Merge branch 'develop' into merge_dsv3_pr

Author: hushenwei2000

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

@@ -12,7 +12,7 @@ on:
   #   types: [opened, synchronize, reopened]
   #   branches: [develop]
   schedule:
-    - cron: "5 0 * * 0"     # every Sunday at 00:05
+    - cron: "0 16 * * 5"   # UTC 16:00 = Beijing Time, Saturday at 00:00
   workflow_dispatch:        # allow to manually trigger the workflow
 
 concurrency:
@@ -25,14 +25,13 @@ env:
   BRANCH: ${{ github.event.pull_request.base.ref }}
   GITHUB_EVENT_NAME: ${{ github.event_name }}
   TASK: PaddleFormers-CI-${{ github.event.pull_request.number }}-build-image
-  image_base: "iregistry.baidu-int.com/paddlecloud/base-images:paddlecloud-ubuntu18.04-gcc8.2-cuda11.8-cudnn8.6-nccl2.15.5"
-  no_proxy: "localhost,bj.bcebos.com,su.bcebos.com,bcebos.com,apiin.im.baidu.com,gitee.com,aliyun.com,.baidu.com,.tuna.tsinghua.edu.cn"
+  image_base: "ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddle:cuda126-dev"
 
 jobs:
   build-ci-docker-images:
     name: build-ci-docker-images
     if: github.ref == 'refs/heads/develop'
-    runs-on: [self-hosted, ernie-8gpu]
+    runs-on: [self-hosted, ernie-8gpu-1]
     outputs:
       run_build: ${{ steps.set_flags.outputs.run_build }}
       run_downstream: ${{ steps.set_flags.outputs.run_downstream }}
@@ -45,7 +44,7 @@ jobs:
         run: |
           container_name=${TASK}-$(date +%Y%m%d-%H%M%S)
           echo "container_name=${container_name}" >> ${{ github.env }}
-          docker_image="${image_base}-latest"
+          docker_image="${image_base}"
           docker run -d -t --name ${container_name} --net=host -v /dev/shm:/dev/shm --shm-size=32G \
             -v $work_dir/../../..:$work_dir/../../.. \
             -v $work_dir:/workspace \
@@ -55,7 +54,7 @@ jobs:
             -e COMMIT_ID \
             -e work_dir \
             -e no_proxy \
-            -w /workspace --runtime=nvidia ${docker_image}
+            -w /workspace --runtime=nvidia --privileged ${docker_image}
 
       - name: Download Code
         env:
@@ -139,10 +138,9 @@ jobs:
       - name: Write Dockerfile Inline
         run: |
           cat <<EOF > Dockerfile
-          FROM iregistry.baidu-int.com/paddlecloud/base-images:paddlecloud-ubuntu18.04-gcc8.2-cuda11.8-cudnn8.6-nccl2.15.5-libsndfile
-          RUN unlink /usr/bin/python3 && ln -sf python3.10 /usr/bin/python3
-          RUN rm -rf /etc/pip.conf
+          FROM ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddle:cuda126-dev
           RUN python -m pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple
+          RUN python -m pip config set global.trusted-host pypi.tuna.tsinghua.edu.cn
           RUN python -m pip install bce-python-sdk==0.8.74
           COPY PaddleFormers/ /tmp/paddleformers/
           WORKDIR /tmp/paddleformers
@@ -153,7 +151,6 @@ jobs:
           RUN python -m pip install dist/*.whl
           WORKDIR /
           RUN rm -rf /tmp/paddleformers
-          CMD ["python3"]
           EOF
       
       - name: Build Docker Image
@@ -180,7 +177,7 @@ jobs:
   clean-ci-image:
     name: clean-ci-image
     needs: [build-ci-docker-images, test-ci-docker-images]
-    runs-on: [self-hosted, ernie-8gpu]
+    runs-on: [self-hosted, ernie-8gpu-1]
     steps:
       - name: Remove CI Image
         env:
@@ -203,5 +200,5 @@ jobs:
                 echo "$CONTAINERS"
             fi
           fi
-          # 安全删除悬空镜像（无悬空镜像时跳过）
+          # Safely remove dangling images (skip if no dangling images exist)
           docker images -f "dangling=true" -q | xargs -r docker rmi -f

.github/workflows/ce-build.yml

@@ -4,15 +4,10 @@ on:
   push:
     branches:
       - develop
-      
-concurrency:
-  group: build-${{ github.workflow }}
-  cancel-in-progress: true
 
 env:
-  COMMIT_ID: ${{ github.event.pull_request.head.sha }}
-  TASK: PaddleFormers-CE-${{ github.event.pull_request.number }}-build
-  BRANCH: ${{ github.event.pull_request.base.ref }}
+  COMMIT_ID: ${{ github.sha }}
+  TASK: PaddleFormers-CE-${{ github.sha }}-build
   CE_name: build-ce
   no_proxy: "localhost,bj.bcebos.com,su.bcebos.com,bcebos.com,apiin.im.baidu.com,gitee.com,aliyun.com,.baidu.com,.tuna.tsinghua.edu.cn"
 
@@ -23,18 +18,18 @@ defaults:
 jobs:
   build-ce:
     name: build-ce
-    runs-on: [self-hosted, ernie-8gpu]
+    runs-on: [self-hosted, paddleformers]
     steps:
       - name: Determine Image Name
         run: |
-          echo "IMAGE_NAME=iregistry.baidu-int.com/paddlecloud/base-images:paddlecloud-ubuntu18.04-gcc8.2-cuda11.8-cudnn8.6-nccl2.15.5-latest" >> "$GITHUB_ENV"
+          echo "IMAGE_NAME=ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddle:cuda126-dev-latest" >> "$GITHUB_ENV"
 
       - name: Run Container
         env:
           work_dir: ${{ github.workspace }}
           FLAGS_dynamic_static_unified_comm: "True"
           python_version: "3.10"
-          paddle: https://paddle-qa.bj.bcebos.com/paddle-pipeline/Develop-GpuSome-LinuxCentos-Gcc82-Cuda118-Cudnn86-Trt85-Py310-CENN-Compile/latest/paddlepaddle_gpu-0.0.0-cp310-cp310-linux_x86_64.whl
+          paddle: https://paddle-qa.bj.bcebos.com/paddle-pipeline/Develop-GpuAll-LinuxCentos-Gcc11-Cuda126-Cudnn95-Trt105-Py310-Compile/latest/paddlepaddle_gpu-0.0.0-cp310-cp310-linux_x86_64.whl
         run: |
           container_name=${TASK}-$(date +%Y%m%d-%H%M%S)
           echo "container_name=${container_name}" >> "$GITHUB_ENV"
@@ -43,7 +38,6 @@ jobs:
             -v $work_dir:/workspace \
             -v /home/.cache/pip:/home/.cache/pip \
             -e BRANCH \
-            -e AGILE_COMPILE_BRANCH \
             -e COMMIT_ID \
             -e work_dir \
             -e ce_scripts \
@@ -52,7 +46,7 @@ jobs:
             -e paddle \
             -e FLAGS_dynamic_static_unified_comm \
             -e python_version \
-            -w /workspace --runtime=nvidia $IMAGE_NAME
+            -w /workspace --runtime=nvidia --privileged $IMAGE_NAME
             
       - name: Download Code
         run: |
@@ -68,40 +62,37 @@ jobs:
           git config --global user.email "[email protected]"
           git pull
           git submodule update --init --recursive --force
+          git checkout -b $COMMIT_ID $COMMIT_ID
           git log --pretty=oneline -10
           '
 
       - name: Test
         run: |
           docker exec -t $container_name /bin/bash -c '
           ldconfig
-          unlink /usr/bin/python3
-          ln -sf $(which python${python_version}) /usr/bin/python3
           pip config set global.cache-dir "/home/.cache/pip"
           set -e
           cd /workspace/PaddleFormers && git config --global --add safe.directory $PWD
           bash scripts/regression/build.sh
           '
           
-      - name: Upload Logs
+      - name: Upload Products
         if: always()
         env:
           home_path: ${{ github.workspace }}/../../..
           bos_file: ${{ github.workspace }}/../../../bos/BosClient.py
-          allure_file: ${{ github.workspace }}/../../../allure-2.19.0/bin/allure
         run: |
           docker exec -t $container_name /bin/bash -c '
           if [ ! -f "${{ env.bos_file }}" ]; then
             wget -q --no-proxy -O ${{ env.home_path }}/bos_new.tar.gz https://xly-devops.bj.bcebos.com/home/bos_new.tar.gz --no-check-certificate
             mkdir ${{ env.home_path }}/bos
             tar xf ${{ env.home_path }}/bos_new.tar.gz -C ${{ env.home_path }}/bos
           fi
-          bos_prefix="schedule/$(date +%Y%m%d)"
-          cd /workspace/PaddleFormers/build_logs
-          for FILE in /workspace/PaddleFormers/build_logs/*; do
+          cd /workspace/PaddleFormers/upload
+          for FILE in /workspace/PaddleFormers/upload/*; do
             file=$(basename "$FILE")
-            python ${{ env.bos_file }} $file paddle-github-action/PR/PaddleFormers/build/${bos_prefix}/logs
-            echo "$file: https://paddle-github-action.bj.bcebos.com/PR/PaddleFormers/build/${bos_prefix}/logs/$file"
+            python ${{ env.bos_file }} $file paddleformers/wheels
+            echo "$file: https://paddleformers.bj.bcebos.com/wheels/$file"
           done
           '
 

.github/workflows/ce-deadlink.yml

@@ -2,7 +2,7 @@ name: Deadlink CE
 
 on:
   schedule:
-    - cron: "0 8 * * 6"     # every Saturday at 6:00 AM
+    - cron: "0 8 * * 6"     # every Saturday at 16:00
   workflow_dispatch:        # allow to manually trigger the workflow
 
 concurrency:
@@ -24,18 +24,18 @@ jobs:
   deadlink-ce:
     name: deadlink-ce
     if: github.ref == 'refs/heads/develop'
-    runs-on: [self-hosted, ernie-8gpu]
+    runs-on: [self-hosted, paddleformers]
     steps:
       - name: Determine Image Name
         run: |
-          echo "IMAGE_NAME=iregistry.baidu-int.com/paddlecloud/base-images:paddlecloud-ubuntu18.04-gcc8.2-cuda11.8-cudnn8.6-nccl2.15.5-latest" >> "$GITHUB_ENV"
+          echo "IMAGE_NAME=ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddle:cuda126-dev-latest" >> "$GITHUB_ENV"
 
       - name: Run Container
         env:
           work_dir: ${{ github.workspace }}
           FLAGS_dynamic_static_unified_comm: "True"
           python_version: "3.10"
-          paddle_whl: https://paddle-qa.bj.bcebos.com/paddle-pipeline/Develop-GpuSome-LinuxCentos-Gcc82-Cuda118-Cudnn86-Trt85-Py310-CENN-Compile/latest/paddlepaddle_gpu-0.0.0-cp310-cp310-linux_x86_64.whl
+          paddle_whl: https://paddle-qa.bj.bcebos.com/paddle-pipeline/Develop-GpuAll-LinuxCentos-Gcc11-Cuda126-Cudnn95-Trt105-Py310-Compile/latest/paddlepaddle_gpu-0.0.0-cp310-cp310-linux_x86_64.whl
         run: |
           container_name=${TASK}-$(date +%Y%m%d-%H%M%S)
           echo "container_name=${container_name}" >> "$GITHUB_ENV"
@@ -44,7 +44,6 @@ jobs:
             -v $work_dir:/workspace \
             -v /home/.cache/pip:/home/.cache/pip \
             -e BRANCH \
-            -e AGILE_COMPILE_BRANCH \
             -e COMMIT_ID \
             -e work_dir \
             -e ce_scripts \
@@ -53,7 +52,7 @@ jobs:
             -e paddle_whl \
             -e FLAGS_dynamic_static_unified_comm \
             -e python_version \
-            -w /workspace --runtime=nvidia $IMAGE_NAME
+            -w /workspace --runtime=nvidia --privileged $IMAGE_NAME
             
       - name: Download Code
         run: |
@@ -81,20 +80,18 @@ jobs:
         run: |
           docker exec -t $container_name /bin/bash -c '
           ldconfig
-          unlink /usr/bin/python3
-          ln -sf $(which python${python_version}) /usr/bin/python3
           pip config set global.cache-dir "/home/.cache/pip"
           set -e
           python -m pip install beautifulsoup4 openpyxl
-          bash run.sh PaddleFormers develop [email protected] PaddleFormers死链检测汇总报告
+          source $work_dir/../../../proxy
+          bash run.sh PaddleFormers develop [email protected] "PaddleFormers Broken Link Check Summary Report"
           '
           
       - name: Upload Logs
         if: always()
         env:
           home_path: ${{ github.workspace }}/../../..
           bos_file: ${{ github.workspace }}/../../../bos/BosClient.py
-          allure_file: ${{ github.workspace }}/../../../allure-2.19.0/bin/allure
         run: |
           docker exec -t $container_name /bin/bash -c '
           if [ ! -f "${{ env.bos_file }}" ]; then
@@ -103,8 +100,8 @@ jobs:
             tar xf ${{ env.home_path }}/bos_new.tar.gz -C ${{ env.home_path }}/bos
           fi
           bos_prefix="schedule/$(date +%Y%m%d)"
-          cd /workspace/PaddleFormers/deadlink_logs
-          for FILE in /workspace/PaddleFormers/deadlink_logs/*; do
+          cd /workspace/result
+          for FILE in /workspace/result/*; do
             file=$(basename "$FILE")
             python ${{ env.bos_file }} $file paddle-github-action/PR/PaddleFormers/deadlink/${bos_prefix}/logs
             echo "$file: https://paddle-github-action.bj.bcebos.com/PR/PaddleFormers/deadlink/${bos_prefix}/logs/$file"

.github/workflows/ce-unittest-gpu.yml

@@ -2,8 +2,18 @@ name: Unittest GPU CE
 
 on:
   schedule:
-    - cron: "0 6 * * 6"
+    - cron: "0 2 * * 6"     # every Saturday at 10:00
   workflow_dispatch:        # allow to manually trigger the workflow
+    inputs:
+      paddle_whl:
+        description: "paddle_whl"
+        required: false
+        default: "https://paddle-qa.bj.bcebos.com/paddle-pipeline/Develop-GpuAll-LinuxCentos-Gcc11-Cuda126-Cudnn95-Trt105-Py310-Compile/latest/paddlepaddle_gpu-0.0.0-cp310-cp310-linux_x86_64.whl"
+        type: string
+      repo_branch:
+        description: "repo_branch"
+        required: false
+        type: string
 
 concurrency:
   group: unittest-${{ github.workflow }}
@@ -25,19 +35,19 @@ defaults:
 jobs:
   unittest-gpu-ce:
     name: unittest-gpu-ce
-    if: github.ref == 'refs/heads/develop'
     runs-on: [self-hosted, ernie-8gpu]
     steps:
       - name: Determine Image Name
         run: |
-          echo "IMAGE_NAME=iregistry.baidu-int.com/paddlecloud/base-images:paddlecloud-ubuntu18.04-gcc8.2-cuda11.8-cudnn8.6-nccl2.15.5-latest" >> "$GITHUB_ENV"
+          echo "IMAGE_NAME=ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddle:cuda126-dev-latest" >> "$GITHUB_ENV"
 
       - name: Run Container
         env:
           work_dir: ${{ github.workspace }}
           FLAGS_dynamic_static_unified_comm: "True"
           python_version: "3.10"
-          paddle_whl: https://paddle-qa.bj.bcebos.com/paddle-pipeline/Develop-GpuSome-LinuxCentos-Gcc82-Cuda118-Cudnn86-Trt85-Py310-CENN-Compile/latest/paddlepaddle_gpu-0.0.0-cp310-cp310-linux_x86_64.whl
+          paddle_whl: ${{ github.event.inputs.paddle_whl || 'https://paddle-qa.bj.bcebos.com/paddle-pipeline/Develop-GpuAll-LinuxCentos-Gcc11-Cuda126-Cudnn95-Trt105-Py310-Compile/latest/paddlepaddle_gpu-0.0.0-cp310-cp310-linux_x86_64.whl' }}
+          repo_branch: ${{ github.event.inputs.repo_branch || '' }}
         run: |
           container_name=${TASK}-$(date +%Y%m%d-%H%M%S)
           echo "container_name=${container_name}" >> "$GITHUB_ENV"
@@ -53,9 +63,10 @@ jobs:
             -e no_proxy \
             -e CE_name \
             -e paddle_whl \
+            -e repo_branch \
             -e FLAGS_dynamic_static_unified_comm \
             -e python_version \
-            -w /workspace --runtime=nvidia $IMAGE_NAME
+            -w /workspace --runtime=nvidia --privileged $IMAGE_NAME
             
       - name: Download Code
         run: |
@@ -69,7 +80,15 @@ jobs:
           cd PaddleFormers
           git config --global user.name "PaddleCE"
           git config --global user.email "[email protected]"
-          git pull
+          if [ -n "$repo_branch" ]; then
+            echo "Switching to branch: $repo_branch"
+            git fetch origin $repo_branch
+            git checkout $repo_branch || git checkout -b $repo_branch origin/$repo_branch
+            git pull
+          else
+            echo "No repo_branch provided, just pulling latest changes"
+            git pull
+          fi
           git submodule update --init --recursive --force
           git log --pretty=oneline -10
           '
@@ -78,12 +97,10 @@ jobs:
         run: |
           docker exec -t $container_name /bin/bash -c '
           ldconfig
-          unlink /usr/bin/python3
-          ln -sf $(which python${python_version}) /usr/bin/python3
           pip config set global.cache-dir "/home/.cache/pip"
           set -e
           cd /workspace/PaddleFormers && git config --global --add safe.directory $PWD
-          timeout 2h bash scripts/unit_test/ci_unit.sh ${paddle} true
+          timeout 2h bash scripts/unit_test/ci_unit.sh ${paddle_whl} true
           '
           
       - name: Upload Allure-reports & Logs