Recover vllm-ascend dev image (#209)

### What this PR does / why we need it?
Recover vllm-ascend dev image

### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
CI passed

Signed-off-by: Yikun Jiang <yikunkero@gmail.com>

.github/workflows/vllm_ascend_test.yaml

@@ -119,7 +119,5 @@ jobs:
           pytest -sv tests
 
       - name: Run vllm-project/vllm test
-        env:
-          HF_TOKEN: ${{ secrets.HF_TOKEN }}
         run: |
           pytest -sv

Dockerfile

@@ -23,7 +23,7 @@ ARG PIP_INDEX_URL="https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple"
 ENV DEBIAN_FRONTEND=noninteractive
 
 RUN apt-get update -y && \
-    apt-get install -y python3-pip git vim && \
+    apt-get install -y python3-pip git vim net-tools && \
     rm -rf /var/cache/apt/* && \
     rm -rf /var/lib/apt/lists/*
 
@@ -33,15 +33,19 @@ COPY . /workspace/vllm-ascend/
 
 RUN pip config set global.index-url ${PIP_INDEX_URL}
 
-# Install vLLM main
+# Install vLLM
 ARG VLLM_REPO=https://github.com/vllm-project/vllm.git
-RUN git clone --depth 1 $VLLM_REPO /workspace/vllm
+ARG VLLM_TAG=main
+RUN git clone --depth 1 $VLLM_REPO --branch $VLLM_TAG /workspace/vllm
 RUN VLLM_TARGET_DEVICE="empty" python3 -m pip install /workspace/vllm/
 
-# Install vllm-ascend main
+# Install vllm-ascend
 RUN python3 -m pip install /workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/
 
-# Install modelscope (for fast download) and ray/net-tools (for multinode)
-RUN python3 -m pip install modelscope ray net-tools
+# Install torch-npu
+RUN bash /workspace/vllm-ascend/pta_install.sh
+
+# Install modelscope (for fast download) and ray (for multinode)
+RUN python3 -m pip install modelscope ray
 
 CMD ["/bin/bash"]

README.md

@@ -36,7 +36,7 @@ By using vLLM Ascend plugin, popular open-source models, including Transformer-l
 - Software:
   * Python >= 3.9
   * CANN >= 8.0.0
-  * PyTorch >= 2.5.1, torch-npu >= 2.5.1.dev20250218
+  * PyTorch >= 2.5.1, torch-npu >= 2.5.1.dev20250226
   * vLLM (the same version as vllm-ascend)
 
 Find more about how to setup your environment step by step in [here](docs/source/installation.md).

README.zh.md

@@ -36,7 +36,7 @@ vLLM 昇腾插件 (`vllm-ascend`) 是一个让vLLM在Ascend NPU无缝运行的
 - 软件：
   * Python >= 3.9
   * CANN >= 8.0.RC2
-  * PyTorch >= 2.5.1, torch-npu >= 2.5.1.dev20250218
+  * PyTorch >= 2.5.1, torch-npu >= 2.5.1.dev20250226
   * vLLM (与vllm-ascend版本一致)
 
 在[此处](docs/source/installation.md)，您可以了解如何逐步准备环境。
@@ -97,4 +97,4 @@ vllm-ascend有主干分支和开发分支。
 
 ## 许可证
 
-Apache 许可证 2.0，如 [LICENSE](./LICENSE) 文件中所示。
+Apache 许可证 2.0，如 [LICENSE](./LICENSE) 文件中所示。

docs/source/conf.py

@@ -70,8 +70,8 @@ myst_substitutions = {
     'vllm_ascend_version': 'main',
     # the newest release version of vllm-ascend and matched vLLM, used in pip install.
     # This value should be updated when cut down release.
-    'pip_vllm_ascend_version': "v0.7.1rc1",
-    'pip_vllm_version': "v0.7.1",
+    'pip_vllm_ascend_version': "0.7.3rc1",
+    'pip_vllm_version': "0.7.3",
 }
 
 # Add any paths that contain templates here, relative to this directory.

docs/source/installation.md

@@ -113,21 +113,37 @@ Once it's done, you can start to set up `vllm` and `vllm-ascend`.
 :selected:
 :sync: pip
 
-You can install `vllm` and `vllm-ascend` from **pre-built wheel**:
+You can install `vllm` and `vllm-ascend` from **pre-built wheel** (**Unreleased yet**, please build from source code):
 
 ```{code-block} bash
    :substitutions:
 
-# Install vllm from source, since `pip install vllm` doesn't work on CPU currently.
-# It'll be fixed in the next vllm release, e.g. v0.7.3.
-git clone --branch |pip_vllm_version| https://github.com/vllm-project/vllm
+# Install vllm-project/vllm from pypi
+pip install vllm==|pip_vllm_version|
 
+# Install vllm-project/vllm-ascend from pypi.
+pip install vllm-ascend==|pip_vllm_ascend_version| --extra-index https://download.pytorch.org/whl/cpu/
+```
+
+or build from **source code**:
+
+```{code-block} bash
+   :substitutions:
+
+# Install vLLM
+git clone --depth 1 --branch |vllm_version| https://github.com/vllm-project/vllm
 cd vllm
 VLLM_TARGET_DEVICE=empty pip install . --extra-index https://download.pytorch.org/whl/cpu/
 
-# Install vllm-ascend from pypi.
-pip install vllm-ascend==|pip_vllm_ascend_version| --extra-index https://download.pytorch.org/whl/cpu/
+# Install vLLM Ascend
+git clone  --depth 1 --branch |vllm_ascend_version| https://github.com/vllm-project/vllm-ascend.git
+cd vllm-ascend
+pip install -e . --extra-index https://download.pytorch.org/whl/cpu/
+```
 
+Current version depends on a unreleased `torch-npu`, you need to install manually:
+
+```
 # Once the packages are installed, you need to install `torch-npu` manually,
 # because that vllm-ascend relies on an unreleased version of torch-npu.
 # This step will be removed in the next vllm-ascend release.
@@ -140,25 +156,10 @@ pip install vllm-ascend==|pip_vllm_ascend_version| --extra-index https://downloa
 #
 mkdir pta
 cd pta
-wget https://pytorch-package.obs.cn-north-4.myhuaweicloud.com/pta/Daily/v2.5.1/20250218.4/pytorch_v2.5.1_py310.tar.gz
+wget https://pytorch-package.obs.cn-north-4.myhuaweicloud.com/pta/Daily/v2.5.1/20250226.4/pytorch_v2.5.1_py310.tar.gz
 tar -xvf pytorch_v2.5.1_py310.tar.gz
 pip install ./torch_npu-2.5.1.dev20250226-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
 ```
-
-or build from **source code**:
-
-```{code-block} bash
-   :substitutions:
-
-git clone --depth 1 --branch |vllm_version| https://github.com/vllm-project/vllm
-cd vllm
-VLLM_TARGET_DEVICE=empty pip install . --extra-index https://download.pytorch.org/whl/cpu/
-
-git clone  --depth 1 --branch |vllm_ascend_version| https://github.com/vllm-project/vllm-ascend.git
-cd vllm-ascend
-pip install -e . --extra-index https://download.pytorch.org/whl/cpu/
-```
-
 ::::
 
 ::::{tab-item} Using docker

pta_install.sh

@@ -0,0 +1,15 @@
+#!/bin/bash
+mkdir pta
+cd pta || exit
+wget https://pytorch-package.obs.cn-north-4.myhuaweicloud.com/pta/Daily/v2.5.1/20250226.4/pytorch_v2.5.1_py310.tar.gz
+tar -zxvf pytorch_v2.5.1_py310.tar.gz
+
+if [ "$(uname -i)" == "aarch64" ]
+then
+    pip install ./torch_npu-2.5.1.dev20250226-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
+else
+    pip install ./torch_npu-2.5.1.dev20250226-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
+fi
+
+cd ..
+rm -rf pta