Dockerfile

FROM alpine:3.7

# Based on https://github.com/tatsushid/docker-alpine-py3-tensorflow-jupyter/blob/master/Dockerfile
# Changes:
# - Bumping versions of Bazel and Tensorflow
# - Add -Xmx to the Java params when building Bazel
# - Disable TF_GENERATE_BACKTRACE and TF_GENERATE_STACKTRACE

ENV JAVA_HOME /usr/lib/jvm/java-1.8-openjdk
ENV LOCAL_RESOURCES 2048,.5,1.0
ENV BAZEL_VERSION 0.10.0
RUN apk add --no-cache python3 python3-tkinter py3-numpy py3-numpy-f2py freetype libpng libjpeg-turbo imagemagick graphviz git
RUN apk add --no-cache --virtual=.build-deps \
        bash \
        cmake \
        curl \
        freetype-dev \
        g++ \
        libjpeg-turbo-dev \
        libpng-dev \
        linux-headers \
        make \
        musl-dev \
        openblas-dev \
        openjdk8 \
        patch \
        perl \
        python3-dev \
        py-numpy-dev \
        rsync \
        sed \
        swig \
        zip \
    && cd /tmp \
    && pip3 install --no-cache-dir wheel \
    && $(cd /usr/bin && ln -s python3 python)

# Bazel download
RUN curl -SLO https://github.com/bazelbuild/bazel/releases/download/${BAZEL_VERSION}/bazel-${BAZEL_VERSION}-dist.zip \
    && mkdir bazel-${BAZEL_VERSION} \
    && unzip -qd bazel-${BAZEL_VERSION} bazel-${BAZEL_VERSION}-dist.zip

# Bazel install
RUN cd bazel-${BAZEL_VERSION} \
    && sed -i -e 's/-classpath/-J-Xmx8192m -J-Xms128m -classpath/g' scripts/bootstrap/compile.sh \
    && bash compile.sh \
    && cp -p output/bazel /usr/bin/

# Download Tensorflow
ENV TENSORFLOW_VERSION 1.7.0
RUN cd /tmp \
    && curl -SL https://github.com/tensorflow/tensorflow/archive/v${TENSORFLOW_VERSION}.tar.gz \
        | tar xzf -

# Build Tensorflow
RUN cd /tmp/tensorflow-${TENSORFLOW_VERSION} \
    && : musl-libc does not have "secure_getenv" function \
    && sed -i -e '/JEMALLOC_HAVE_SECURE_GETENV/d' third_party/jemalloc.BUILD \
    && sed -i -e '/define TF_GENERATE_BACKTRACE/d' tensorflow/core/platform/default/stacktrace.h \
    && sed -i -e '/define TF_GENERATE_STACKTRACE/d' tensorflow/core/platform/stacktrace_handler.cc \
    && PYTHON_BIN_PATH=/usr/bin/python \
        PYTHON_LIB_PATH=/usr/lib/python3.6/site-packages \
        CC_OPT_FLAGS="-march=native" \
        TF_NEED_JEMALLOC=1 \
        TF_NEED_GCP=0 \
        TF_NEED_HDFS=0 \
        TF_NEED_S3=0 \
        TF_ENABLE_XLA=0 \
        TF_NEED_GDR=0 \
        TF_NEED_VERBS=0 \
        TF_NEED_OPENCL=0 \
        TF_NEED_CUDA=0 \
        TF_NEED_MPI=0 \
        bash configure
RUN cd /tmp/tensorflow-${TENSORFLOW_VERSION} \
    && bazel build --jobs 6 -c opt --local_resources ${LOCAL_RESOURCES} //tensorflow/tools/pip_package:build_pip_package
RUN cd /tmp/tensorflow-${TENSORFLOW_VERSION} \
    && ./bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg
RUN cp /tmp/tensorflow_pkg/tensorflow-${TENSORFLOW_VERSION}-cp36-cp36m-linux_x86_64.whl /root

# Make sure it's built properly
RUN pip3 install --no-cache-dir /root/tensorflow-${TENSORFLOW_VERSION}-cp36-cp36m-linux_x86_64.whl \
    && python3 -c 'import tensorflow'