vllm0.9.2的cpu版本用docker部署qwen3:0.6B

现在Docker镜像仓库已经没有cpu低版本的vllm来部署大语言模型，我用源码来重新构建了一个docker镜像，然后部署了Qwen3:0.6B的模型，在cpu上部署的。

1. 拉取vllm源码：https://github.com/vllm-project/vllm.git

2. 用git命令切换到v0.9.2版本

3. 进入源码的docker文件夹下，有一个Dockerfile.cpu文件，这就是构建镜像的脚本

4. 我这边直接以Dockerfile.cpu构建镜像报错了，打开该文件，发现里面安装依赖时用的是requirements/cpu.txt

5. 我这边报错是transformers的错误，我在cpu.txt中加了一行

transformers==4.51.1

6. 然后再以Dockerfile.cpu来构建镜像，就可以了

docker build -f ./docker/Dockerfile.cpu  -t vllm-cpu-env:v0.9.2 --build-arg VLLM_CPU_DISABLE_AVX512=true --shm-size=4g .

7. Dockerfile.cpu文件的内容如下，我这里记录下来：

# This vLLM Dockerfile is used to construct image that can build and run vLLM on x86 CPU platform.
#
# Build targets:
#   vllm-openai (default): used for serving deployment
#   vllm-test: used for CI tests
#   vllm-dev: used for development
#
# Build arguments:
#   PYTHON_VERSION=3.12 (default)|3.11|3.10|3.9
#   VLLM_CPU_DISABLE_AVX512=false (default)|true
#   VLLM_CPU_AVX512BF16=false (default)|true
#   VLLM_CPU_AVX512VNNI=false (default)|true
#

######################### BASE IMAGE #########################
FROM ubuntu:22.04 AS base

WORKDIR /workspace/

ARG PYTHON_VERSION=3.12
ARG PIP_EXTRA_INDEX_URL="https://download.pytorch.org/whl/cpu"

ENV LD_PRELOAD=""

# Install minimal dependencies and uv
RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
    --mount=type=cache,target=/var/lib/apt,sharing=locked \
    apt-get update -y \
    && apt-get install -y --no-install-recommends ccache git curl wget ca-certificates \
        gcc-12 g++-12 libtcmalloc-minimal4 libnuma-dev ffmpeg libsm6 libxext6 libgl1 jq lsof \
    && update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-12 10 --slave /usr/bin/g++ g++ /usr/bin/g++-12 \
    && curl -LsSf https://astral.sh/uv/install.sh | sh \
    && echo 'export PATH="/root/.local/bin:$PATH"' >> /root/.bashrc \
    && /root/.local/bin/uv --version 

# 立即更新 PATH 并设置缓存和编译相关环境变量
ENV CCACHE_DIR=/root/.cache/ccache
ENV CMAKE_CXX_COMPILER_LAUNCHER=ccache
ENV PATH="/root/.local/bin:$PATH" 

# 创建虚拟环境（此时 uv 已在 PATH 中）
ENV VIRTUAL_ENV="/opt/venv"
ENV UV_PYTHON_INSTALL_DIR=/opt/uv/python
RUN uv venv --python ${PYTHON_VERSION} --seed ${VIRTUAL_ENV}
ENV PATH="$VIRTUAL_ENV/bin:$PATH" 

ENV UV_HTTP_TIMEOUT=500

# Install Python dependencies 
ENV PIP_EXTRA_INDEX_URL=${PIP_EXTRA_INDEX_URL}
ENV UV_EXTRA_INDEX_URL=${PIP_EXTRA_INDEX_URL}
ENV UV_INDEX_STRATEGY="unsafe-best-match"
ENV UV_LINK_MODE="copy"
RUN --mount=type=cache,target=/root/.cache/uv \
    --mount=type=bind,src=requirements/common.txt,target=requirements/common.txt \
    --mount=type=bind,src=requirements/cpu.txt,target=requirements/cpu.txt \
    uv pip install --upgrade pip && \
    uv pip install -r requirements/cpu.txt

ENV LD_PRELOAD="/usr/lib/x86_64-linux-gnu/libtcmalloc_minimal.so.4:/opt/venv/lib/libiomp5.so:$LD_PRELOAD"

RUN echo 'ulimit -c 0' >> ~/.bashrc

######################### BUILD IMAGE #########################
FROM base AS vllm-build

ARG GIT_REPO_CHECK=0
# Support for building with non-AVX512 vLLM: docker build --build-arg VLLM_CPU_DISABLE_AVX512="true" ...
ARG VLLM_CPU_DISABLE_AVX512=0
ENV VLLM_CPU_DISABLE_AVX512=${VLLM_CPU_DISABLE_AVX512}
# Support for building with AVX512BF16 ISA: docker build --build-arg VLLM_CPU_AVX512BF16="true" ...
ARG VLLM_CPU_AVX512BF16=0
ENV VLLM_CPU_AVX512BF16=${VLLM_CPU_AVX512BF16}
# Support for building with AVX512VNNI ISA: docker build --build-arg VLLM_CPU_AVX512VNNI="true" ...
ARG VLLM_CPU_AVX512VNNI=0
ENV VLLM_CPU_AVX512VNNI=${VLLM_CPU_AVX512VNNI}

WORKDIR /workspace/vllm

RUN --mount=type=cache,target=/root/.cache/uv \
    --mount=type=bind,src=requirements/cpu-build.txt,target=requirements/build.txt \
    uv pip install -r requirements/build.txt

COPY . .
RUN --mount=type=bind,source=.git,target=.git \
    if [ "$GIT_REPO_CHECK" != 0 ]; then bash tools/check_repo.sh ; fi

RUN --mount=type=cache,target=/root/.cache/uv \
    --mount=type=cache,target=/root/.cache/ccache \
    --mount=type=cache,target=/workspace/vllm/.deps,sharing=locked \
    --mount=type=bind,source=.git,target=.git \
    VLLM_TARGET_DEVICE=cpu python3 setup.py bdist_wheel 

######################### TEST DEPS #########################
FROM base AS vllm-test-deps

WORKDIR /workspace/vllm

RUN --mount=type=bind,src=requirements/test.in,target=requirements/test.in \
    cp requirements/test.in requirements/cpu-test.in && \
    sed -i '/mamba_ssm/d' requirements/cpu-test.in && \
    sed -i 's/torch==.*/torch==2.6.0/g' requirements/cpu-test.in && \
    sed -i 's/torchaudio.*/torchaudio/g' requirements/cpu-test.in && \
    sed -i 's/torchvision.*/torchvision/g' requirements/cpu-test.in && \
    uv pip compile requirements/cpu-test.in -o requirements/cpu-test.txt --index-strategy unsafe-best-match --torch-backend cpu

RUN --mount=type=cache,target=/root/.cache/uv \
    uv pip install -r requirements/cpu-test.txt 

######################### DEV IMAGE #########################
FROM vllm-build AS vllm-dev

WORKDIR /workspace/vllm

RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
    --mount=type=cache,target=/var/lib/apt,sharing=locked \
    apt-get install -y --no-install-recommends vim numactl xz-utils

# install development dependencies (for testing)
RUN --mount=type=cache,target=/root/.cache/uv \
    uv pip install -e tests/vllm_test_utils 

RUN --mount=type=cache,target=/root/.cache/uv \
    --mount=type=cache,target=/root/.cache/ccache \
    --mount=type=bind,source=.git,target=.git \
    VLLM_TARGET_DEVICE=cpu python3 setup.py develop 

COPY --from=vllm-test-deps /workspace/vllm/requirements/cpu-test.txt requirements/test.txt

RUN --mount=type=cache,target=/root/.cache/uv \
    uv pip install -r requirements/dev.txt && \
    pre-commit install --hook-type pre-commit --hook-type commit-msg

ENTRYPOINT ["bash"]

######################### TEST IMAGE #########################
FROM vllm-test-deps AS vllm-test

WORKDIR /workspace/

RUN --mount=type=cache,target=/root/.cache/uv \
    --mount=type=bind,from=vllm-build,src=/workspace/vllm/dist,target=dist \
    uv pip install dist/*.whl

ADD ./tests/ ./tests/
ADD ./examples/ ./examples/
ADD ./benchmarks/ ./benchmarks/
ADD ./vllm/collect_env.py .
ADD ./.buildkite/ ./.buildkite/

# install development dependencies (for testing)
RUN --mount=type=cache,target=/root/.cache/uv \
    uv pip install -e tests/vllm_test_utils 

ENTRYPOINT ["bash"]

######################### RELEASE IMAGE #########################
FROM base AS vllm-openai

WORKDIR /workspace/

RUN --mount=type=cache,target=/root/.cache/uv \
    --mount=type=cache,target=/root/.cache/ccache \
    --mount=type=bind,from=vllm-build,src=/workspace/vllm/dist,target=dist \
    uv pip install dist/*.whl

ENTRYPOINT ["python3", "-m", "vllm.entrypoints.openai.api_server"]

8. 运行Qwen3:0.6B模型

sudo docker run -v ./Qwen:/home/models --name vllm-qwen -p 9091:8000  vllm-cpu-env:v0.9.2 --model  /home/models/Qwen3-0.6B  --served-model-name qwen3:0.6b   --enable-reasoning --reasoning-parser deepseek_r1 --max-model-len 30720 --gpu-memory-utilization 0.95  --disable-log-stats

9. 之后我还用这个镜像部署过qwen的嵌入模型，还部署过重排模型

# 嵌入模型 
docker run \
  -v ./Qwen:/home/models  \
  --name vllm-embedding \
  -p 9092:8000 \
  vllm-cpu-env:v0.9.2 \
  --model /home/models/Qwen3-Embedding-0.6B \
  --served-model-name qwen3-embedding:0.6b \
  --task embed
  --max-model-len 30720 
  --gpu-memory-utilization 0.95 
  --disable-log-stats


# 重排模型
docker run \
  -v ./BAAI:/home/models  \
  --name vllm-reranker \
  -p 9095:8000 \
  vllm-cpu-env:v0.9.2 \
  --model /home/models/bge-reranker-v2-m3 \
  --served-model-name bge-reranker-v2-m3 \
  --task score
  --disable-log-stats

10. 后面有GPU环境了，我还用vllm的GPU版本镜像部署过多模态模型

docker run --gpus '"device=0,1"' \
  -p 33003:8000 \
  --ipc=host \
  -v /mnt/inaisfs/user-fs/iei/Qwen3.6-35B-A3B:/home/models/Qwen3.6-35B-A3B \
  vllm/vllm-openai:v0.22.0-cu129 \
    --model  /home/models/Qwen3.6-35B-A3B \
    --served-model-name Qwen/Qwen3.6-35B-A3B \
    --tensor-parallel-size 2 \
    --reasoning-parser qwen3 \
    --enable-prefix-caching \
    --max-model-len 262144