tmp

Signed-off-by: Yikun Jiang <[email protected]>

Author: Yikun

.github/workflows/accuracy_test.yaml

@@ -46,7 +46,6 @@ on:
           - all
           - Qwen/Qwen2.5-7B-Instruct
           - Qwen/Qwen2.5-VL-7B-Instruct
-          - Qwen/Qwen3-8B-Base
         default: 'all'
 
 # Bash shells do not use ~/.profile or ~/.bashrc so these shells need to be explicitly
@@ -59,7 +58,13 @@ defaults:
 jobs:
   model_tests:
     # test will be triggered when tag '*-accuracy-test' & 'ready-for-test' or schedule job
-    if: ${{ contains(github.event.pull_request.labels.*.name, 'dense-accuracy-test') && contains(github.event.pull_request.labels.*.name, 'ready-for-test') || github.event_name == 'workflow_dispatch' }}
+    if:  >-
+      ${{
+      (contains(github.event.pull_request.labels.*.name, 'vl-accuracy-test') ||
+      contains(github.event.pull_request.labels.*.name, 'dense-accuracy-test')) &&
+      contains(github.event.pull_request.labels.*.name, 'ready-for-test') ||
+      github.event_name == 'workflow_dispatch'
+      }}
     name: Model Test - ${{ matrix.model_name }}
     runs-on: 'linux-arm64-npu-2'
     strategy:
@@ -72,9 +77,9 @@ jobs:
         #   - dense-accuracy-test: Qwen/Qwen2.5-7B-Instruct
         #   - vl-accuracy-test: Qwen/Qwen2.5-VL-7B-Instruct
         include: ${{ fromJSON(
-          github.event_name == 'workflow_dispatch' ||
-          (github.event.inputs.models == 'all' && '[{"model_name":"Qwen/Qwen2.5-7B-Instruct"},{"model_name":"Qwen/Qwen2.5-VL-7B-Instruct"},{"model_name":"Qwen/Qwen3-8B-Base"}]') ||
-          (github.event.inputs.models != '' && '[{"model_name":"${{ github.event.inputs.models }}"}]') ||
+          (github.event.inputs.models == 'all' && '[{"model_name":"Qwen/Qwen2.5-7B-Instruct"},{"model_name":"meta-llama/Llama-3.1-8B-Instruct"},{"model_name":"Qwen/Qwen2.5-VL-7B-Instruct"}, {"model_name":"Qwen/Qwen3-8B-Base"}]') ||
+          (github.event.inputs.models == 'Qwen/Qwen2.5-7B-Instruct' && '[{"model_name":"Qwen/Qwen2.5-7B-Instruct"}]') ||
+          (github.event.inputs.models == 'Qwen/Qwen2.5-VL-7B-Instruct' && '[{"model_name":"Qwen/Qwen2.5-VL-7B-Instruct"}]') ||
           contains(github.event.pull_request.labels.*.name, 'dense-accuracy-test') && '[{"model_name":"Qwen/Qwen2.5-7B-Instruct"}]' ||
           contains(github.event.pull_request.labels.*.name, 'vl-accuracy-test') && '[{"model_name":"Qwen/Qwen2.5-VL-7B-Instruct"}]'
          ) }}
@@ -86,6 +91,7 @@ jobs:
         HF_ENDPOINT: https://hf-mirror.com
         HF_TOKEN: ${{ secrets.HF_TOKEN }}
         DATASET_SOURCE: ModelScope
+        VLLM_USE_MODELSCOPE: True
 
     steps:
       - name: Checkout repository
@@ -114,26 +120,27 @@ jobs:
         with:
           repository: vllm-project/vllm
           path: ./vllm-empty
-          ref: ${{ github.event.inputs.vllm-version }}
+          ref: ${{ github.event.inputs.vllm-version  || 'v0.9.0' }}
 
-      - name: Install vllm-project/vllm from source
-        working-directory: ./vllm-empty
-        run: VLLM_TARGET_DEVICE=empty pip install -e .
-      
+      # - name: Install vllm-project/vllm from source
+      #   working-directory: ./vllm-empty
+      #   run: VLLM_TARGET_DEVICE=empty pip install -e .
 
       - name: Checkout vllm-project/vllm-ascend repo
         uses: actions/checkout@v4
         with:
           repository: vllm-project/vllm-ascend
           path: ./vllm-ascend
+          # 1. If version specified (work_dispatch), do specified branch test
+          # 2. If no version (labeled PR), do PR accuracy test
           ref: ${{ github.event.inputs.vllm-ascend-version }}
           fetch-depth: 0
 
-      - name: Install vllm-project/vllm-ascend
-        working-directory: ./vllm-ascend
-        run: |
-          pip install -r requirements-dev.txt
-          pip install -e .
+      # - name: Install vllm-project/vllm-ascend
+      #   working-directory: ./vllm-ascend
+      #   run: |
+      #     pip install -r requirements-dev.txt
+      #     pip install -e .
 
       - name: Install lm-eval, ray, and datasets
         run: |
@@ -175,16 +182,19 @@ jobs:
           PYTORCH_NPU_ALLOC_CONF: max_split_size_mb:256
         run: |
           model_base_name=$(basename ${{ matrix.model_name }})
+          echo "model_base_name=$model_base_name"
           echo "model_base_name=$model_base_name" >> $GITHUB_OUTPUT
           mkdir -p ./accuracy/V0
-          python ./scripts/run_accuracy.py \
-            --model "${{ matrix.model_name }}" \
-            --output "./accuracy/V0/${model_base_name}.md" \
-            --vllm_ascend_version "${{ github.event.inputs.vllm-ascend-version }}" \
-            --cann_version "${{ env.CANN_VERSION }}" \
-            --torch_npu_version "${{ env.TORCH_NPU_VERSION }}" \
-            --torch_version "${{ env.TORCH_VERSION }}" \
-            --vllm_version "${{ env.VLLM_VERSION }}"
+          echo "test111" > ./accuracy/V0/${model_base_name}.md
+          # python ./scripts/run_accuracy.py \
+          #   --model "${{ matrix.model_name }}" \
+          #   --output "./accuracy/V0/${model_base_name}.md" \
+          #   --vllm_ascend_version "${{ github.event.inputs.vllm-ascend-version || 'current' }}" \
+          #   --cann_version "${{ env.CANN_VERSION }}" \
+          #   --torch_npu_version "${{ env.TORCH_NPU_VERSION }}" \
+          #   --torch_version "${{ env.TORCH_VERSION }}" \
+          #   --vllm_version "${{ env.VLLM_VERSION }}"
+          cat ./accuracy/V0/${model_base_name}.md
 
       - name: Upload Report for V0
         uses: actions/upload-artifact@v4