为机器学习创建一个 Docker 容器

Question

我不是 Docker 专家，我正在尝试为机器学习项目创建一个容器。 这主要是出于学术目的，因为我正在研究机器学习。

我写了一个 dockerfile（和一个 devcontainer.json 文件来在 vscode 中打开容器），它运行良好，直到我添加了构建 tensorflow 的行。我发现了三个问题，但我不知道我遗漏了什么：

• 我收到关于 Bazel 安装不是发行版的警告
• 当building阶段到了./configure我无法通过提示问题进行配置
• 我在构建阶段遇到错误， ERROR: The project you're trying to build requires Bazel 5.3.0 (specified in /docklearning/tensorflow/.bazelversion), but it wasn't found in /usr/bin.

这是 Dockerfile：

FROM nvidia/cuda:12.0.0-runtime-ubuntu22.04 as base

ARG USER_UID=1000

#switch to non-interactive frontend
ENV DEBIAN_FRONTEND=noninteractive

WORKDIR /docklearning
ADD . /docklearning

# Install packages
RUN apt-get update -q && apt-get install -q -y --no-install-recommends \
    apt-transport-https curl gnupg apt-utils wget gcc g++ npm unzip build-essential ca-certificates curl git gh \
    make nano iproute2 nano openssh-client openssl procps \
    software-properties-common bzip2 subversion neofetch \
    fontconfig && \
    curl -fsSL https://bazel.build/bazel-release.pub.gpg | gpg --dearmor >bazel-archive-keyring.gpg && \
    mv bazel-archive-keyring.gpg /usr/share/keyrings && \
    echo "deb [arch=amd64 signed-by=/usr/share/keyrings/bazel-archive-keyring.gpg] https://storage.googleapis.com/bazel-apt stable jdk1.8" \
    | tee /etc/apt/sources.list.d/bazel.list && \
    apt update && apt install -q -y bazel && \
    apt-get full-upgrade -q -y && \
    cd ~ && \
    wget https://github.com/ryanoasis/nerd-fonts/releases/download/v2.1.0/Meslo.zip && \
    mkdir -p .local/share/fonts && \
    unzip Meslo.zip -d .local/share/fonts && \
    cd .local/share/fonts && rm *Windows* && \
    cd ~ && \
    rm Meslo.zip && \
    fc-cache -fv && \
    apt-get install -y zsh zsh-doc chroma
    
# Install anaconda
RUN wget https://repo.anaconda.com/archive/Anaconda3-2022.10-Linux-x86_64.sh -O Anaconda.sh && \
    /bin/bash Anaconda.sh -b -p /opt/conda && \
    rm Anaconda.sh

# Install LSD for ls substitute and clean up
RUN wget https://github.com/Peltoche/lsd/releases/download/0.23.1/lsd_0.23.1_amd64.deb -P /tmp && \
    dpkg -i /tmp/lsd_0.23.1_amd64.deb && \
    rm /tmp/lsd_0.23.1_amd64.deb && \
    apt-get autoremove -y && \
    apt-get autoclean && \
    rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*/apt/lists/* && \
    useradd -r -m -s /bin/bash -u ${USER_UID} docklearning

# Add conda to PATH
ENV PATH=/opt/conda/bin:$PATH

ENV HOME=/home/docklearning

# Make zsh default shell
RUN chsh -s /usr/bin/zsh docklearning

# link conda to /etc/profile.d
RUN ln -s /opt/conda/etc/profile.d/conda.sh /etc/profile.d/conda.sh

# Update Conda Base env
RUN conda update -n base --all -y && \
    conda install -c conda-forge bandit opt_einsum keras-preprocessing && \
    conda clean -a -q -y
     
# Download Tensorflow from github and build from source
RUN git clone https://github.com/tensorflow/tensorflow.git && \
    cd tensorflow && \
    ./configure && \
    bazel build --config=opt --config=cuda --cxxopt="-mavx2" //tensorflow/tools/pip_package:build_pip_package && \
    ./bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg && \
    pip install /tmp/tensorflow_pkg/tensorflow-*.whl && \
    cd .. && \
    rm -rf tensorflow


# Create shell history file
RUN mkdir ${HOME}/zsh_history && \
    chown docklearning ${HOME}/zsh_history && \
    mkdir ${HOME}/.ssh

# Switch to internal user
USER docklearning
WORKDIR ${HOME}

# Copy user configuration files
COPY --chown=docklearning ./config/.aliases.sh ./
COPY --chown=docklearning ./config/.bashrc ./
COPY --chown=docklearning ./config/.nanorc ./

# Configure Zsh for internal user
ENV ZSH=${HOME}/.oh-my-zsh
ENV ZSH_CUSTOM=${ZSH}/custom
ENV ZSH_PLUGINS=${ZSH_CUSTOM}/plugins
ENV ZSH_THEMES=${ZSH_CUSTOM}/themes
RUN wget -qO- https://raw.githubusercontent.com/ohmyzsh/ohmyzsh/master/tools/install.sh | zsh || true
RUN git clone --single-branch --branch 'master' --depth 1 https://github.com/zsh-users/zsh-syntax-highlighting.git ${ZSH_PLUGINS}/zsh-syntax-highlighting \
    && git clone --single-branch --branch 'master' --depth 1 https://github.com/zsh-users/zsh-autosuggestions ${ZSH_PLUGINS}/zsh-autosuggestions \
    && git clone --single-branch --depth 1 https://github.com/romkatv/powerlevel10k.git ${ZSH_THEMES}/powerlevel10k
COPY --chown=docklearning ./config/.p10k.zsh ./
COPY --chown=docklearning ./config/.zshrc ./


CMD [ "/bin/zsh" ]

Answer 1

我不会说这个关于 docker 或 TensorFlow 的问题。它是关于 package 安装和操作系统配置的。

在这种情况下，您需要首先通过明确指定来修复 bazel 版本：

apt install -q -y bazel-5.3.0

但是 package 非常有线并且不会创建符号链接，因此您需要创建它：

ln -s /bin/bazel-5.3.0 /bin/bazel

您可以通过运行它来验证是否安装了 bazel：

`bazel --version
 bazel 5.3.0`