ddhmed/vllm repository overview

########################################################
# sudo docker build -t ddhmed/vllm:20240603 .
# sudo ufw allow 5000
# sudo docker run --rm -v $PWD/models:/models -p 5000:5000 ddhmed/vllm:20240603
# sudo docker run --gpus all --rm -v $PWD/models:/models -p 5000:5000 ddhmed/vllm:20240603
# sudo docker run -it --gpus all --rm -v $PWD/models:/models -p 5000:5000 ddhmed/vllm:20240603 /bin/bash
########################################################
# 使用NVIDIA提供的CUDA基础镜像
FROM nvidia/cuda:11.8.0-cudnn8-runtime-ubuntu20.04

### 作者和邮箱
MAINTAINER ddhmed [email protected]

USER root

# 设置时区，避免交互式提示
ENV TZ=Asia/Shanghai
RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo $TZ > /etc/timezone

# 安装基本工具和Python
RUN apt-get update && apt-get install -y \
    python3 \
    python3-pip \
    && rm -rf /var/lib/apt/lists/*

# 创建一个符号链接，使 python3 命令可以用 python 调用
RUN ln -s /usr/bin/python3 /usr/bin/python

# 升级pip
RUN pip3 install --upgrade pip

# install base packages
RUN pip install --no-cache-dir \
    vllm \
    transformers -i https://pypi.tuna.tsinghua.edu.cn/simple

RUN pip install gradio -i https://pypi.tuna.tsinghua.edu.cn/simple
    
# 清理不必要的文件
RUN apt-get clean && \
    rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*

### 设置工作目录 
WORKDIR /models

# 暴露端口
EXPOSE 8000 7860

# 启动 vllm 模型服务和 Gradio 应用
CMD ["sh", "-c", "CUDA_VISIBLE_DEVICES=0 python -m vllm.entrypoints.openai.api_server --model './model' --served-model-name 'openchat' --port 8000 & python app.py"]