Update some workflow files (#41892)

* update

* update

* final check

* final check

* final clean

---------

Co-authored-by: ydshieh <[email protected]>

Author: ydshieh

.github/workflows/benchmark.yml

@@ -28,7 +28,7 @@ jobs:
       (github.event_name == 'pull_request' && contains( github.event.pull_request.labels.*.name, 'run-benchmark') )||
       (github.event_name == 'push' && github.ref == 'refs/heads/main')
     container:
-      image: huggingface/transformers-pytorch-gpu
+      image: huggingface/transformers-all-latest-gpu
       options: --gpus all --privileged --ipc host
     steps:
       - name: Get repo

.github/workflows/benchmark_v2_a10_caller.yml

@@ -9,7 +9,7 @@ jobs:
     uses: ./.github/workflows/benchmark_v2.yml
     with:
       runner: aws-g5-4xlarge-cache-use1-public-80
-      container_image: huggingface/transformers-pytorch-gpu
+      container_image: huggingface/transformers-all-latest-gpu
       container_options: --gpus all --privileged --ipc host --shm-size "16gb"
       commit_sha: ${{ github.sha }}
       run_id: ${{ github.run_id }}

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

@@ -45,33 +45,20 @@ jobs:
             REF=main
           push: true
           tags: huggingface/transformers-all-latest-gpu${{ inputs.image_postfix }}
-      # Push CI images still need to be re-built daily
-      -
-        name: Build and push (for Push CI) in a daily basis
-        # This condition allows `schedule` events, or `push` events that trigger this workflow NOT via `workflow_call`.
-        # The later case is useful for manual image building for debugging purpose. Use another tag in this case!
-        if: inputs.image_postfix != '-push-ci'
-        uses: docker/build-push-action@v5
-        with:
-          context: ./docker/transformers-all-latest-gpu
-          build-args: |
-            REF=main
-          push: true
-          tags: huggingface/transformers-all-latest-gpu-push-ci
 
       - name: Post to Slack
         if: always()
         uses: huggingface/hf-workflows/.github/actions/post-slack@main
         with:
           slack_channel: ${{ secrets.CI_SLACK_CHANNEL_DOCKER }}
-          title: 🤗 Results of the transformers-all-latest-gpu-push-ci docker build
+          title: 🤗 Results of the transformers-all-latest-gpu docker build
           status: ${{ job.status }}
           slack_token: ${{ secrets.SLACK_CIFEEDBACK_BOT_TOKEN }}
 
-  latest-torch-deepspeed-docker:
-    name: "Latest PyTorch + DeepSpeed"
+  flash-attn-ci-image:
+    name: "PyTorch with Flash Attn [dev]"
     runs-on:
-      group: aws-g4dn-2xlarge-cache
+      group: aws-general-8-plus
     steps:
       -
         name: Set up Docker Buildx
@@ -89,26 +76,28 @@ jobs:
         name: Build and push
         uses: docker/build-push-action@v5
         with:
-          context: ./docker/transformers-pytorch-deepspeed-latest-gpu
+          context: ./docker/transformers-all-latest-gpu
           build-args: |
-            REF=main
+            REF=update_dockerfile
+            PYTORCH=2.8.0
+            TORCHCODEC=0.7.0
+            FLASH_ATTN=yes
           push: true
-          tags: huggingface/transformers-pytorch-deepspeed-latest-gpu${{ inputs.image_postfix }}
+          tags: huggingface/transformers-all-latest-gpu${{ inputs.image_postfix }}:flash-attn
 
       - name: Post to Slack
         if: always()
         uses: huggingface/hf-workflows/.github/actions/post-slack@main
         with:
-          slack_channel: ${{ secrets.CI_SLACK_CHANNEL_DOCKER}}
-          title: 🤗 Results of the transformers-pytorch-deepspeed-latest-gpu docker build
+          slack_channel: ${{ secrets.CI_SLACK_CHANNEL_DOCKER }}
+          title: 🤗 Results of the transformers-all-latest-gpu docker build
           status: ${{ job.status }}
           slack_token: ${{ secrets.SLACK_CIFEEDBACK_BOT_TOKEN }}
 
-  # Can't build 2 images in a single job `latest-torch-deepspeed-docker` (for `nvcr.io/nvidia`)
-  latest-torch-deepspeed-docker-for-push-ci-daily-build:
-    name: "Latest PyTorch + DeepSpeed (Push CI - Daily Build)"
+  latest-torch-deepspeed-docker:
+    name: "Latest PyTorch + DeepSpeed"
     runs-on:
-      group: aws-general-8-plus
+      group: aws-g4dn-2xlarge-cache
     steps:
       -
         name: Set up Docker Buildx
@@ -122,33 +111,27 @@ jobs:
         with:
           username: ${{ secrets.DOCKERHUB_USERNAME }}
           password: ${{ secrets.DOCKERHUB_PASSWORD }}
-      # Push CI images still need to be re-built daily
       -
-        name: Build and push (for Push CI) in a daily basis
-        # This condition allows `schedule` events, or `push` events that trigger this workflow NOT via `workflow_call`.
-        # The later case is useful for manual image building for debugging purpose. Use another tag in this case!
-        if: inputs.image_postfix != '-push-ci'
+        name: Build and push
         uses: docker/build-push-action@v5
         with:
           context: ./docker/transformers-pytorch-deepspeed-latest-gpu
           build-args: |
             REF=main
           push: true
-          tags: huggingface/transformers-pytorch-deepspeed-latest-gpu-push-ci
+          tags: huggingface/transformers-pytorch-deepspeed-latest-gpu${{ inputs.image_postfix }}
 
       - name: Post to Slack
         if: always()
         uses: huggingface/hf-workflows/.github/actions/post-slack@main
         with:
-          slack_channel: ${{ secrets.CI_SLACK_CHANNEL_DOCKER }}
-          title: 🤗 Results of the transformers-pytorch-deepspeed-latest-gpu-push-ci docker build
+          slack_channel: ${{ secrets.CI_SLACK_CHANNEL_DOCKER}}
+          title: 🤗 Results of the transformers-pytorch-deepspeed-latest-gpu docker build
           status: ${{ job.status }}
           slack_token: ${{ secrets.SLACK_CIFEEDBACK_BOT_TOKEN }}
 
   doc-builder:
     name: "Doc builder"
-    # Push CI doesn't need this image
-    if: inputs.image_postfix != '-push-ci'
     runs-on:
       group: aws-general-8-plus
     steps:
@@ -181,44 +164,6 @@ jobs:
           status: ${{ job.status }}
           slack_token: ${{ secrets.SLACK_CIFEEDBACK_BOT_TOKEN }}
 
-  latest-pytorch:
-    name: "Latest PyTorch [dev]"
-    # Push CI doesn't need this image
-    if: inputs.image_postfix != '-push-ci'
-    runs-on:
-      group: aws-general-8-plus
-    steps:
-      -
-        name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v3
-      -
-        name: Check out code
-        uses: actions/checkout@v4
-      -
-        name: Login to DockerHub
-        uses: docker/login-action@v3
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_PASSWORD }}
-      -
-        name: Build and push
-        uses: docker/build-push-action@v5
-        with:
-          context: ./docker/transformers-pytorch-gpu
-          build-args: |
-            REF=main
-          push: true
-          tags: huggingface/transformers-pytorch-gpu
-
-      - name: Post to Slack
-        if: always()
-        uses: huggingface/hf-workflows/.github/actions/post-slack@main
-        with:
-          slack_channel: ${{ secrets.CI_SLACK_CHANNEL_DOCKER }}
-          title: 🤗 Results of the huggingface/transformers-pytorch-gpudocker build
-          status: ${{ job.status }}
-          slack_token: ${{ secrets.SLACK_CIFEEDBACK_BOT_TOKEN }}
-
   latest-pytorch-amd:
     name: "Latest PyTorch (AMD) [dev]"
     runs-on:
@@ -245,26 +190,13 @@ jobs:
             REF=main
           push: true
           tags: huggingface/transformers-pytorch-amd-gpu${{ inputs.image_postfix }}
-      # Push CI images still need to be re-built daily
-      -
-        name: Build and push (for Push CI) in a daily basis
-        # This condition allows `schedule` events, or `push` events that trigger this workflow NOT via `workflow_call`.
-        # The later case is useful for manual image building for debugging purpose. Use another tag in this case!
-        if: inputs.image_postfix != '-push-ci'
-        uses: docker/build-push-action@v5
-        with:
-          context: ./docker/transformers-pytorch-amd-gpu
-          build-args: |
-            REF=main
-          push: true
-          tags: huggingface/transformers-pytorch-amd-gpu-push-ci
 
       - name: Post to Slack
         if: always()
         uses: huggingface/hf-workflows/.github/actions/post-slack@main
         with:
           slack_channel: ${{ secrets.CI_SLACK_CHANNEL_DOCKER }}
-          title: 🤗 Results of the huggingface/transformers-pytorch-amd-gpu-push-ci build
+          title: 🤗 Results of the huggingface/transformers-pytorch-amd-gpu build
           status: ${{ job.status }}
           slack_token: ${{ secrets.SLACK_CIFEEDBACK_BOT_TOKEN }}
 
@@ -294,19 +226,6 @@ jobs:
             REF=main
           push: true
           tags: huggingface/transformers-pytorch-deepspeed-amd-gpu${{ inputs.image_postfix }}
-      # Push CI images still need to be re-built daily
-      -
-        name: Build and push (for Push CI) in a daily basis
-        # This condition allows `schedule` events, or `push` events that trigger this workflow NOT via `workflow_call`.
-        # The later case is useful for manual image building for debugging purpose. Use another tag in this case!
-        if: inputs.image_postfix != '-push-ci'
-        uses: docker/build-push-action@v5
-        with:
-          context: ./docker/transformers-pytorch-deepspeed-amd-gpu
-          build-args: |
-            REF=main
-          push: true
-          tags: huggingface/transformers-pytorch-deepspeed-amd-gpu-push-ci
 
       - name: Post to Slack
         if: always()
@@ -319,8 +238,6 @@ jobs:
 
   latest-quantization-torch-docker:
     name: "Latest Pytorch + Quantization [dev]"
-     # Push CI doesn't need this image
-    if: inputs.image_postfix != '-push-ci'
     runs-on:
       group: aws-general-8-plus
     steps:

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

@@ -149,7 +149,7 @@ jobs:
     with:
       job: run_models_gpu
       slack_report_channel: "#transformers-ci-push"
-      docker: huggingface/transformers-all-latest-gpu
+      docker: huggingface/transformers-all-latest-gpu:flash-attn
       ci_event: push
       report_repo_id: hf-internal-testing/transformers_ci_push
       commit_sha: ${{ github.sha }}