Bump torch-npu version to 2.5.1.post1.dev20250528

Signed-off-by: Icey <1790571317@qq.com>
Signed-off-by: MengqingCao <cmq0113@163.com>

Author: wxsIcey

.github/workflows/accuracy_test.yaml

@@ -173,6 +173,8 @@ jobs:
 
       - name: Install vllm-project/vllm-ascend
         working-directory: ./vllm-ascend
+        env:
+          PIP_EXTRA_INDEX_URL: https://mirrors.huaweicloud.com/ascend/repos/pypi
         run: |
           pip install -r requirements-dev.txt
           pip install -e .

.github/workflows/image_openeuler.yml

@@ -19,6 +19,12 @@ on:
       - '.github/workflows/image_openeuler.yml'
       - 'Dockerfile.openEuler'
       - 'vllm_ascend/**'
+      - 'setup.py'
+      - 'pyproject.toml'
+      - 'requirements.txt'
+      - 'cmake/**'
+      - 'CMakeLists.txt'
+      - 'csrc/**'
   push:
     # Publish image when tagging, the Dockerfile in tag will be build as tag image
     branches:

.github/workflows/image_ubuntu.yml

@@ -19,6 +19,12 @@ on:
       - '.github/workflows/image_ubuntu.yml'
       - 'Dockerfile'
       - 'vllm_ascend/**'
+      - 'setup.py'
+      - 'pyproject.toml'
+      - 'requirements.txt'
+      - 'cmake/**'
+      - 'CMakeLists.txt'
+      - 'csrc/**'
   push:
     # Publish image when tagging, the Dockerfile in tag will be build as tag image
     branches:

.github/workflows/nightly_benchmarks.yaml

@@ -115,6 +115,8 @@ jobs:
           VLLM_TARGET_DEVICE=empty pip install -e .
 
       - name: Install vllm-project/vllm-ascend
+        env:
+          PIP_EXTRA_INDEX_URL: https://mirrors.huaweicloud.com/ascend/repos/pypi
         run: |
           pip install -e .
           pip install -r benchmarks/requirements-bench.txt

.github/workflows/vllm_ascend_test.yaml

@@ -114,6 +114,57 @@ jobs:
           echo "::add-matcher::.github/workflows/matchers/mypy.json"
           tools/mypy.sh 1 ${{ matrix.python-version }}
 
+  ut:
+    needs: [lint]
+    name: unit test
+    if: ${{ needs.lint.result == 'success' }}
+    runs-on: ubuntu-latest
+    container:
+      image: m.daocloud.io/quay.io/ascend/cann:8.1.rc1-910b-ubuntu22.04-py3.10
+      env:
+        VLLM_LOGGING_LEVEL: ERROR
+        VLLM_USE_MODELSCOPE: True
+    strategy:
+      matrix:
+        vllm_version: [main, v0.9.1]
+    steps:
+      - name: Install packages
+        run: |
+          apt-get update -y
+          apt-get install -y python3-pip git vim wget net-tools gcc g++ cmake libnuma-dev
+
+      - name: Checkout vllm-project/vllm repo
+        uses: actions/checkout@v4
+        with:
+          repository: vllm-project/vllm
+          ref: ${{ matrix.vllm_version }}
+          path: ./vllm-empty
+
+      - name: Install vllm-project/vllm from source
+        working-directory: ./vllm-empty
+        run: |
+          VLLM_TARGET_DEVICE=empty python3 -m pip install . --extra-index https://download.pytorch.org/whl/cpu/
+          python3 -m pip uninstall -y triton
+
+      - name: Checkout vllm-project/vllm-ascend repo
+        uses: actions/checkout@v4
+
+      - name: Install vllm-project/vllm-ascend
+        run: |
+          export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi
+          export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/x86_64-linux/devlib
+          python3 -m pip install -r requirements-dev.txt --extra-index https://download.pytorch.org/whl/cpu/
+          python3 -m pip install -v . --extra-index https://download.pytorch.org/whl/cpu/
+
+      - name: Run unit test for V1 Engine
+        env:
+          VLLM_USE_V1: 1
+          VLLM_WORKER_MULTIPROC_METHOD: spawn
+          TORCH_DEVICE_BACKEND_AUTOLOAD: 0
+        run: |
+          export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/x86_64-linux/devlib
+          pytest -sv tests/ut
+
   e2e:
     needs: [lint]
     if: ${{ needs.lint.result == 'success' }}
@@ -164,6 +215,8 @@ jobs:
           VLLM_TARGET_DEVICE=empty pip install -e .
 
       - name: Install vllm-project/vllm-ascend
+        env:
+          PIP_EXTRA_INDEX_URL: https://mirrors.huaweicloud.com/ascend/repos/pypi
         run: |
           pip install -r requirements-dev.txt
           pip install -v -e .
@@ -267,6 +320,8 @@ jobs:
           VLLM_TARGET_DEVICE=empty pip install -e .
 
       - name: Install vllm-project/vllm-ascend
+        env:
+          PIP_EXTRA_INDEX_URL: https://mirrors.huaweicloud.com/ascend/repos/pypi
         run: |
           pip install -r requirements-dev.txt
           pip install -v -e .

.github/workflows/vllm_ascend_test_long_term.yaml

@@ -88,6 +88,8 @@ jobs:
           VLLM_TARGET_DEVICE=empty pip install -e .
 
       - name: Install vllm-project/vllm-ascend
+        env:
+          PIP_EXTRA_INDEX_URL: https://mirrors.huaweicloud.com/ascend/repos/pypi
         run: |
           pip install -r requirements-dev.txt
           pip install -v -e .

.github/workflows/vllm_ascend_test_pd.yaml

@@ -97,6 +97,8 @@ jobs:
           VLLM_TARGET_DEVICE=empty pip install -e .
 
       - name: Install vllm-project/vllm-ascend
+        env:
+          PIP_EXTRA_INDEX_URL: https://mirrors.huaweicloud.com/ascend/repos/pypi
         run: |
           pip install -r requirements-dev.txt
           pip install -v -e .

Dockerfile

@@ -46,7 +46,8 @@ RUN VLLM_TARGET_DEVICE="empty" python3 -m pip install -v -e /vllm-workspace/vllm
 
 # Install vllm-ascend
 # Append `libascend_hal.so` path (devlib) to LD_LIBRARY_PATH
-RUN source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
+RUN export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi && \
+    source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
     source /usr/local/Ascend/nnal/atb/set_env.sh && \
     export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
     python3 -m pip install -v -e /vllm-workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/ && \

Dockerfile.openEuler

@@ -43,7 +43,8 @@ RUN VLLM_TARGET_DEVICE="empty" python3 -m pip install -e /vllm-workspace/vllm/ -
     python3 -m pip cache purge
 
 # Install vllm-ascend
-RUN source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
+RUN export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi && \
+    source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
     source /usr/local/Ascend/nnal/atb/set_env.sh && \
     export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
     python3 -m pip install -v -e /vllm-workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/ && \

README.md

@@ -38,7 +38,7 @@ By using vLLM Ascend plugin, popular open-source models, including Transformer-l
 - Software:
   * Python >= 3.9, < 3.12
   * CANN >= 8.1.RC1
-  * PyTorch >= 2.5.1, torch-npu >= 2.5.1
+  * PyTorch >= 2.5.1, torch-npu >= 2.5.1.post1.dev20250528
   * vLLM (the same version as vllm-ascend)
 
 ## Getting Started

README.zh.md

@@ -39,7 +39,7 @@ vLLM 昇腾插件 (`vllm-ascend`) 是一个由社区维护的让vLLM在Ascend NP
 - 软件：
   * Python >= 3.9, < 3.12
   * CANN >= 8.1.RC1
-  * PyTorch >= 2.5.1, torch-npu >= 2.5.1
+  * PyTorch >= 2.5.1, torch-npu >= 2.5.1.post1.dev20250528
   * vLLM (与vllm-ascend版本一致)
 
 ## 开始使用

docs/source/installation.md

@@ -9,11 +9,11 @@ This document describes how to install vllm-ascend manually.
 - A hardware with Ascend NPU. It's usually the Atlas 800 A2 series.
 - Software:
 
-    | Software  | Supported version | Note                                   |
-    |-----------|-------------------|----------------------------------------| 
-    | CANN      | >= 8.1.RC1        | Required for vllm-ascend and torch-npu |
-    | torch-npu | >= 2.5.1          | Required for vllm-ascend               |
-    | torch     | >= 2.5.1          | Required for torch-npu and vllm        |
+    | Software      | Supported version                | Note                                      |
+    |---------------|----------------------------------|-------------------------------------------|
+    | CANN          | >= 8.1.RC1                       | Required for vllm-ascend and torch-npu    |
+    | torch-npu     | >= 2.5.1.post1.dev20250528       | Required for vllm-ascend                  |
+    | torch         | >= 2.5.1                         | Required for torch-npu and vllm           |
 
 You have 2 way to install:
 - **Using pip**: first prepare env manually or via CANN image, then install `vllm-ascend` using pip.
@@ -156,6 +156,7 @@ cd ..
 # Install vLLM Ascend
 git clone  --depth 1 --branch |vllm_ascend_version| https://github.com/vllm-project/vllm-ascend.git
 cd vllm-ascend
+export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi
 pip install -v -e .
 cd ..
 ```

requirements.txt

@@ -10,8 +10,6 @@ pyyaml
 scipy
 setuptools>=64
 setuptools-scm>=8
---extra-index-url https://mirrors.huaweicloud.com/ascend/repos/pypi
-torch-npu==2.5.1.post1.dev20250528
 torch>=2.5.1
 torchvision<0.21.0
 wheel
@@ -22,3 +20,8 @@ quart
 
 # Required for N-gram speculative decoding
 numba
+
+# Install torch_npu
+--pre
+--extra-index-url https://mirrors.huaweicloud.com/ascend/repos/pypi
+torch-npu==2.5.1.post1.dev20250528

setup.py

@@ -152,7 +152,7 @@ def configure(self, ext: CMakeExtension) -> None:
             # if pybind11 is installed via pip
             pybind11_cmake_path = (subprocess.check_output(
                 [python_executable, "-m", "pybind11",
-                 "--cmake"]).decode().strip())
+                 "--cmakedir"]).decode().strip())
         except subprocess.CalledProcessError as e:
             # else specify pybind11 path installed from source code on CI container
             raise RuntimeError(f"CMake configuration failed: {e}")