv.org/abs/2212.02623) (table 1) for all task prefixes.
-- One can also fine-tune [`UdopEncoderModel`], which is the encoder-only part of UDOP, which can be seen as a LayoutLMv3-like Transformer encoder. For discriminative tasks, one can just add a linear classifier on top of it and fine-tune it on a labeled dataset.
+- One can use [`UdopProcessor`] to prepare images and text for the model. By default, this class uses the Tesseract engine to extract a list of words
+and boxes (coordinates) from a given document. Its functionality is equivalent to that of [`LayoutLMv3Processor`], hence it supports passing either
+`apply_ocr=False` in case you prefer to use your own OCR engine or `apply_ocr=True` in case you want the default OCR engine to be used.
 
 This model was contributed by [nielsr](https://huggingface.co/nielsr).
 The original code can be found [here](https://github.com/microsoft/UDOP).
 
-## Resources
-
-A list of official Hugging Face and community (indicated by 🌎) resources to help you get started with UDOP. If
-you're interested in submitting a resource to be included here, please feel free to open a Pull Request and we'll
-review it! The resource should ideally demonstrate something new instead of duplicating an existing resource.
-
-- Demo notebooks regarding UDOP can be found [here](https://github.com/NielsRogge/Transformers-Tutorials/tree/master/UDOP) that show how
-to fine-tune UDOP on a custom dataset as well as inference. 🌎
-- [Document question answering task guide](../tasks/document_question_answering)
 
 ## UdopConfig
 

docs/source/en/perf_infer_gpu_one.md

@@ -47,7 +47,6 @@ FlashAttention-2 is currently supported for the following architectures:
 * [GPTNeo](https://huggingface.co/docs/transformers/model_doc/gpt_neo#transformers.GPTNeoModel)
 * [GPTNeoX](https://huggingface.co/docs/transformers/model_doc/gpt_neox#transformers.GPTNeoXModel)
 * [GPT-J](https://huggingface.co/docs/transformers/model_doc/gptj#transformers.GPTJModel)
-* [Idefics2](https://huggingface.co/docs/transformers/model_doc/idefics2#transformers.Idefics2Model)
 * [Falcon](https://huggingface.co/docs/transformers/model_doc/falcon#transformers.FalconModel)
 * [Llama](https://huggingface.co/docs/transformers/model_doc/llama#transformers.LlamaModel)
 * [Llava](https://huggingface.co/docs/transformers/model_doc/llava)
@@ -97,8 +96,8 @@ model_id = "tiiuae/falcon-7b"
 tokenizer = AutoTokenizer.from_pretrained(model_id)
 
 model = AutoModelForCausalLM.from_pretrained(
-    model_id,
-    torch_dtype=torch.bfloat16,
+    model_id, 
+    torch_dtype=torch.bfloat16, 
     attn_implementation="flash_attention_2",
 )
 ```
@@ -110,7 +109,7 @@ FlashAttention-2 can only be used when the model's dtype is `fp16` or `bf16`. Ma
 <br>
 
 You can also set `use_flash_attention_2=True` to enable FlashAttention-2 but it is deprecated in favor of `attn_implementation="flash_attention_2"`.
-
+  
 </Tip>
 
 FlashAttention-2 can be combined with other optimization techniques like quantization to further speedup inference. For example, you can combine FlashAttention-2 with 8-bit or 4-bit quantization:
@@ -124,14 +123,14 @@ tokenizer = AutoTokenizer.from_pretrained(model_id)
 
 # load in 8bit
 model = AutoModelForCausalLM.from_pretrained(
-    model_id,
+    model_id, 
     load_in_8bit=True,
     attn_implementation="flash_attention_2",
 )
 
 # load in 4bit
 model = AutoModelForCausalLM.from_pretrained(
-    model_id,
+    model_id, 
     load_in_4bit=True,
     attn_implementation="flash_attention_2",
 )

docs/source/en/pipeline_tutorial.md

@@ -270,13 +270,11 @@ For example, if you use this [invoice image](https://huggingface.co/spaces/impir
 >>> from transformers import pipeline
 
 >>> vqa = pipeline(model="impira/layoutlm-document-qa")
->>> output = vqa(
+>>> vqa(
 ...     image="https://huggingface.co/spaces/impira/docquery/resolve/2359223c1837a7587402bda0f2643382a6eefeab/invoice.png",
 ...     question="What is the invoice number?",
 ... )
->>> output[0]["score"] = round(output[0]["score"], 3)
->>> output
-[{'score': 0.425, 'answer': 'us-001', 'start': 16, 'end': 16}]
+[{'score': 0.42515, 'answer': 'us-001', 'start': 16, 'end': 16}]
 ```
 
 <Tip>

docs/source/en/task_summary.md

@@ -326,7 +326,7 @@ Document question answering is a task that answers natural language questions fr
 >>> from PIL import Image
 >>> import requests
 
->>> url = "https://huggingface.co/datasets/hf-internal-testing/example-documents/resolve/main/jpeg_images/2.jpg"
+>>> url = "https://datasets-server.huggingface.co/assets/hf-internal-testing/example-documents/--/hf-internal-testing--example-documents/test/2/image/image.jpg"
 >>> image = Image.open(requests.get(url, stream=True).raw)
 
 >>> doc_question_answerer = pipeline("document-question-answering", model="magorshunov/layoutlm-invoices")

docs/source/en/tasks/language_modeling.md

@@ -37,7 +37,7 @@ You can finetune other architectures for causal language modeling following the
 Choose one of the following architectures:
 
 <!--This tip is automatically generated by `make fix-copies`, do not fill manually!-->
-[BART](../model_doc/bart), [BERT](../model_doc/bert), [Bert Generation](../model_doc/bert-generation), [BigBird](../model_doc/big_bird), [BigBird-Pegasus](../model_doc/bigbird_pegasus), [BioGpt](../model_doc/biogpt), [Blenderbot](../model_doc/blenderbot), [BlenderbotSmall](../model_doc/blenderbot-small), [BLOOM](../model_doc/bloom), [CamemBERT](../model_doc/camembert), [CodeLlama](../model_doc/code_llama), [CodeGen](../model_doc/codegen), [Cohere](../model_doc/cohere), [CPM-Ant](../model_doc/cpmant), [CTRL](../model_doc/ctrl), [Data2VecText](../model_doc/data2vec-text), [ELECTRA](../model_doc/electra), [ERNIE](../model_doc/ernie), [Falcon](../model_doc/falcon), [Fuyu](../model_doc/fuyu), [Gemma](../model_doc/gemma), [GIT](../model_doc/git), [GPT-Sw3](../model_doc/gpt-sw3), [OpenAI GPT-2](../model_doc/gpt2), [GPTBigCode](../model_doc/gpt_bigcode), [GPT Neo](../model_doc/gpt_neo), [GPT NeoX](../model_doc/gpt_neox), [GPT NeoX Japanese](../model_doc/gpt_neox_japanese), [GPT-J](../model_doc/gptj), [LLaMA](../model_doc/llama), [Mamba](../model_doc/mamba), [Marian](../model_doc/marian), [mBART](../model_doc/mbart), [MEGA](../model_doc/mega), [Megatron-BERT](../model_doc/megatron-bert), [Mistral](../model_doc/mistral), [Mixtral](../model_doc/mixtral), [MPT](../model_doc/mpt), [MusicGen](../model_doc/musicgen), [MusicGen Melody](../model_doc/musicgen_melody), [MVP](../model_doc/mvp), [OpenLlama](../model_doc/open-llama), [OpenAI GPT](../model_doc/openai-gpt), [OPT](../model_doc/opt), [Pegasus](../model_doc/pegasus), [Persimmon](../model_doc/persimmon), [Phi](../model_doc/phi), [PLBart](../model_doc/plbart), [ProphetNet](../model_doc/prophetnet), [QDQBert](../model_doc/qdqbert), [Qwen2](../model_doc/qwen2), [Qwen2MoE](../model_doc/qwen2_moe), [RecurrentGemma](../model_doc/recurrent_gemma), [Reformer](../model_doc/reformer), [RemBERT](../model_doc/rembert), [RoBERTa](../model_doc/roberta), [RoBERTa-PreLayerNorm](../model_doc/roberta-prelayernorm), [RoCBert](../model_doc/roc_bert), [RoFormer](../model_doc/roformer), [RWKV](../model_doc/rwkv), [Speech2Text2](../model_doc/speech_to_text_2), [StableLm](../model_doc/stablelm), [Starcoder2](../model_doc/starcoder2), [Transformer-XL](../model_doc/transfo-xl), [TrOCR](../model_doc/trocr), [Whisper](../model_doc/whisper), [XGLM](../model_doc/xglm), [XLM](../model_doc/xlm), [XLM-ProphetNet](../model_doc/xlm-prophetnet), [XLM-RoBERTa](../model_doc/xlm-roberta), [XLM-RoBERTa-XL](../model_doc/xlm-roberta-xl), [XLNet](../model_doc/xlnet), [X-MOD](../model_doc/xmod)
+[BART](../model_doc/bart), [BERT](../model_doc/bert), [Bert Generation](../model_doc/bert-generation), [BigBird](../model_doc/big_bird), [BigBird-Pegasus](../model_doc/bigbird_pegasus), [BioGpt](../model_doc/biogpt), [Blenderbot](../model_doc/blenderbot), [BlenderbotSmall](../model_doc/blenderbot-small), [BLOOM](../model_doc/bloom), [CamemBERT](../model_doc/camembert), [CodeLlama](../model_doc/code_llama), [CodeGen](../model_doc/codegen), [Cohere](../model_doc/cohere), [CPM-Ant](../model_doc/cpmant), [CTRL](../model_doc/ctrl), [Data2VecText](../model_doc/data2vec-text), [ELECTRA](../model_doc/electra), [ERNIE](../model_doc/ernie), [Falcon](../model_doc/falcon), [Fuyu](../model_doc/fuyu), [Gemma](../model_doc/gemma), [GIT](../model_doc/git), [GPT-Sw3](../model_doc/gpt-sw3), [OpenAI GPT-2](../model_doc/gpt2), [GPTBigCode](../model_doc/gpt_bigcode), [GPT Neo](../model_doc/gpt_neo), [GPT NeoX](../model_doc/gpt_neox), [GPT NeoX Japanese](../model_doc/gpt_neox_japanese), [GPT-J](../model_doc/gptj), [LLaMA](../model_doc/llama), [Mamba](../model_doc/mamba), [Marian](../model_doc/marian), [mBART](../model_doc/mbart), [MEGA](../model_doc/mega), [Megatron-BERT](../model_doc/megatron-bert), [Mistral](../model_doc/mistral), [Mixtral](../model_doc/mixtral), [MPT](../model_doc/mpt), [MusicGen](../model_doc/musicgen), [MusicGen Melody](../model_doc/musicgen_melody), [MVP](../model_doc/mvp), [OpenLlama](../model_doc/open-llama), [OpenAI GPT](../model_doc/openai-gpt), [OPT](../model_doc/opt), [Pegasus](../model_doc/pegasus), [Persimmon](../model_doc/persimmon), [Phi](../model_doc/phi), [PLBart](../model_doc/plbart), [ProphetNet](../model_doc/prophetnet), [QDQBert](../model_doc/qdqbert), [Qwen2](../model_doc/qwen2), [Qwen2MoE](../model_doc/qwen2_moe), [Reformer](../model_doc/reformer), [RemBERT](../model_doc/rembert), [RoBERTa](../model_doc/roberta), [RoBERTa-PreLayerNorm](../model_doc/roberta-prelayernorm), [RoCBert](../model_doc/roc_bert), [RoFormer](../model_doc/roformer), [RWKV](../model_doc/rwkv), [Speech2Text2](../model_doc/speech_to_text_2), [StableLm](../model_doc/stablelm), [Starcoder2](../model_doc/starcoder2), [Transformer-XL](../model_doc/transfo-xl), [TrOCR](../model_doc/trocr), [Whisper](../model_doc/whisper), [XGLM](../model_doc/xglm), [XLM](../model_doc/xlm), [XLM-ProphetNet](../model_doc/xlm-prophetnet), [XLM-RoBERTa](../model_doc/xlm-roberta), [XLM-RoBERTa-XL](../model_doc/xlm-roberta-xl), [XLNet](../model_doc/xlnet), [X-MOD](../model_doc/xmod)
 
 
 

docs/source/en/tasks/prompting.md

@@ -80,7 +80,7 @@ Run inference with decoder-only models with the `text-generation` pipeline:
 >>> prompt = "Hello, I'm a language model"
 
 >>> generator(prompt, max_length = 30)
-[{'generated_text': "Hello, I'm a language model programmer so you can use some of my stuff. But you also need some sort of a C program to run."}]
+[{'generated_text': "Hello, I'm a language model expert, so I'm a big believer in the concept that I know very well and then I try to look into"}]
 ```
 
 To run inference with an encoder-decoder, use the `text2text-generation` pipeline:
@@ -284,7 +284,7 @@ the leading word or phrase (`"Answer:"`) to nudge the model to start generating
 
 >>> for seq in sequences:
 ...     print(f"Result: {seq['generated_text']}")
-Result: Modern tools often used to make gazpacho include
+Result: Modern tools are used, such as immersion blenders
 ```
 
 #### Reasoning

docs/source/en/tasks/semantic_segmentation.md

@@ -28,9 +28,8 @@ In this guide, we will:
 
 Before you begin, make sure you have all the necessary libraries installed:
 
-```py
-# uncomment to install the necessary libraries
-!pip install -q datasets transformers evaluate accelerate
+```bash
+pip install -q datasets transformers evaluate
 ```
 
 We encourage you to log in to your Hugging Face account so you can upload and share your model with the community. When prompted, enter your token to log in:
@@ -237,9 +236,6 @@ Then take a look at an example:
 {'image': <PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=512x683 at 0x7F9B0C201F90>,
  'annotation': <PIL.PngImagePlugin.PngImageFile image mode=L size=512x683 at 0x7F9B0C201DD0>,
  'scene_category': 368}
-
-# view the image
->>> train_ds[0]["image"]
 ```
 
 - `image`: a PIL image of the scene.
@@ -667,19 +663,15 @@ Congratulations! You have fine-tuned your model and shared it on the 🤗 Hub. Y
 </tf>
 </frameworkcontent>
 
+
 ### Inference
 
 Great, now that you've finetuned a model, you can use it for inference!
 
-Reload the dataset and load an image for inference.
+Load an image for inference:
 
 ```py
->>> from datasets import load_dataset
-
->>> ds = load_dataset("scene_parse_150", split="train[:50]")
->>> ds = ds.train_test_split(test_size=0.2)
->>> test_ds = ds["test"]
->>> image = ds["test"][0]["image"]
+>>> image = ds[0]["image"]
 >>> image
 ```
 
@@ -757,166 +749,7 @@ Next, rescale the logits to the original image size and apply argmax on the clas
 </tf>
 </frameworkcontent>
 
-To visualize the results, load the [dataset color palette](https://github.com/tensorflow/models/blob/3f1ca33afe3c1631b733ea7e40c294273b9e406d/research/deeplab/utils/get_dataset_colormap.py#L51) as `ade_palette()` that maps each class to their RGB values.
-
-```py
-def ade_palette():
-  return np.asarray([
-      [0, 0, 0],
-      [120, 120, 120],
-      [180, 120, 120],
-      [6, 230, 230],
-      [80, 50, 50],
-      [4, 200, 3],
-      [120, 120, 80],
-      [140, 140, 140],
-      [204, 5, 255],
-      [230, 230, 230],
-      [4, 250, 7],
-      [224, 5, 255],
-      [235, 255, 7],
-      [150, 5, 61],
-      [120, 120, 70],
-      [8, 255, 51],
-      [255, 6, 82],
-      [143, 255, 140],
-      [204, 255, 4],
-      [255, 51, 7],
-      [204, 70, 3],
-      [0, 102, 200],
-      [61, 230, 250],
-      [255, 6, 51],
-      [11, 102, 255],
-      [255, 7, 71],
-      [255, 9, 224],
-      [9, 7, 230],
-      [220, 220, 220],
-      [255, 9, 92],
-      [112, 9, 255],
-      [8, 255, 214],
-      [7, 255, 224],
-      [255, 184, 6],
-      [10, 255, 71],
-      [255, 41, 10],
-      [7, 255, 255],
-      [224, 255, 8],
-      [102, 8, 255],
-      [255, 61, 6],
-      [255, 194, 7],
-      [255, 122, 8],
-      [0, 255, 20],
-      [255, 8, 41],
-      [255, 5, 153],
-      [6, 51, 255],
-      [235, 12, 255],
-      [160, 150, 20],
-      [0, 163, 255],
-      [140, 140, 140],
-      [250, 10, 15],
-      [20, 255, 0],
-      [31, 255, 0],
-      [255, 31, 0],
-      [255, 224, 0],
-      [153, 255, 0],
-      [0, 0, 255],
-      [255, 71, 0],
-      [0, 235, 255],
-      [0, 173, 255],
-      [31, 0, 255],
-      [11, 200, 200],
-      [255, 82, 0],
-      [0, 255, 245],
-      [0, 61, 255],
-      [0, 255, 112],
-      [0, 255, 133],
-      [255, 0, 0],
-      [255, 163, 0],
-      [255, 102, 0],
-      [194, 255, 0],
-      [0, 143, 255],
-      [51, 255, 0],
-      [0, 82, 255],
-      [0, 255, 41],
-      [0, 255, 173],
-      [10, 0, 255],
-      [173, 255, 0],
-      [0, 255, 153],
-      [255, 92, 0],
-      [255, 0, 255],
-      [255, 0, 245],
-      [255, 0, 102],
-      [255, 173, 0],
-      [255, 0, 20],
-      [255, 184, 184],
-      [0, 31, 255],
-      [0, 255, 61],
-      [0, 71, 255],
-      [255, 0, 204],
-      [0, 255, 194],
-      [0, 255, 82],
-      [0, 10, 255],
-      [0, 112, 255],
-      [51, 0, 255],
-      [0, 194, 255],
-      [0, 122, 255],
-      [0, 255, 163],
-      [255, 153, 0],
-      [0, 255, 10],
-      [255, 112, 0],
-      [143, 255, 0],
-      [82, 0, 255],
-      [163, 255, 0],
-      [255, 235, 0],
-      [8, 184, 170],
-      [133, 0, 255],
-      [0, 255, 92],
-      [184, 0, 255],
-      [255, 0, 31],
-      [0, 184, 255],
-      [0, 214, 255],
-      [255, 0, 112],
-      [92, 255, 0],
-      [0, 224, 255],
-      [112, 224, 255],
-      [70, 184, 160],
-      [163, 0, 255],
-      [153, 0, 255],
-      [71, 255, 0],
-      [255, 0, 163],
-      [255, 204, 0],
-      [255, 0, 143],
-      [0, 255, 235],
-      [133, 255, 0],
-      [255, 0, 235],
-      [245, 0, 255],
-      [255, 0, 122],
-      [255, 245, 0],
-      [10, 190, 212],
-      [214, 255, 0],
-      [0, 204, 255],
-      [20, 0, 255],
-      [255, 255, 0],
-      [0, 153, 255],
-      [0, 41, 255],
-      [0, 255, 204],
-      [41, 0, 255],
-      [41, 255, 0],
-      [173, 0, 255],
-      [0, 245, 255],
-      [71, 0, 255],
-      [122, 0, 255],
-      [0, 255, 184],
-      [0, 92, 255],
-      [184, 255, 0],
-      [0, 133, 255],
-      [255, 214, 0],
-      [25, 194, 194],
-      [102, 255, 0],
-      [92, 0, 255],
-  ])
-```
-
-Then you can combine and plot your image and the predicted segmentation map:
+To visualize the results, load the [dataset color palette](https://github.com/tensorflow/models/blob/3f1ca33afe3c1631b733ea7e40c294273b9e406d/research/deeplab/utils/get_dataset_colormap.py#L51) as `ade_palette()` that maps each class to their RGB values. Then you can combine and plot your image and the predicted segmentation map:
 
 ```py
 >>> import matplotlib.pyplot as plt

docs/source/en/tasks/translation.md

@@ -348,10 +348,7 @@ The simplest way to try out your finetuned model for inference is to use it in a
 ```py
 >>> from transformers import pipeline
 
-# Change `xx` to the language of the input and `yy` to the language of the desired output.
-# Examples: "en" for English, "fr" for French, "de" for German, "es" for Spanish, "zh" for Chinese, etc; translation_en_to_fr translates English to French
-# You can view all the lists of languages here - https://huggingface.co/languages
->>> translator = pipeline("translation_xx_to_yy", model="my_awesome_opus_books_model")
+>>> translator = pipeline("translation", model="my_awesome_opus_books_model")
 >>> translator(text)
 [{'translation_text': 'Legumes partagent des ressources avec des bactéries azotantes.'}]
 ```

docs/source/es/add_new_pipeline.md

@@ -212,10 +212,14 @@ from transformers import pipeline
 classifier = pipeline("pair-classification", model="sgugger/finetuned-bert-mrpc")
 ```
 
-Ahora podemos compartirlo en el Hub usando el método `save_pretrained`:
+Ahora podemos compartirlo en el Hub usando el método `save_pretrained` (guardar pre-entrenado) en un `Repository`:
 
 ```py
-classifier.push_to_hub("test-dynamic-pipeline")
+from huggingface_hub import Repository
+
+repo = Repository("test-dynamic-pipeline", clone_from="{your_username}/test-dynamic-pipeline")
+classifier.save_pretrained("test-dynamic-pipeline")
+repo.push_to_hub()
 ```
 
 Esto copiará el archivo donde definiste `PairClassificationPipeline` dentro de la carpeta `"test-dynamic-pipeline"`,

docs/source/es/task_summary.md

@@ -325,7 +325,7 @@ Las respuestas a preguntas de documentos es una tarea que responde preguntas en
 >>> from PIL import Image
 >>> import requests
 
->>> url = "https://huggingface.co/datasets/hf-internal-testing/example-documents/resolve/main/jpeg_images/2.jpg"
+>>> url = "https://datasets-server.huggingface.co/assets/hf-internal-testing/example-documents/--/hf-internal-testing--example-documents/test/2/image/image.jpg"
 >>> image = Image.open(requests.get(url, stream=True).raw)
 
 >>> doc_question_answerer = pipeline("document-question-answering", model="magorshunov/layoutlm-invoices")

docs/source/hi/pipeline_tutorial.md

@@ -270,13 +270,11 @@ NLP कार्यों के लिए [`pipeline`] का उपयोग
 >>> from transformers import pipeline
 
 >>> vqa = pipeline(model="impira/layoutlm-document-qa")
->>> output = vqa(
+>>> vqa(
 ...     image="https://huggingface.co/spaces/impira/docquery/resolve/2359223c1837a7587402bda0f2643382a6eefeab/invoice.png",
 ...     question="What is the invoice number?",
 ... )
->>> output[0]["score"] = round(output[0]["score"], 3)
->>> output
-[{'score': 0.425, 'answer': 'us-001', 'start': 16, 'end': 16}]
+[{'score': 0.42515, 'answer': 'us-001', 'start': 16, 'end': 16}]
 ```
 
 <Tip>

docs/source/it/add_new_pipeline.md

@@ -202,10 +202,14 @@ from transformers import pipeline
 classifier = pipeline("pair-classification", model="sgugger/finetuned-bert-mrpc")
 ```
 
-Successivamente possiamo condividerlo sull'Hub usando il metodo `push_to_hub`
+Successivamente possiamo condividerlo sull'Hub usando il metodo `save_pretrained` in un `Repository`:
 
 ```py
-classifier.push_to_hub("test-dynamic-pipeline")
+from huggingface_hub import Repository
+
+repo = Repository("test-dynamic-pipeline", clone_from="{your_username}/test-dynamic-pipeline")
+classifier.save_pretrained("test-dynamic-pipeline")
+repo.push_to_hub()
 ```
 
 Questo codice copierà il file dove è stato definitp `PairClassificationPipeline` all'interno della cartella `"test-dynamic-pipeline"`,

docs/source/ja/model_doc/code_llama.md

@@ -94,8 +94,7 @@ def remove_non_ascii(s: str) -> str:
 >>> import torch
 
 >>> generator = pipeline("text-generation",model="codellama/CodeLlama-7b-hf",torch_dtype=torch.float16, device_map="auto")
->>> generator('def remove_non_ascii(s: str) -> str:\n    """ <FILL_ME>\n    return result', max_new_tokens = 128)
-[{'generated_text': 'def remove_non_ascii(s: str) -> str:\n    """ <FILL_ME>\n    return resultRemove non-ASCII characters from a string. """\n    result = ""\n    for c in s:\n        if ord(c) < 128:\n            result += c'}]
+>>> generator('def remove_non_ascii(s: str) -> str:\n    """ <FILL_ME>\n    return result', max_new_tokens = 128, return_type = 1)
 ```
 
 内部では、トークナイザーが [`<FILL_ME>` によって自動的に分割](https://huggingface.co/docs/transformers/main/model_doc/code_llama#transformers.CodeLlamaTokenizer.fill_token) して、[ に続く書式設定された入力文字列を作成します。オリジナルのトレーニング パターン](https://github.com/facebookresearch/codellama/blob/cb51c14ec761370ba2e2bc351374a79265d0465e/llama/generation.py#L402)。これは、パターンを自分で準備するよりも堅牢です。トークンの接着など、デバッグが非常に難しい落とし穴を回避できます。このモデルまたは他のモデルに必要な CPU および GPU メモリの量を確認するには、その値を決定するのに役立つ [この計算ツール](https://huggingface.co/spaces/hf-accelerate/model-memory-usage) を試してください。

docs/source/ja/pipeline_tutorial.md

@@ -246,13 +246,11 @@ for out in pipe(KeyDataset(dataset, "audio")):
 >>> from transformers import pipeline
 
 >>> vqa = pipeline(model="impira/layoutlm-document-qa")
->>> output = vqa(
+>>> vqa(
 ...     image="https://huggingface.co/spaces/impira/docquery/resolve/2359223c1837a7587402bda0f2643382a6eefeab/invoice.png",
 ...     question="What is the invoice number?",
 ... )
->>> output[0]["score"] = round(output[0]["score"], 3)
->>> output
-[{'score': 0.425, 'answer': 'us-001', 'start': 16, 'end': 16}]
+[{'score': 0.42515, 'answer': 'us-001', 'start': 16, 'end': 16}]
 ```
 
 <Tip>

docs/source/ja/task_summary.md

@@ -340,7 +340,7 @@ score: 0.9327, start: 30, end: 54, answer: huggingface/transformers
 >>> from PIL import Image
 >>> import requests
 
->>> url = "https://huggingface.co/datasets/hf-internal-testing/example-documents/resolve/main/jpeg_images/2.jpg"
+>>> url = "https://datasets-server.huggingface.co/assets/hf-internal-testing/example-documents/--/hf-internal-testing--example-documents/test/2/image/image.jpg"
 >>> image = Image.open(requests.get(url, stream=True).raw)
 
 >>> doc_question_answerer = pipeline("document-question-answering", model="magorshunov/layoutlm-invoices")

docs/source/ja/tasks/translation.md

@@ -349,10 +349,7 @@ TensorFlow でモデルを微調整するには、オプティマイザー関数
 ```py
 >>> from transformers import pipeline
 
-# Change `xx` to the language of the input and `yy` to the language of the desired output.
-# Examples: "en" for English, "fr" for French, "de" for German, "es" for Spanish, "zh" for Chinese, etc; translation_en_to_fr translates English to French
-# You can view all the lists of languages here - https://huggingface.co/languages
->>> translator = pipeline("translation_xx_to_yy", model="my_awesome_opus_books_model")
+>>> translator = pipeline("translation", model="my_awesome_opus_books_model")
 >>> translator(text)
 [{'translation_text': 'Legumes partagent des ressources avec des bactéries azotantes.'}]
 ```

docs/source/ko/_toctree.yml

@@ -29,8 +29,7 @@
     title: 대규모 언어 모델로 생성하기
   title: 튜토리얼
 - sections:
-  - isExpanded: false
-    sections:
+  - sections:
       - local: tasks/sequence_classification
         title: 텍스트 분류
       - local: tasks/token_classification
@@ -48,15 +47,15 @@
       - local: tasks/multiple_choice
         title: 객관식 문제(Multiple Choice)
     title: 자연어처리
-  - isExpanded: false
-    sections:
+    isExpanded: false
+  - sections:
       - local: tasks/audio_classification
         title: 오디오 분류
       - local: tasks/asr
         title: 자동 음성 인식
     title: 오디오
-  - isExpanded: false
-    sections:
+    isExpanded: false
+  - sections:
       - local: tasks/image_classification
         title: 이미지 분류
       - local: tasks/semantic_segmentation
@@ -71,114 +70,83 @@
         title: 제로샷(zero-shot) 이미지 분류
       - local: tasks/monocular_depth_estimation
         title: 단일 영상 기반 깊이 추정
-      - local: in_translation
-        title: (번역중) Image-to-Image
-      - local: in_translation
-        title: (번역중) Image Feature Extraction
-      - local: in_translation
-        title: (번역중) Mask Generation
-      - local: in_translation
-        title: (번역중) Knowledge Distillation for Computer Vision
     title: 컴퓨터 비전
-  - isExpanded: false
-    sections:
+    isExpanded: false
+  - sections:
       - local: tasks/image_captioning
         title: 이미지 캡셔닝
       - local: tasks/document_question_answering
         title: 문서 질의 응답(Document Question Answering)
       - local: tasks/visual_question_answering
         title: 시각적 질의응답 (Visual Question Answering)
-      - local: in_translation
-        title: (번역중) Text to speech
     title: 멀티모달
-  - isExpanded: false
-    sections:
-    - local: generation_strategies
-      title: 텍스트 생성 전략 사용자 정의
-    title: 생성
-  - isExpanded: false
-    sections:
-    - local: in_translation
-      title: (번역중) Image tasks with IDEFICS
-    - local: in_translation
-      title: (번역중) LLM prompting guide
-    title: (번역중) 프롬프팅
+    isExpanded: false
   title: 태스크 가이드
 - sections:
-  - local: fast_tokenizers
-    title: 🤗 Tokenizers 라이브러리에서 토크나이저 사용하기
-  - local: multilingual
-    title: 다국어 모델 추론하기
-  - local: create_a_model
-    title: 모델별 API 사용하기
-  - local: custom_models
-    title: 사용자 정의 모델 공유하기
-  - local: in_translation
-    title: (번역중) Templates for chat models
-  - local: in_translation
-    title: (번역중) Trainer
-  - local: sagemaker
-    title: Amazon SageMaker에서 학습 실행하기
-  - local: serialization
-    title: ONNX로 내보내기
-  - local: tflite
-    title: TFLite로 내보내기
-  - local: torchscript
-    title: TorchScript로 내보내기
-  - local: in_translation
-    title: (번역중) Benchmarks
-  - local: in_translation
-    title: (번역중) Notebooks with examples
-  - local: community
-    title: 커뮤니티 리소스
-  - local: custom_tools
-    title: 사용자 정의 도구와 프롬프트
-  - local: troubleshooting
-    title: 문제 해결
-  - local: in_translation
-    title: (번역중) Contribute new quantization method
+    - local: fast_tokenizers
+      title: 🤗 Tokenizers 라이브러리에서 토크나이저 사용하기
+    - local: multilingual
+      title: 다국어 모델 추론하기
+    - local: generation_strategies
+      title: 텍스트 생성 전략 사용자 정의
+    - local: create_a_model
+      title: 모델별 API 사용하기
+    - local: custom_models
+      title: 사용자 정의 모델 공유하기
+    - local: sagemaker
+      title: Amazon SageMaker에서 학습 실행하기
+    - local: serialization
+      title: ONNX로 내보내기
+    - local: tflite
+      title: TFLite로 내보내기
+    - local: torchscript
+      title: TorchScript로 내보내기
+    - local: in_translation
+      title: (번역중) Benchmarks
+    - local: in_translation
+      title: (번역중) Notebooks with examples
+    - local: community
+      title: 커뮤니티 리소스
+    - local: custom_tools
+      title: 사용자 정의 도구와 프롬프트
+    - local: troubleshooting
+      title: 문제 해결
   title: (번역중) 개발자 가이드
 - sections:
-  - local: performance
-    title: 성능 및 확장성
-  - local: in_translation
-    title: (번역중) Quantization
-  - sections:
+    - local: performance
+      title: 성능 및 확장성
     - local: in_translation
       title: (번역중) Training on one GPU
     - local: perf_train_gpu_many
       title: 다중 GPU에서 훈련 진행하기
-    - local: in_translation
-      title: (번역중) Fully Sharded Data Parallel
-    - local: in_translation
-      title: (번역중) DeepSpeed
     - local: perf_train_cpu
       title: CPU에서 훈련
     - local: perf_train_cpu_many
       title: 다중 CPU에서 훈련하기
+    - local: in_translation
+      title: (번역중) Training on TPUs
     - local: perf_train_tpu_tf
       title: TensorFlow로 TPU에서 훈련하기
     - local: in_translation
-      title: (번역중) PyTorch training on Apple silicon
-    - local: perf_hardware
-      title: 훈련용 사용자 맞춤형 하드웨어
-    - local: hpo_train
-      title: Trainer API를 사용한 하이퍼파라미터 탐색
-    title: (번역중) 효율적인 학습 기술들
-  - sections:
+      title: (번역중) Training on Specialized Hardware
     - local: perf_infer_cpu
       title: CPU로 추론하기
     - local: perf_infer_gpu_one
       title: 하나의 GPU를 활용한 추론
-    title: 추론 최적화하기
-  - local: big_models
-    title: 대형 모델을 인스턴스화
-  - local: debugging
-    title: 디버깅
-  - local: tf_xla
-    title: TensorFlow 모델을 위한 XLA 통합
-  - local: in_translation
-    title: (번역중) Optimize inference using `torch.compile()`
+    - local: perf_infer_gpu_many
+      title: 다중 GPU에서 추론
+    - local: in_translation
+      title: (번역중) Inference on Specialized Hardware
+    - local: perf_hardware
+      title: 훈련용 사용자 맞춤형 하드웨어
+    - local: big_models
+      title: 대형 모델을 인스턴스화
+    - local: debugging
+      title: 디버깅
+    - local: hpo_train
+      title: Trainer API를 사용한 하이퍼파라미터 탐색
+    - local: tf_xla
+      title: TensorFlow 모델을 위한 XLA 통합
   title: (번역중) 성능 및 확장성
 - sections:
     - local: contributing
@@ -194,6 +162,7 @@
     - local: pr_checks
       title: Pull Request에 대한 검사
   title: (번역중) 기여하기
+
 - sections:
   - local: philosophy
     title: 이념과 목표
@@ -219,17 +188,11 @@
     title: 추론 웹 서버를 위한 파이프라인
   - local: model_memory_anatomy
     title: 모델 학습 해부하기
-  - local: in_translation
-    title: (번역중) Getting the most out of LLMs
   title: (번역중) 개념 가이드
 - sections:
   - sections:
-    - local: in_translation
-      title: (번역중) Agents and Tools
     - local: in_translation
       title: (번역중) Auto Classes
-    - local: in_translation
-      title: (번역중) Backbones
     - local: in_translation
       title: (번역중) Callbacks
     - local: in_translation
@@ -261,7 +224,7 @@
     - local: in_translation
       title: (번역중) Trainer
     - local: in_translation
-      title: (번역중) DeepSpeed
+      title: (번역중) DeepSpeed Integration
     - local: in_translation
       title: (번역중) Feature Extractor
     - local: in_translation

docs/source/ko/add_new_pipeline.md

@@ -203,10 +203,14 @@ from transformers import pipeline
 classifier = pipeline("pair-classification", model="sgugger/finetuned-bert-mrpc")
 ```
 
-그런 다음 `push_to_hub` 메소드를 사용하여 허브에 공유할 수 있습니다:
+그런 다음 `Repository`의 `save_pretrained` 메소드를 사용하여 허브에 공유할 수 있습니다:
 
 ```py
-classifier.push_to_hub("test-dynamic-pipeline")
+from huggingface_hub import Repository
+
+repo = Repository("test-dynamic-pipeline", clone_from="{your_username}/test-dynamic-pipeline")
+classifier.save_pretrained("test-dynamic-pipeline")
+repo.push_to_hub()
 ```
 
 이렇게 하면 "test-dynamic-pipeline" 폴더 내에 `PairClassificationPipeline`을 정의한 파일이 복사되며, 파이프라인의 모델과 토크나이저도 저장한 후, `{your_username}/test-dynamic-pipeline` 저장소에 있는 모든 것을 푸시합니다.

docs/source/ko/perf_infer_gpu_many.md

@@ -0,0 +1,27 @@
+<!--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
+
+⚠️ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
+rendered properly in your Markdown viewer.
+
+-->
+
+# 다중 GPU에서 효율적인 추론 [[efficient-inference-on-a-multiple-gpus]]
+
+이 문서에는 다중 GPU에서 효율적으로 추론하는 방법에 대한 정보가 포함되어 있습니다.
+<Tip>
+
+참고: 다중 GPU 설정은 [단일 GPU 섹션](./perf_infer_gpu_one)에서 설명된 대부분의 전략을 사용할 수 있습니다. 그러나 더 나은 활용을 위해 간단한 기법들을 알아야 합니다.
+
+</Tip>
+
+## 더 빠른 추론을 위한 `BetterTransformer` [[bettertransformer-for-faster-inference]]
+
+우리는 최근 텍스트, 이미지 및 오디오 모델에 대한 다중 GPU에서 더 빠른 추론을 위해 `BetterTransformer`를 통합했습니다. 자세한 내용은 이 통합에 대한 [문서](https://huggingface.co/docs/optimum/bettertransformer/overview)를 확인하십시오.

docs/source/ko/tasks/translation.md

@@ -346,10 +346,7 @@ TensorFlow에서 모델을 파인튜닝하려면 우선 optimizer 함수, 학습
 ```py
 >>> from transformers import pipeline
 
-# Change `xx` to the language of the input and `yy` to the language of the desired output. 
-# Examples: "en" for English, "fr" for French, "de" for German, "es" for Spanish, "zh" for Chinese, etc; translation_en_to_fr translates English to French
-# You can view all the lists of languages here - https://huggingface.co/languages
->>> translator = pipeline("translation_xx_to_yy", model="my_awesome_opus_books_model")
+>>> translator = pipeline("translation", model="my_awesome_opus_books_model")
 >>> translator(text)
 [{'translation_text': 'Legumes partagent des ressources avec des bactéries azotantes.'}]
 ```

docs/source/zh/pipeline_tutorial.md

@@ -257,13 +257,11 @@ for out in pipe(KeyDataset(dataset, "audio")):
 >>> from transformers import pipeline
 
 >>> vqa = pipeline(model="impira/layoutlm-document-qa")
->>> output = vqa(
+>>> vqa(
 ...     image="https://huggingface.co/spaces/impira/docquery/resolve/2359223c1837a7587402bda0f2643382a6eefeab/invoice.png",
 ...     question="What is the invoice number?",
 ... )
->>> output[0]["score"] = round(output[0]["score"], 3)
->>> output
-[{'score': 0.425, 'answer': 'us-001', 'start': 16, 'end': 16}]
+[{'score': 0.42515, 'answer': 'us-001', 'start': 16, 'end': 16}]
 ```
 
 <Tip>

docs/source/zh/task_summary.md

@@ -332,7 +332,7 @@ score: 0.9327, start: 30, end: 54, answer: huggingface/transformers
 >>> from PIL import Image
 >>> import requests
 
->>> url = "https://huggingface.co/datasets/hf-internal-testing/example-documents/resolve/main/jpeg_images/2.jpg"
+>>> url = "https://datasets-server.huggingface.co/assets/hf-internal-testing/example-documents/--/hf-internal-testing--example-documents/test/2/image/image.jpg"
 >>> image = Image.open(requests.get(url, stream=True).raw)
 
 >>> doc_question_answerer = pipeline("document-question-answering", model="magorshunov/layoutlm-invoices")

examples/pytorch/_tests_requirements.txt

@@ -25,4 +25,3 @@ torchaudio
 jiwer
 librosa
 evaluate >= 0.2.0
-albumentations

examples/pytorch/semantic-segmentation/README.md

@@ -97,10 +97,6 @@ The script leverages the [🤗 Trainer API](https://huggingface.co/docs/transfor
 
 Here we show how to fine-tune a [SegFormer](https://huggingface.co/nvidia/mit-b0) model on the [segments/sidewalk-semantic](https://huggingface.co/datasets/segments/sidewalk-semantic) dataset:
 
-In order to use `segments/sidewalk-semantic`: 
- - Log in to Hugging Face with `huggingface-cli login` (token can be accessed [here](https://huggingface.co/settings/tokens)).
- - Accept terms of use for `sidewalk-semantic` on [dataset page](https://huggingface.co/datasets/segments/sidewalk-semantic).
-
 ```bash
 python run_semantic_segmentation.py \
     --model_name_or_path nvidia/mit-b0 \
@@ -109,6 +105,7 @@ python run_semantic_segmentation.py \
     --remove_unused_columns False \
     --do_train \
     --do_eval \
+    --evaluation_strategy steps \
     --push_to_hub \
     --push_to_hub_model_id segformer-finetuned-sidewalk-10k-steps \
     --max_steps 10000 \

examples/pytorch/semantic-segmentation/requirements.txt

@@ -1,6 +1,4 @@
+git://github.com/huggingface/accelerate.git
 datasets >= 2.0.0
 torch >= 1.3
-accelerate
-evaluate
-Pillow
-albumentations
+evaluate

examples/pytorch/semantic-segmentation/run_semantic_segmentation.py

@@ -16,20 +16,21 @@
 import json
 import logging
 import os
+import random
 import sys
 import warnings
 from dataclasses import dataclass, field
-from functools import partial
 from typing import Optional
 
-import albumentations as A
 import evaluate
 import numpy as np
 import torch
-from albumentations.pytorch import ToTensorV2
 from datasets import load_dataset
 from huggingface_hub import hf_hub_download
+from PIL import Image
 from torch import nn
+from torchvision import transforms
+from torchvision.transforms import functional
 
 import transformers
 from transformers import (
@@ -56,19 +57,118 @@ check_min_version("4.40.0.dev0")
 require_version("datasets>=2.0.0", "To fix: pip install -r examples/pytorch/semantic-segmentation/requirements.txt")
 
 
-def reduce_labels_transform(labels: np.ndarray, **kwargs) -> np.ndarray:
-    """Set `0` label as with value 255 and then reduce all other labels by 1.
+def pad_if_smaller(img, size, fill=0):
+    size = (size, size) if isinstance(size, int) else size
+    original_width, original_height = img.size
+    pad_height = size[1] - original_height if original_height < size[1] else 0
+    pad_width = size[0] - original_width if original_width < size[0] else 0
+    img = functional.pad(img, (0, 0, pad_width, pad_height), fill=fill)
+    return img
 
-    Example:
-        Initial class labels:         0 - background; 1 - road; 2 - car;
-        Transformed class labels:   255 - background; 0 - road; 1 - car;
 
-    **kwargs are required to use this function with albumentations.
-    """
-    labels[labels == 0] = 255
-    labels = labels - 1
-    labels[labels == 254] = 255
-    return labels
+class Compose:
+    def __init__(self, transforms):
+        self.transforms = transforms
+
+    def __call__(self, image, target):
+        for t in self.transforms:
+            image, target = t(image, target)
+        return image, target
+
+
+class Identity:
+    def __init__(self):
+        pass
+
+    def __call__(self, image, target):
+        return image, target
+
+
+class Resize:
+    def __init__(self, size):
+        self.size = size
+
+    def __call__(self, image, target):
+        image = functional.resize(image, self.size)
+        target = functional.resize(target, self.size, interpolation=transforms.InterpolationMode.NEAREST)
+        return image, target
+
+
+class RandomResize:
+    def __init__(self, min_size, max_size=None):
+        self.min_size = min_size
+        if max_size is None:
+            max_size = min_size
+        self.max_size = max_size
+
+    def __call__(self, image, target):
+        size = random.randint(self.min_size, self.max_size)
+        image = functional.resize(image, size)
+        target = functional.resize(target, size, interpolation=transforms.InterpolationMode.NEAREST)
+        return image, target
+
+
+class RandomCrop:
+    def __init__(self, size):
+        self.size = size if isinstance(size, tuple) else (size, size)
+
+    def __call__(self, image, target):
+        image = pad_if_smaller(image, self.size)
+        target = pad_if_smaller(target, self.size, fill=255)
+        crop_params = transforms.RandomCrop.get_params(image, self.size)
+        image = functional.crop(image, *crop_params)
+        target = functional.crop(target, *crop_params)
+        return image, target
+
+
+class RandomHorizontalFlip:
+    def __init__(self, flip_prob):
+        self.flip_prob = flip_prob
+
+    def __call__(self, image, target):
+        if random.random() < self.flip_prob:
+            image = functional.hflip(image)
+            target = functional.hflip(target)
+        return image, target
+
+
+class PILToTensor:
+    def __call__(self, image, target):
+        image = functional.pil_to_tensor(image)
+        target = torch.as_tensor(np.array(target), dtype=torch.int64)
+        return image, target
+
+
+class ConvertImageDtype:
+    def __init__(self, dtype):
+        self.dtype = dtype
+
+    def __call__(self, image, target):
+        image = functional.convert_image_dtype(image, self.dtype)
+        return image, target
+
+
+class Normalize:
+    def __init__(self, mean, std):
+        self.mean = mean
+        self.std = std
+
+    def __call__(self, image, target):
+        image = functional.normalize(image, mean=self.mean, std=self.std)
+        return image, target
+
+
+class ReduceLabels:
+    def __call__(self, image, target):
+        if not isinstance(target, np.ndarray):
+            target = np.array(target).astype(np.uint8)
+        # avoid using underflow conversion
+        target[target == 0] = 255
+        target = target - 1
+        target[target == 254] = 255
+
+        target = Image.fromarray(target)
+        return image, target
 
 
 @dataclass
@@ -265,7 +365,7 @@ def main():
     id2label = {int(k): v for k, v in id2label.items()}
     label2id = {v: str(k) for k, v in id2label.items()}
 
-    # Load the mean IoU metric from the evaluate package
+    # Load the mean IoU metric from the datasets package
     metric = evaluate.load("mean_iou", cache_dir=model_args.cache_dir)
 
     # Define our compute_metrics function. It takes an `EvalPrediction` object (a namedtuple with a
@@ -324,62 +424,64 @@ def main():
         token=model_args.token,
         trust_remote_code=model_args.trust_remote_code,
     )
-    # `reduce_labels` is a property of dataset labels, in case we use image_processor
-    # pretrained on another dataset we should override the default setting
-    image_processor.do_reduce_labels = data_args.reduce_labels
 
-    # Define transforms to be applied to each image and target.
+    # Define torchvision transforms to be applied to each image + target.
+    # Not that straightforward in torchvision: https://github.com/pytorch/vision/issues/9
+    # Currently based on official torchvision references: https://github.com/pytorch/vision/blob/main/references/segmentation/transforms.py
     if "shortest_edge" in image_processor.size:
         # We instead set the target size as (shortest_edge, shortest_edge) to here to ensure all images are batchable.
-        height, width = image_processor.size["shortest_edge"], image_processor.size["shortest_edge"]
+        size = (image_processor.size["shortest_edge"], image_processor.size["shortest_edge"])
     else:
-        height, width = image_processor.size["height"], image_processor.size["width"]
-    train_transforms = A.Compose(
+        size = (image_processor.size["height"], image_processor.size["width"])
+    train_transforms = Compose(
         [
-            A.Lambda(
-                name="reduce_labels",
-                mask=reduce_labels_transform if data_args.reduce_labels else None,
-                p=1.0,
-            ),
-            # pad image with 255, because it is ignored by loss
-            A.PadIfNeeded(min_height=height, min_width=width, border_mode=0, value=255, p=1.0),
-            A.RandomCrop(height=height, width=width, p=1.0),
-            A.HorizontalFlip(p=0.5),
-            A.Normalize(mean=image_processor.image_mean, std=image_processor.image_std, max_pixel_value=255.0, p=1.0),
-            ToTensorV2(),
+            ReduceLabels() if data_args.reduce_labels else Identity(),
+            RandomCrop(size=size),
+            RandomHorizontalFlip(flip_prob=0.5),
+            PILToTensor(),
+            ConvertImageDtype(torch.float),
+            Normalize(mean=image_processor.image_mean, std=image_processor.image_std),
         ]
     )
-    val_transforms = A.Compose(
+    # Define torchvision transform to be applied to each image.
+    # jitter = ColorJitter(brightness=0.25, contrast=0.25, saturation=0.25, hue=0.1)
+    val_transforms = Compose(
         [
-            A.Lambda(
-                name="reduce_labels",
-                mask=reduce_labels_transform if data_args.reduce_labels else None,
-                p=1.0,
-            ),
-            A.Resize(height=height, width=width, p=1.0),
-            A.Normalize(mean=image_processor.image_mean, std=image_processor.image_std, max_pixel_value=255.0, p=1.0),
-            ToTensorV2(),
+            ReduceLabels() if data_args.reduce_labels else Identity(),
+            Resize(size=size),
+            PILToTensor(),
+            ConvertImageDtype(torch.float),
+            Normalize(mean=image_processor.image_mean, std=image_processor.image_std),
         ]
     )
 
-    def preprocess_batch(example_batch, transforms: A.Compose):
+    def preprocess_train(example_batch):
         pixel_values = []
         labels = []
         for image, target in zip(example_batch["image"], example_batch["label"]):
-            transformed = transforms(image=np.array(image.convert("RGB")), mask=np.array(target))
-            pixel_values.append(transformed["image"])
-            labels.append(transformed["mask"])
+            image, target = train_transforms(image.convert("RGB"), target)
+            pixel_values.append(image)
+            labels.append(target)
 
         encoding = {}
-        encoding["pixel_values"] = torch.stack(pixel_values).to(torch.float)
-        encoding["labels"] = torch.stack(labels).to(torch.long)
+        encoding["pixel_values"] = torch.stack(pixel_values)
+        encoding["labels"] = torch.stack(labels)
 
         return encoding
 
-    # Preprocess function for dataset should have only one argument,
-    # so we use partial to pass the transforms
-    preprocess_train_batch_fn = partial(preprocess_batch, transforms=train_transforms)
-    preprocess_val_batch_fn = partial(preprocess_batch, transforms=val_transforms)
+    def preprocess_val(example_batch):
+        pixel_values = []
+        labels = []
+        for image, target in zip(example_batch["image"], example_batch["label"]):
+            image, target = val_transforms(image.convert("RGB"), target)
+            pixel_values.append(image)
+            labels.append(target)
+
+        encoding = {}
+        encoding["pixel_values"] = torch.stack(pixel_values)
+        encoding["labels"] = torch.stack(labels)
+
+        return encoding
 
     if training_args.do_train:
         if "train" not in dataset:
@@ -389,7 +491,7 @@ def main():
                 dataset["train"].shuffle(seed=training_args.seed).select(range(data_args.max_train_samples))
             )
         # Set the training transforms
-        dataset["train"].set_transform(preprocess_train_batch_fn)
+        dataset["train"].set_transform(preprocess_train)
 
     if training_args.do_eval:
         if "validation" not in dataset:
@@ -399,7 +501,7 @@ def main():
                 dataset["validation"].shuffle(seed=training_args.seed).select(range(data_args.max_eval_samples))
             )
         # Set the validation transforms
-        dataset["validation"].set_transform(preprocess_val_batch_fn)
+        dataset["validation"].set_transform(preprocess_val)
 
     # Initialize our trainer
     trainer = Trainer(

examples/pytorch/semantic-segmentation/run_semantic_segmentation_no_trainer.py

@@ -18,10 +18,9 @@ import argparse
 import json
 import math
 import os
-from functools import partial
+import random
 from pathlib import Path
 
-import albumentations as A
 import datasets
 import evaluate
 import numpy as np
@@ -29,10 +28,12 @@ import torch
 from accelerate import Accelerator
 from accelerate.logging import get_logger
 from accelerate.utils import set_seed
-from albumentations.pytorch import ToTensorV2
 from datasets import load_dataset
 from huggingface_hub import HfApi, hf_hub_download
+from PIL import Image
 from torch.utils.data import DataLoader
+from torchvision import transforms
+from torchvision.transforms import functional
 from tqdm.auto import tqdm
 
 import transformers
@@ -56,23 +57,123 @@ logger = get_logger(__name__)
 require_version("datasets>=2.0.0", "To fix: pip install -r examples/pytorch/semantic-segmentation/requirements.txt")
 
 
-def reduce_labels_transform(labels: np.ndarray, **kwargs) -> np.ndarray:
-    """Set `0` label as with value 255 and then reduce all other labels by 1.
+def pad_if_smaller(img, size, fill=0):
+    min_size = min(img.size)
+    if min_size < size:
+        original_width, original_height = img.size
+        pad_height = size - original_height if original_height < size else 0
+        pad_width = size - original_width if original_width < size else 0
+        img = functional.pad(img, (0, 0, pad_width, pad_height), fill=fill)
+    return img
 
-    Example:
-        Initial class labels:         0 - background; 1 - road; 2 - car;
-        Transformed class labels:   255 - background; 0 - road; 1 - car;
 
-    **kwargs are required to use this function with albumentations.
-    """
-    labels[labels == 0] = 255
-    labels = labels - 1
-    labels[labels == 254] = 255
-    return labels
+class Compose:
+    def __init__(self, transforms):
+        self.transforms = transforms
+
+    def __call__(self, image, target):
+        for t in self.transforms:
+            image, target = t(image, target)
+        return image, target
+
+
+class Identity:
+    def __init__(self):
+        pass
+
+    def __call__(self, image, target):
+        return image, target
+
+
+class Resize:
+    def __init__(self, size):
+        self.size = size
+
+    def __call__(self, image, target):
+        image = functional.resize(image, self.size)
+        target = functional.resize(target, self.size, interpolation=transforms.InterpolationMode.NEAREST)
+        return image, target
+
+
+class RandomResize:
+    def __init__(self, min_size, max_size=None):
+        self.min_size = min_size
+        if max_size is None:
+            max_size = min_size
+        self.max_size = max_size
+
+    def __call__(self, image, target):
+        size = random.randint(self.min_size, self.max_size)
+        image = functional.resize(image, size)
+        target = functional.resize(target, size, interpolation=transforms.InterpolationMode.NEAREST)
+        return image, target
+
+
+class RandomCrop:
+    def __init__(self, size):
+        self.size = size
+
+    def __call__(self, image, target):
+        image = pad_if_smaller(image, self.size)
+        target = pad_if_smaller(target, self.size, fill=255)
+        crop_params = transforms.RandomCrop.get_params(image, (self.size, self.size))
+        image = functional.crop(image, *crop_params)
+        target = functional.crop(target, *crop_params)
+        return image, target
+
+
+class RandomHorizontalFlip:
+    def __init__(self, flip_prob):
+        self.flip_prob = flip_prob
+
+    def __call__(self, image, target):
+        if random.random() < self.flip_prob:
+            image = functional.hflip(image)
+            target = functional.hflip(target)
+        return image, target
+
+
+class PILToTensor:
+    def __call__(self, image, target):
+        image = functional.pil_to_tensor(image)
+        target = torch.as_tensor(np.array(target), dtype=torch.int64)
+        return image, target
+
+
+class ConvertImageDtype:
+    def __init__(self, dtype):
+        self.dtype = dtype
+
+    def __call__(self, image, target):
+        image = functional.convert_image_dtype(image, self.dtype)
+        return image, target
+
+
+class Normalize:
+    def __init__(self, mean, std):
+        self.mean = mean
+        self.std = std
+
+    def __call__(self, image, target):
+        image = functional.normalize(image, mean=self.mean, std=self.std)
+        return image, target
+
+
+class ReduceLabels:
+    def __call__(self, image, target):
+        if not isinstance(target, np.ndarray):
+            target = np.array(target).astype(np.uint8)
+        # avoid using underflow conversion
+        target[target == 0] = 255
+        target = target - 1
+        target[target == 254] = 255
+
+        target = Image.fromarray(target)
+        return image, target
 
 
 def parse_args():
-    parser = argparse.ArgumentParser(description="Finetune a transformers model on a image semantic segmentation task")
+    parser = argparse.ArgumentParser(description="Finetune a transformers model on a text classification task")
     parser.add_argument(
         "--model_name_or_path",
         type=str,
@@ -317,58 +418,69 @@ def main():
     model = AutoModelForSemanticSegmentation.from_pretrained(
         args.model_name_or_path, config=config, trust_remote_code=args.trust_remote_code
     )
-    # `reduce_labels` is a property of dataset labels, in case we use image_processor
-    # pretrained on another dataset we should override the default setting
-    image_processor.do_reduce_labels = args.reduce_labels
 
-    # Define transforms to be applied to each image and target.
+    # Preprocessing the datasets
+    # Define torchvision transforms to be applied to each image + target.
+    # Not that straightforward in torchvision: https://github.com/pytorch/vision/issues/9
+    # Currently based on official torchvision references: https://github.com/pytorch/vision/blob/main/references/segmentation/transforms.py
     if "shortest_edge" in image_processor.size:
         # We instead set the target size as (shortest_edge, shortest_edge) to here to ensure all images are batchable.
-        height, width = image_processor.size["shortest_edge"], image_processor.size["shortest_edge"]
+        size = (image_processor.size["shortest_edge"], image_processor.size["shortest_edge"])
     else:
-        height, width = image_processor.size["height"], image_processor.size["width"]
-    train_transforms = A.Compose(
+        size = (image_processor.size["height"], image_processor.size["width"])
+    train_transforms = Compose(
         [
-            A.Lambda(name="reduce_labels", mask=reduce_labels_transform if args.reduce_labels else None, p=1.0),
-            # pad image with 255, because it is ignored by loss
-            A.PadIfNeeded(min_height=height, min_width=width, border_mode=0, value=255, p=1.0),
-            A.RandomCrop(height=height, width=width, p=1.0),
-            A.HorizontalFlip(p=0.5),
-            A.Normalize(mean=image_processor.image_mean, std=image_processor.image_std, max_pixel_value=255.0, p=1.0),
-            ToTensorV2(),
+            ReduceLabels() if args.reduce_labels else Identity(),
+            RandomCrop(size=size),
+            RandomHorizontalFlip(flip_prob=0.5),
+            PILToTensor(),
+            ConvertImageDtype(torch.float),
+            Normalize(mean=image_processor.image_mean, std=image_processor.image_std),
         ]
     )
-    val_transforms = A.Compose(
+    # Define torchvision transform to be applied to each image.
+    # jitter = ColorJitter(brightness=0.25, contrast=0.25, saturation=0.25, hue=0.1)
+    val_transforms = Compose(
         [
-            A.Lambda(name="reduce_labels", mask=reduce_labels_transform if args.reduce_labels else None, p=1.0),
-            A.Resize(height=height, width=width, p=1.0),
-            A.Normalize(mean=image_processor.image_mean, std=image_processor.image_std, max_pixel_value=255.0, p=1.0),
-            ToTensorV2(),
+            ReduceLabels() if args.reduce_labels else Identity(),
+            Resize(size=size),
+            PILToTensor(),
+            ConvertImageDtype(torch.float),
+            Normalize(mean=image_processor.image_mean, std=image_processor.image_std),
         ]
     )
 
-    def preprocess_batch(example_batch, transforms: A.Compose):
+    def preprocess_train(example_batch):
         pixel_values = []
         labels = []
         for image, target in zip(example_batch["image"], example_batch["label"]):
-            transformed = transforms(image=np.array(image.convert("RGB")), mask=np.array(target))
-            pixel_values.append(transformed["image"])
-            labels.append(transformed["mask"])
+            image, target = train_transforms(image.convert("RGB"), target)
+            pixel_values.append(image)
+            labels.append(target)
 
         encoding = {}
-        encoding["pixel_values"] = torch.stack(pixel_values).to(torch.float)
-        encoding["labels"] = torch.stack(labels).to(torch.long)
+        encoding["pixel_values"] = torch.stack(pixel_values)
+        encoding["labels"] = torch.stack(labels)
 
         return encoding
 
-    # Preprocess function for dataset should have only one input argument,
-    # so we use partial to pass transforms
-    preprocess_train_batch_fn = partial(preprocess_batch, transforms=train_transforms)
-    preprocess_val_batch_fn = partial(preprocess_batch, transforms=val_transforms)
+    def preprocess_val(example_batch):
+        pixel_values = []
+        labels = []
+        for image, target in zip(example_batch["image"], example_batch["label"]):
+            image, target = val_transforms(image.convert("RGB"), target)
+            pixel_values.append(image)
+            labels.append(target)
+
+        encoding = {}
+        encoding["pixel_values"] = torch.stack(pixel_values)
+        encoding["labels"] = torch.stack(labels)
+
+        return encoding
 
     with accelerator.main_process_first():
-        train_dataset = dataset["train"].with_transform(preprocess_train_batch_fn)
-        eval_dataset = dataset["validation"].with_transform(preprocess_val_batch_fn)
+        train_dataset = dataset["train"].with_transform(preprocess_train)
+        eval_dataset = dataset["validation"].with_transform(preprocess_val)
 
     train_dataloader = DataLoader(
         train_dataset, shuffle=True, collate_fn=default_data_collator, batch_size=args.per_device_train_batch_size
@@ -614,7 +726,7 @@ def main():
                 f"eval_{k}": v.tolist() if isinstance(v, np.ndarray) else v for k, v in eval_metrics.items()
             }
             with open(os.path.join(args.output_dir, "all_results.json"), "w") as f:
-                json.dump(all_results, f, indent=2)
+                json.dump(all_results, f)
 
 
 if __name__ == "__main__":

examples/pytorch/test_pytorch_examples.py

@@ -372,7 +372,6 @@ class ExamplesTests(TestCasePlus):
             --predict_with_generate
             --source_lang en_XX
             --target_lang ro_RO
-            --max_source_length 512
         """.split()
 
         with patch.object(sys, "argv", testargs):

examples/pytorch/text-classification/run_glue_no_trainer.py

@@ -327,9 +327,6 @@ def main():
     tokenizer = AutoTokenizer.from_pretrained(
         args.model_name_or_path, use_fast=not args.use_slow_tokenizer, trust_remote_code=args.trust_remote_code
     )
-    if tokenizer.pad_token is None:
-        tokenizer.pad_token = tokenizer.eos_token
-    config.pad_token_id = tokenizer.pad_token_id
     model = AutoModelForSequenceClassification.from_pretrained(
         args.model_name_or_path,
         from_tf=bool(".ckpt" in args.model_name_or_path),

examples/research_projects/tapex/wikisql_utils.py

@@ -21,7 +21,7 @@ import re
 
 # The following script is adapted from the script of TaPas.
 # Original: https://github.com/google-research/tapas/master/wikisql_utils.py
-from typing import Any, List
+from typing import Any, List, Text
 
 
 EMPTY_ANSWER = "none"
@@ -114,7 +114,7 @@ class _Operator(enum.Enum):
 class _Condition:
     """Represents an SQL where clauses (e.g A = "a" or B > 5)."""
 
-    column: str
+    column: Text
     operator: _Operator
     cmp_value: Any
 

pyproject.toml

@@ -1,18 +1,16 @@
 [tool.ruff]
-line-length = 119
-
-[tool.ruff.lint]
 # Never enforce `E501` (line length violations).
 ignore = ["C901", "E501", "E741", "F402", "F823" ]
 select = ["C", "E", "F", "I", "W"]
+line-length = 119
 
 # Ignore import violations in all `__init__.py` files.
-[tool.ruff.lint.per-file-ignores]
+[tool.ruff.per-file-ignores]
 "__init__.py" = ["E402", "F401", "F403", "F811"]
 "src/transformers/file_utils.py" = ["F401"]
 "src/transformers/utils/dummy_*.py" = ["F401"]
 
-[tool.ruff.lint.isort]
+[tool.ruff.isort]
 lines-after-imports = 2
 known-first-party = ["transformers"]
 
@@ -35,4 +33,4 @@ doctest_glob="**/*.md"
 markers = [
     "flash_attn_test: marks tests related to flash attention (deselect with '-m \"not flash_attn_test\"')",
     "bitsandbytes: select (or deselect with `not`) bitsandbytes integration tests",
-]
+]

src/transformers/__init__.py

@@ -488,11 +488,9 @@ _import_structure = {
         "GPTSanJapaneseConfig",
         "GPTSanJapaneseTokenizer",
     ],
-    "models.graphormer": ["GRAPHORMER_PRETRAINED_CONFIG_ARCHIVE_MAP", "GraphormerConfig"],
-    "models.grounding_dino": [
-        "GROUNDING_DINO_PRETRAINED_CONFIG_ARCHIVE_MAP",
-        "GroundingDinoConfig",
-        "GroundingDinoProcessor",
+    "models.graphormer": [
+        "GRAPHORMER_PRETRAINED_CONFIG_ARCHIVE_MAP",
+        "GraphormerConfig",
     ],
     "models.groupvit": [
         "GROUPVIT_PRETRAINED_CONFIG_ARCHIVE_MAP",
@@ -507,7 +505,6 @@ _import_structure = {
         "IDEFICS_PRETRAINED_CONFIG_ARCHIVE_MAP",
         "IdeficsConfig",
     ],
-    "models.idefics2": ["Idefics2Config"],
     "models.imagegpt": ["IMAGEGPT_PRETRAINED_CONFIG_ARCHIVE_MAP", "ImageGPTConfig"],
     "models.informer": ["INFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP", "InformerConfig"],
     "models.instructblip": [
@@ -746,7 +743,6 @@ _import_structure = {
         "RealmConfig",
         "RealmTokenizer",
     ],
-    "models.recurrent_gemma": ["RecurrentGemmaConfig"],
     "models.reformer": ["REFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP", "ReformerConfig"],
     "models.regnet": ["REGNET_PRETRAINED_CONFIG_ARCHIVE_MAP", "RegNetConfig"],
     "models.rembert": ["REMBERT_PRETRAINED_CONFIG_ARCHIVE_MAP", "RemBertConfig"],
@@ -1334,9 +1330,7 @@ else:
     _import_structure["models.flava"].extend(["FlavaFeatureExtractor", "FlavaImageProcessor", "FlavaProcessor"])
     _import_structure["models.fuyu"].extend(["FuyuImageProcessor", "FuyuProcessor"])
     _import_structure["models.glpn"].extend(["GLPNFeatureExtractor", "GLPNImageProcessor"])
-    _import_structure["models.grounding_dino"].extend(["GroundingDinoImageProcessor"])
     _import_structure["models.idefics"].extend(["IdeficsImageProcessor"])
-    _import_structure["models.idefics2"].extend(["Idefics2ImageProcessor"])
     _import_structure["models.imagegpt"].extend(["ImageGPTFeatureExtractor", "ImageGPTImageProcessor"])
     _import_structure["models.layoutlmv2"].extend(["LayoutLMv2FeatureExtractor", "LayoutLMv2ImageProcessor"])
     _import_structure["models.layoutlmv3"].extend(["LayoutLMv3FeatureExtractor", "LayoutLMv3ImageProcessor"])
@@ -2396,14 +2390,6 @@ else:
             "GraphormerPreTrainedModel",
         ]
     )
-    _import_structure["models.grounding_dino"].extend(
-        [
-            "GROUNDING_DINO_PRETRAINED_MODEL_ARCHIVE_LIST",
-            "GroundingDinoForObjectDetection",
-            "GroundingDinoModel",
-            "GroundingDinoPreTrainedModel",
-        ]
-    )
     _import_structure["models.groupvit"].extend(
         [
             "GROUPVIT_PRETRAINED_MODEL_ARCHIVE_LIST",
@@ -2443,15 +2429,6 @@ else:
             "IdeficsProcessor",
         ]
     )
-    _import_structure["models.idefics2"].extend(
-        [
-            "IDEFICS2_PRETRAINED_MODEL_ARCHIVE_LIST",
-            "Idefics2ForConditionalGeneration",
-            "Idefics2Model",
-            "Idefics2PreTrainedModel",
-            "Idefics2Processor",
-        ]
-    )
     _import_structure["models.imagegpt"].extend(
         [
             "IMAGEGPT_PRETRAINED_MODEL_ARCHIVE_LIST",
@@ -3138,13 +3115,6 @@ else:
             "load_tf_weights_in_realm",
         ]
     )
-    _import_structure["models.recurrent_gemma"].extend(
-        [
-            "RecurrentGemmaForCausalLM",
-            "RecurrentGemmaModel",
-            "RecurrentGemmaPreTrainedModel",
-        ]
-    )
     _import_structure["models.reformer"].extend(
         [
             "REFORMER_PRETRAINED_MODEL_ARCHIVE_LIST",
@@ -5402,11 +5372,9 @@ if TYPE_CHECKING:
         GPTSanJapaneseConfig,
         GPTSanJapaneseTokenizer,
     )
-    from .models.graphormer import GRAPHORMER_PRETRAINED_CONFIG_ARCHIVE_MAP, GraphormerConfig
-    from .models.grounding_dino import (
-        GROUNDING_DINO_PRETRAINED_CONFIG_ARCHIVE_MAP,
-        GroundingDinoConfig,
-        GroundingDinoProcessor,
+    from .models.graphormer import (
+        GRAPHORMER_PRETRAINED_CONFIG_ARCHIVE_MAP,
+        GraphormerConfig,
     )
     from .models.groupvit import (
         GROUPVIT_PRETRAINED_CONFIG_ARCHIVE_MAP,
@@ -5421,7 +5389,6 @@ if TYPE_CHECKING:
         IDEFICS_PRETRAINED_CONFIG_ARCHIVE_MAP,
         IdeficsConfig,
     )
-    from .models.idefics2 import Idefics2Config
     from .models.imagegpt import IMAGEGPT_PRETRAINED_CONFIG_ARCHIVE_MAP, ImageGPTConfig
     from .models.informer import INFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP, InformerConfig
     from .models.instructblip import (
@@ -5658,7 +5625,6 @@ if TYPE_CHECKING:
         RealmConfig,
         RealmTokenizer,
     )
-    from .models.recurrent_gemma import RecurrentGemmaConfig
     from .models.reformer import REFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP, ReformerConfig
     from .models.regnet import REGNET_PRETRAINED_CONFIG_ARCHIVE_MAP, RegNetConfig
     from .models.rembert import REMBERT_PRETRAINED_CONFIG_ARCHIVE_MAP, RemBertConfig
@@ -6220,9 +6186,7 @@ if TYPE_CHECKING:
         )
         from .models.fuyu import FuyuImageProcessor, FuyuProcessor
         from .models.glpn import GLPNFeatureExtractor, GLPNImageProcessor
-        from .models.grounding_dino import GroundingDinoImageProcessor
         from .models.idefics import IdeficsImageProcessor
-        from .models.idefics2 import Idefics2ImageProcessor
         from .models.imagegpt import ImageGPTFeatureExtractor, ImageGPTImageProcessor
         from .models.layoutlmv2 import (
             LayoutLMv2FeatureExtractor,
@@ -7139,12 +7103,6 @@ if TYPE_CHECKING:
             GraphormerModel,
             GraphormerPreTrainedModel,
         )
-        from .models.grounding_dino import (
-            GROUNDING_DINO_PRETRAINED_MODEL_ARCHIVE_LIST,
-            GroundingDinoForObjectDetection,
-            GroundingDinoModel,
-            GroundingDinoPreTrainedModel,
-        )
         from .models.groupvit import (
             GROUPVIT_PRETRAINED_MODEL_ARCHIVE_LIST,
             GroupViTModel,
@@ -7176,13 +7134,6 @@ if TYPE_CHECKING:
             IdeficsPreTrainedModel,
             IdeficsProcessor,
         )
-        from .models.idefics2 import (
-            IDEFICS2_PRETRAINED_MODEL_ARCHIVE_LIST,
-            Idefics2ForConditionalGeneration,
-            Idefics2Model,
-            Idefics2PreTrainedModel,
-            Idefics2Processor,
-        )
         from .models.imagegpt import (
             IMAGEGPT_PRETRAINED_MODEL_ARCHIVE_LIST,
             ImageGPTForCausalImageModeling,
@@ -7736,11 +7687,6 @@ if TYPE_CHECKING:
             RealmScorer,
             load_tf_weights_in_realm,
         )
-        from .models.recurrent_gemma import (
-            RecurrentGemmaForCausalLM,
-            RecurrentGemmaModel,
-            RecurrentGemmaPreTrainedModel,
-        )
         from .models.reformer import (
             REFORMER_PRETRAINED_MODEL_ARCHIVE_LIST,
             ReformerAttention,

src/transformers/generation/candidate_generator.py

@@ -148,11 +148,6 @@ class AssistedCandidateGenerator(CandidateGenerator):
         self.generation_config.return_dict_in_generate = True
         self.generation_config.output_scores = True
 
-        # avoid unnecessary warnings that min_length is larger than max_new_tokens
-        self.main_model_min_length = self.generation_config.min_length
-        self.generation_config.min_length = 0
-        self.generation_config.min_new_tokens = None
-
     def get_candidates(self, input_ids: torch.LongTensor) -> Tuple[torch.LongTensor, Optional[torch.FloatTensor]]:
         """
         Fetches the candidates to be tried for the current input.
@@ -171,7 +166,6 @@ class AssistedCandidateGenerator(CandidateGenerator):
         # Don't generate more than `max_length - 1` candidates since the target model generates one extra token.
         new_cur_len = input_ids.shape[-1]
         max_new_tokens = min(int(self.num_assistant_tokens), self.generation_config.max_length - new_cur_len - 1)
-        min_new_tokens = max(min(max_new_tokens, self.main_model_min_length - new_cur_len), 0)
         if max_new_tokens == 0:
             return input_ids, None
 
@@ -192,7 +186,6 @@ class AssistedCandidateGenerator(CandidateGenerator):
         # 2. Forecast next N tokens using the assistant model.
         assistant_generation_kwargs = {
             self.input_ids_key: input_ids,
-            "min_new_tokens": min_new_tokens,
             "max_new_tokens": max_new_tokens,
             "generation_config": self.generation_config,
             "logits_processor": self.logits_processor,

src/transformers/generation/utils.py

@@ -1173,56 +1173,6 @@ class GenerationMixin:
                     UserWarning,
                 )
 
-    def _prepare_generated_length(
-        self,
-        generation_config,
-        has_default_max_length,
-        has_default_min_length,
-        model_input_name,
-        input_ids_length,
-        inputs_tensor,
-    ):
-        """Prepared max and min length in generaion configs to avoid clashes between similar attributes"""
-
-        if generation_config.max_new_tokens is not None:
-            if not has_default_max_length and generation_config.max_length is not None:
-                logger.warning(
-                    f"Both `max_new_tokens` (={generation_config.max_new_tokens}) and `max_length`(="
-                    f"{generation_config.max_length}) seem to have been set. `max_new_tokens` will take precedence. "
-                    "Please refer to the documentation for more information. "
-                    "(https://huggingface.co/docs/transformers/main/en/main_classes/text_generation)"
-                )
-            generation_config.max_length = generation_config.max_new_tokens + input_ids_length
-
-        # if both `inputs_embeds` and `input_ids` are passed, we do not correct the length
-        # otherwise we need total length [inputs-embeds-len + new-tokens-len] to not go beyond indicated `max_length``
-        elif (
-            model_input_name == "inputs_embeds"
-            and input_ids_length != inputs_tensor.shape[1]
-            and not self.config.is_encoder_decoder
-        ):
-            generation_config.max_length -= inputs_tensor.shape[1]
-
-        # same for min length
-        if generation_config.min_new_tokens is not None:
-            if not has_default_min_length:
-                logger.warning(
-                    f"Both `min_new_tokens` (={generation_config.min_new_tokens}) and `min_length`(="
-                    f"{generation_config.min_length}) seem to have been set. `min_new_tokens` will take precedence. "
-                    "Please refer to the documentation for more information. "
-                    "(https://huggingface.co/docs/transformers/main/en/main_classes/text_generation)"
-                )
-            generation_config.min_length = generation_config.min_new_tokens + input_ids_length
-
-        elif (
-            model_input_name == "inputs_embeds"
-            and input_ids_length != inputs_tensor.shape[1]
-            and not self.config.is_encoder_decoder
-        ):
-            generation_config.min_length = max(generation_config.min_length - inputs_tensor.shape[1], 0)
-
-        return generation_config
-
     def _prepare_generation_config(
         self, generation_config: GenerationConfig, **kwargs: Dict
     ) -> Tuple[GenerationConfig, Dict]:
@@ -1468,15 +1418,24 @@ class GenerationMixin:
         # 6. Prepare `max_length` depending on other stopping criteria.
         input_ids_length = input_ids.shape[-1]
         has_default_max_length = kwargs.get("max_length") is None and generation_config.max_length is not None
-        has_default_min_length = kwargs.get("min_length") is None and generation_config.min_length is not None
-        generation_config = self._prepare_generated_length(
-            generation_config=generation_config,
-            has_default_max_length=has_default_max_length,
-            has_default_min_length=has_default_min_length,
-            model_input_name=model_input_name,
-            inputs_tensor=inputs_tensor,
-            input_ids_length=input_ids_length,
-        )
+        if generation_config.max_new_tokens is not None:
+            if not has_default_max_length and generation_config.max_length is not None:
+                logger.warning(
+                    f"Both `max_new_tokens` (={generation_config.max_new_tokens}) and `max_length`(="
+                    f"{generation_config.max_length}) seem to have been set. `max_new_tokens` will take precedence. "
+                    "Please refer to the documentation for more information. "
+                    "(https://huggingface.co/docs/transformers/main/en/main_classes/text_generation)"
+                )
+            generation_config.max_length = generation_config.max_new_tokens + input_ids_length
+
+        # otherwise the total length [inputs-embeds-len + new-tokens-len] will go beyond indicated `max_length``
+        elif (
+            model_input_name == "inputs_embeds"
+            and inputs_tensor.shape[:-1] != input_ids.shape
+            and not self.config.is_encoder_decoder
+        ):
+            generation_config.max_length -= inputs_tensor.shape[1]
+            generation_config.min_length = max(generation_config.min_length - inputs_tensor.shape[1], 0)
 
         if generation_config.cache_implementation in NEED_SETUP_CACHE_CLASSES_MAPPING:
             if generation_config.cache_implementation == "static":
@@ -1552,7 +1511,7 @@ class GenerationMixin:
             )
 
             # 12. run assisted generate
-            result = self._assisted_decoding(
+            result = self.assisted_decoding(
                 input_ids,
                 candidate_generator=candidate_generator,
                 do_sample=generation_config.do_sample,

src/transformers/hf_argparser.py

@@ -14,7 +14,6 @@
 
 import dataclasses
 import json
-import os
 import sys
 import types
 from argparse import ArgumentDefaultsHelpFormatter, ArgumentParser, ArgumentTypeError
@@ -377,9 +376,7 @@ class HfArgumentParser(ArgumentParser):
             raise ValueError(f"Some keys are not used by the HfArgumentParser: {sorted(unused_keys)}")
         return tuple(outputs)
 
-    def parse_json_file(
-        self, json_file: Union[str, os.PathLike], allow_extra_keys: bool = False
-    ) -> Tuple[DataClass, ...]:
+    def parse_json_file(self, json_file: str, allow_extra_keys: bool = False) -> Tuple[DataClass, ...]:
         """
         Alternative helper method that does not use `argparse` at all, instead loading a json file and populating the
         dataclass types.
@@ -401,9 +398,7 @@ class HfArgumentParser(ArgumentParser):
         outputs = self.parse_dict(data, allow_extra_keys=allow_extra_keys)
         return tuple(outputs)
 
-    def parse_yaml_file(
-        self, yaml_file: Union[str, os.PathLike], allow_extra_keys: bool = False
-    ) -> Tuple[DataClass, ...]:
+    def parse_yaml_file(self, yaml_file: str, allow_extra_keys: bool = False) -> Tuple[DataClass, ...]:
         """
         Alternative helper method that does not use `argparse` at all, instead loading a yaml file and populating the
         dataclass types.

src/transformers/image_transforms.py

@@ -749,6 +749,7 @@ def convert_to_rgb(image: ImageInput) -> ImageInput:
     """
     Converts an image to RGB format. Only converts if the image is of type PIL.Image.Image, otherwise returns the image
     as is.
+
     Args:
         image (Image):
             The image to convert.
@@ -758,9 +759,6 @@ def convert_to_rgb(image: ImageInput) -> ImageInput:
     if not isinstance(image, PIL.Image.Image):
         return image
 
-    if image.mode == "RGB":
-        return image
-
     image = image.convert("RGB")
     return image
 

src/transformers/image_utils.py

@@ -311,7 +311,7 @@ def load_image(image: Union[str, "PIL.Image.Image"], timeout: Optional[float] =
         if image.startswith("http://") or image.startswith("https://"):
             # We need to actually check for a real protocol, otherwise it's impossible to use a local file
             # like http_huggingface_co.png
-            image = PIL.Image.open(BytesIO(requests.get(image, timeout=timeout).content))
+            image = PIL.Image.open(requests.get(image, stream=True, timeout=timeout).raw)
         elif os.path.isfile(image):
             image = PIL.Image.open(image)
         else:

src/transformers/integrations/integration_utils.py

@@ -1071,8 +1071,6 @@ class MLflowCallback(TrainerCallback):
             for k, v in logs.items():
                 if isinstance(v, (int, float)):
                     metrics[k] = v
-                elif isinstance(v, torch.Tensor) and v.numel() == 1:
-                    metrics[k] = v.item()
                 else:
                     logger.warning(
                         f'Trainer is attempting to log a value of "{v}" of type {type(v)} for key "{k}" as a metric. '

src/transformers/modeling_attn_mask_utils.py

@@ -305,7 +305,7 @@ def _prepare_4d_causal_attention_mask_for_sdpa(
     attn_mask_converter = AttentionMaskConverter(is_causal=True, sliding_window=sliding_window)
 
     key_value_length = input_shape[-1] + past_key_values_length
-    _, query_length = input_shape
+    batch_size, query_length = input_shape
 
     # torch.jit.trace, symbolic_trace and torchdynamo with fullgraph=True are unable to capture the controlflow `is_causal=attention_mask is None and q_len > 1`
     # used as an SDPA argument. We keep compatibility with these tracing tools by always using SDPA's `attn_mask` argument in case we are tracing.
@@ -316,12 +316,7 @@ def _prepare_4d_causal_attention_mask_for_sdpa(
         or (hasattr(torch, "_dynamo") and torch._dynamo.is_compiling())
     )
 
-    ignore_causal_mask = False
-
-    if attention_mask is None:
-        if sliding_window is None or key_value_length < sliding_window:
-            ignore_causal_mask = not is_tracing
-    elif sliding_window is None or key_value_length < sliding_window:
+    if attention_mask is not None:
         # 4d mask is passed through
         if len(attention_mask.shape) == 4:
             expected_shape = (input_shape[0], 1, input_shape[1], key_value_length)
@@ -340,17 +335,26 @@ def _prepare_4d_causal_attention_mask_for_sdpa(
         elif not is_tracing and torch.all(attention_mask == 1):
             if query_length == 1:
                 # For query_length == 1, causal attention and bi-directional attention are the same.
-                ignore_causal_mask = True
+                attention_mask = None
             elif key_value_length == query_length:
-                ignore_causal_mask = True
+                attention_mask = None
+            else:
+                # Unfortunately, for query_length > 1 and key_value_length != query_length, we cannot generally ignore the attention mask, as SDPA causal mask generation
+                # may be wrong. We will set `is_causal=False` in SDPA and rely on Transformers attention_mask instead, hence not setting it to None here.
+                # Reference: https://github.com/pytorch/pytorch/issues/108108
+                pass
+    elif query_length > 1 and key_value_length != query_length:
+        # See the comment above (https://github.com/pytorch/pytorch/issues/108108).
+        # Ugly: we set it to True here to dispatch in the following controlflow to `to_causal_4d`.
+        attention_mask = True
+    elif is_tracing:
+        raise ValueError(
+            'Attention using SDPA can not be traced with torch.jit.trace when no attention_mask is provided. To solve this issue, please either load your model with the argument `attn_implementation="eager"` or pass an attention_mask input when tracing the model.'
+        )
 
-            # Unfortunately, for query_length > 1 and key_value_length != query_length, we cannot generally ignore the attention mask, as SDPA causal mask generation
-            # may be wrong. We will set `is_causal=False` in SDPA and rely on Transformers attention_mask instead, hence not setting it to None here.
-            # Reference: https://github.com/pytorch/pytorch/issues/108108
-
-    if ignore_causal_mask:
+    if attention_mask is None:
         expanded_4d_mask = None
-    elif attention_mask is None:
+    elif attention_mask is True:
         expanded_4d_mask = attn_mask_converter.to_causal_4d(
             input_shape[0], input_shape[-1], key_value_length, dtype=inputs_embeds.dtype, device=inputs_embeds.device
         )

src/transformers/modeling_utils.py

@@ -1159,12 +1159,9 @@ class ModuleUtilsMixin:
                 # For 4bit models, we need to multiply the number of parameters by 2 as half of the parameters are
                 # used for the 4bit quantization (uint8 tensors are stored)
                 if is_loaded_in_4bit and isinstance(param, bnb.nn.Params4bit):
-                    quant_storage = self.hf_quantizer.quantization_config.bnb_4bit_quant_storage
-                    # For compatibility with older PT version - see: https://github.com/huggingface/peft/pull/1635
-                    nb_params = (
-                        quant_storage.itemsize if hasattr(quant_storage, "itemsize") else quant_storage.element_size()
+                    total_numel.append(
+                        param.numel() * 2 * self.hf_quantizer.quantization_config.bnb_4bit_quant_storage.itemsize
                     )
-                    total_numel.append(param.numel() * 2 * nb_params)
                 else:
                     total_numel.append(param.numel())
 
@@ -2225,7 +2222,7 @@ class PreTrainedModel(nn.Module, ModuleUtilsMixin, GenerationMixin, PushToHubMix
             self._set_gradient_checkpointing(enable=True, gradient_checkpointing_func=gradient_checkpointing_func)
         else:
             self.apply(partial(self._set_gradient_checkpointing, value=True))
-            logger.warning(
+            logger.warn(
                 "You are using an old version of the checkpointing format that is deprecated (We will also silently ignore `gradient_checkpointing_kwargs` in case you passed it)."
                 "Please update to the new format on your modeling file. To use the new format, you need to completely remove the definition of the method `_set_gradient_checkpointing` in your model."
             )
@@ -2273,7 +2270,7 @@ class PreTrainedModel(nn.Module, ModuleUtilsMixin, GenerationMixin, PushToHubMix
             if not _is_using_old_format:
                 self._set_gradient_checkpointing(enable=False)
             else:
-                logger.warning(
+                logger.warn(
                     "You are using an old version of the checkpointing format that is deprecated (We will also silently ignore `gradient_checkpointing_kwargs` in case you passed it)."
                     "Please update to the new format on your modeling file. To use the new format, you need to completely remove the definition of the method `_set_gradient_checkpointing` in your model."
                 )

src/transformers/models/__init__.py

@@ -105,13 +105,11 @@ from . import (
     gptj,
     gptsan_japanese,
     graphormer,
-    grounding_dino,
     groupvit,
     herbert,
     hubert,
     ibert,
     idefics,
-    idefics2,
     imagegpt,
     informer,
     instructblip,
@@ -189,7 +187,6 @@ from . import (
     qwen2_moe,
     rag,
     realm,
-    recurrent_gemma,
     reformer,
     regnet,
     rembert,

src/transformers/models/align/processing_align.py

@@ -57,7 +57,8 @@ class AlignProcessor(ProcessorMixin):
                 `is_split_into_words=True` (to lift the ambiguity with a batch of sequences).
             images (`PIL.Image.Image`, `np.ndarray`, `torch.Tensor`, `List[PIL.Image.Image]`, `List[np.ndarray]`, `List[torch.Tensor]`):
                 The image or batch of images to be prepared. Each image can be a PIL image, NumPy array or PyTorch
-                tensor. Both channels-first and channels-last formats are supported.
+                tensor. In case of a NumPy array/PyTorch tensor, each image should be of shape (C, H, W), where C is a
+                number of channels, H and W are image height and width.
             padding (`bool`, `str` or [`~utils.PaddingStrategy`], *optional*, defaults to `max_length`):
                 Activates and controls padding for tokenization of input text. Choose between [`True` or `'longest'`,
                 `'max_length'`, `False` or `'do_not_pad'`]

src/transformers/models/altclip/processing_altclip.py

@@ -73,7 +73,8 @@ class AltCLIPProcessor(ProcessorMixin):
                 `is_split_into_words=True` (to lift the ambiguity with a batch of sequences).
             images (`PIL.Image.Image`, `np.ndarray`, `torch.Tensor`, `List[PIL.Image.Image]`, `List[np.ndarray]`, `List[torch.Tensor]`):
                 The image or batch of images to be prepared. Each image can be a PIL image, NumPy array or PyTorch
-                tensor. Both channels-first and channels-last formats are supported.
+                tensor. In case of a NumPy array/PyTorch tensor, each image should be of shape (C, H, W), where C is a
+                number of channels, H and W are image height and width.
 
             return_tensors (`str` or [`~utils.TensorType`], *optional*):
                 If set, will return tensors of a particular framework. Acceptable values are:

src/transformers/models/auto/auto_factory.py

@@ -579,7 +579,7 @@ class _BaseAutoModelClass:
             model_class ([`PreTrainedModel`]):
                 The model to register.
         """
-        if hasattr(model_class, "config_class") and str(model_class.config_class) != str(config_class):
+        if hasattr(model_class, "config_class") and model_class.config_class != config_class:
             raise ValueError(
                 "The model class you are passing has a `config_class` attribute that is not consistent with the "
                 f"config class you passed (model has {model_class.config_class} and you passed {config_class}. Fix "

src/transformers/models/auto/configuration_auto.py

@@ -120,12 +120,10 @@ CONFIG_MAPPING_NAMES = OrderedDict(
         ("gptj", "GPTJConfig"),
         ("gptsan-japanese", "GPTSanJapaneseConfig"),
         ("graphormer", "GraphormerConfig"),
-        ("grounding-dino", "GroundingDinoConfig"),
         ("groupvit", "GroupViTConfig"),
         ("hubert", "HubertConfig"),
         ("ibert", "IBertConfig"),
         ("idefics", "IdeficsConfig"),
-        ("idefics2", "Idefics2Config"),
         ("imagegpt", "ImageGPTConfig"),
         ("informer", "InformerConfig"),
         ("instructblip", "InstructBlipConfig"),
@@ -200,7 +198,6 @@ CONFIG_MAPPING_NAMES = OrderedDict(
         ("qwen2_moe", "Qwen2MoeConfig"),
         ("rag", "RagConfig"),
         ("realm", "RealmConfig"),
-        ("recurrent_gemma", "RecurrentGemmaConfig"),
         ("reformer", "ReformerConfig"),
         ("regnet", "RegNetConfig"),
         ("rembert", "RemBertConfig"),
@@ -284,7 +281,6 @@ CONFIG_MAPPING_NAMES = OrderedDict(
     ]
 )
 
-
 MODEL_NAMES_MAPPING = OrderedDict(
     [
         # Add full (and cased) model names here
@@ -386,13 +382,11 @@ MODEL_NAMES_MAPPING = OrderedDict(
         ("gptj", "GPT-J"),
         ("gptsan-japanese", "GPTSAN-japanese"),
         ("graphormer", "Graphormer"),
-        ("grounding-dino", "Grounding DINO"),
         ("groupvit", "GroupViT"),
         ("herbert", "HerBERT"),
         ("hubert", "Hubert"),
         ("ibert", "I-BERT"),
         ("idefics", "IDEFICS"),
-        ("idefics2", "Idefics2"),
         ("imagegpt", "ImageGPT"),
         ("informer", "Informer"),
         ("instructblip", "InstructBLIP"),
@@ -477,7 +471,6 @@ MODEL_NAMES_MAPPING = OrderedDict(
         ("qwen2_moe", "Qwen2MoE"),
         ("rag", "RAG"),
         ("realm", "REALM"),
-        ("recurrent_gemma", "RecurrentGemma"),
         ("reformer", "Reformer"),
         ("regnet", "RegNet"),
         ("rembert", "RemBERT"),

src/transformers/models/auto/image_processing_auto.py

@@ -68,10 +68,8 @@ IMAGE_PROCESSOR_MAPPING_NAMES = OrderedDict(
         ("fuyu", "FuyuImageProcessor"),
         ("git", "CLIPImageProcessor"),
         ("glpn", "GLPNImageProcessor"),
-        ("grounding-dino", "GroundingDinoImageProcessor"),
         ("groupvit", "CLIPImageProcessor"),
         ("idefics", "IdeficsImageProcessor"),
-        ("idefics2", "Idefics2ImageProcessor"),
         ("imagegpt", "ImageGPTImageProcessor"),
         ("instructblip", "BlipImageProcessor"),
         ("kosmos-2", "CLIPImageProcessor"),

src/transformers/models/auto/modeling_auto.py

@@ -115,12 +115,10 @@ MODEL_MAPPING_NAMES = OrderedDict(
         ("gptj", "GPTJModel"),
         ("gptsan-japanese", "GPTSanJapaneseForConditionalGeneration"),
         ("graphormer", "GraphormerModel"),
-        ("grounding-dino", "GroundingDinoModel"),
         ("groupvit", "GroupViTModel"),
         ("hubert", "HubertModel"),
         ("ibert", "IBertModel"),
         ("idefics", "IdeficsModel"),
-        ("idefics2", "Idefics2Model"),
         ("imagegpt", "ImageGPTModel"),
         ("informer", "InformerModel"),
         ("jukebox", "JukeboxModel"),
@@ -185,7 +183,6 @@ MODEL_MAPPING_NAMES = OrderedDict(
         ("qdqbert", "QDQBertModel"),
         ("qwen2", "Qwen2Model"),
         ("qwen2_moe", "Qwen2MoeModel"),
-        ("recurrent_gemma", "RecurrentGemmaModel"),
         ("reformer", "ReformerModel"),
         ("regnet", "RegNetModel"),
         ("rembert", "RemBertModel"),
@@ -288,7 +285,6 @@ MODEL_FOR_PRETRAINING_MAPPING_NAMES = OrderedDict(
         ("gptsan-japanese", "GPTSanJapaneseForConditionalGeneration"),
         ("ibert", "IBertForMaskedLM"),
         ("idefics", "IdeficsForVisionText2Text"),
-        ("idefics2", "Idefics2ForConditionalGeneration"),
         ("layoutlm", "LayoutLMForMaskedLM"),
         ("llava", "LlavaForConditionalGeneration"),
         ("llava_next", "LlavaNextForConditionalGeneration"),
@@ -473,7 +469,6 @@ MODEL_FOR_CAUSAL_LM_MAPPING_NAMES = OrderedDict(
         ("qdqbert", "QDQBertLMHeadModel"),
         ("qwen2", "Qwen2ForCausalLM"),
         ("qwen2_moe", "Qwen2MoeForCausalLM"),
-        ("recurrent_gemma", "RecurrentGemmaForCausalLM"),
         ("reformer", "ReformerModelWithLMHead"),
         ("rembert", "RemBertForCausalLM"),
         ("roberta", "RobertaForCausalLM"),
@@ -680,7 +675,6 @@ MODEL_FOR_VISION_2_SEQ_MAPPING_NAMES = OrderedDict(
         ("blip", "BlipForConditionalGeneration"),
         ("blip-2", "Blip2ForConditionalGeneration"),
         ("git", "GitForCausalLM"),
-        ("idefics2", "Idefics2ForConditionalGeneration"),
         ("instructblip", "InstructBlipForConditionalGeneration"),
         ("kosmos-2", "Kosmos2ForConditionalGeneration"),
         ("llava", "LlavaForConditionalGeneration"),
@@ -757,7 +751,6 @@ MODEL_FOR_OBJECT_DETECTION_MAPPING_NAMES = OrderedDict(
 MODEL_FOR_ZERO_SHOT_OBJECT_DETECTION_MAPPING_NAMES = OrderedDict(
     [
         # Model for Zero Shot Object Detection mapping
-        ("grounding-dino", "GroundingDinoForObjectDetection"),
         ("owlv2", "Owlv2ForObjectDetection"),
         ("owlvit", "OwlViTForObjectDetection"),
     ]

src/transformers/models/auto/processing_auto.py

@@ -61,7 +61,6 @@ PROCESSOR_MAPPING_NAMES = OrderedDict(
         ("groupvit", "CLIPProcessor"),
         ("hubert", "Wav2Vec2Processor"),
         ("idefics", "IdeficsProcessor"),
-        ("idefics2", "Idefics2Processor"),
         ("instructblip", "InstructBlipProcessor"),
         ("kosmos-2", "Kosmos2Processor"),
         ("layoutlmv2", "LayoutLMv2Processor"),

src/transformers/models/auto/tokenization_auto.py

@@ -195,13 +195,11 @@ else:
             ("gpt_neox_japanese", ("GPTNeoXJapaneseTokenizer", None)),
             ("gptj", ("GPT2Tokenizer", "GPT2TokenizerFast" if is_tokenizers_available() else None)),
             ("gptsan-japanese", ("GPTSanJapaneseTokenizer", None)),
-            ("grounding-dino", ("BertTokenizer", "BertTokenizerFast" if is_tokenizers_available() else None)),
             ("groupvit", ("CLIPTokenizer", "CLIPTokenizerFast" if is_tokenizers_available() else None)),
             ("herbert", ("HerbertTokenizer", "HerbertTokenizerFast" if is_tokenizers_available() else None)),
             ("hubert", ("Wav2Vec2CTCTokenizer", None)),
             ("ibert", ("RobertaTokenizer", "RobertaTokenizerFast" if is_tokenizers_available() else None)),
             ("idefics", (None, "LlamaTokenizerFast" if is_tokenizers_available() else None)),
-            ("idefics2", ("LlamaTokenizer", "LlamaTokenizerFast" if is_tokenizers_available() else None)),
             ("instructblip", ("GPT2Tokenizer", "GPT2TokenizerFast" if is_tokenizers_available() else None)),
             ("jukebox", ("JukeboxTokenizer", None)),
             (
@@ -365,13 +363,6 @@ else:
             ),
             ("rag", ("RagTokenizer", None)),
             ("realm", ("RealmTokenizer", "RealmTokenizerFast" if is_tokenizers_available() else None)),
-            (
-                "recurrent_gemma",
-                (
-                    "GemmaTokenizer" if is_sentencepiece_available() else None,
-                    "GemmaTokenizerFast" if is_tokenizers_available() else None,
-                ),
-            ),
             (
                 "reformer",
                 (