# Pulling an already-pulled container image will finish instantly, so
  # repeating the command costs nothing.
  docker pull "$TFCI_DOCKER_IMAGE" || sleep 15
  docker pull "$TFCI_DOCKER_IMAGE" || sleep 15
  docker pull "$TFCI_DOCKER_IMAGE"
fi 

if [[ "$TFCI_DOCKER_REBUILD_ENABLE" == 1 ]]; then
  DOCKER_BUILDKIT=1 docker build --cache-from "$TFCI_DOCKER_IMAGE" -t "$TFCI_DOCKER_IMAGE" $TFCI_DOCKER_REBUILD_ARGS

# different compilation methods. E.g. for a PR to test a new CUDA version, set
# the CACHEBUSTER to the PR number.
build --action_env=CACHEBUSTER=20220325

# Use Python 3.X as installed in container image
build --action_env PYTHON_BIN_PATH="/usr/local/bin/python3"
build --python_path="/usr/local/bin/python3"

# Build TensorFlow v2
build --define=tf_api_version=2 --action_env=TF2_BEHAVIOR=1

# CuDNN: https://docs.nvidia.com/deeplearning/sdk/cudnn-install/index.html#ubuntu-network-installation
libcudnn8=8.9.6.50-1+cuda12.2
libcudnn8-dev=8.9.6.50-1+cuda12.2

# This can be removed once NVIDIA publishes a cuda-12.3.2 Docker image.
# For now it ensures that we install at least version 12.3.107 of PTXAS,
# since 12.3.103 has a bug.

# to tensorflow_cpu. These ARM builds are supposed to have the name "tensorflow"
# despite lacking Nvidia CUDA support.
TFCI_BUILD_PIP_PACKAGE_ARGS="--repo_env=WHEEL_NAME=tensorflow"
TFCI_DOCKER_ENABLE=1
TFCI_DOCKER_IMAGE=gcr.io/tensorflow-sigs/build-arm64:tf-2-16-multi-python
TFCI_DOCKER_PULL_ENABLE=1
TFCI_DOCKER_REBUILD_ARGS="--target=tf ci/official/containers/linux_arm64"
TFCI_INDEX_HTML_ENABLE=1
TFCI_LIB_SUFFIX="-cpu-linux-arm64"

    --build-arg REQUIREMENTS_FILE=jax.requirements.txt \
    --target=$target \
    --cache-from "$IMAGE" \
    -t "$IMAGE"  . && break
  done
  final=$?
  if [ $final -ne 0 ]; then
    exit $final
  fi
  set -e

  if [[ -n "$KOKORO_BUILD_ID" ]]; then
    gcloud auth configure-docker
    docker push "$IMAGE"
  fi

    - name: Update the RBE Configs
      run: |
        function map() {
          # The "digest" that allows us to pull an image is not the digest as
          # returned by the API, but a sha256sum of the entire chunk of image
          # metadata. gcr.io helpfully includes it in the header of the response
          # as docker-content-digest: sha256:[digest]. Note we use egrep to

            echo "No running containers";
          else
            echo "Running container(s) found" && \
            docker stop $running_containers;
          fi
          docker container prune -f
          docker image prune -af
      - name: Clean repository
        shell: bash
        run: find /home/ubuntu/actions-runner/_work/tensorflow/tensorflow/. -name . -o -prune -exec sudo rm -rf -- {} + || true

/usr/bin/$VERSION get-pip.py
/usr/bin/$VERSION -m pip install --no-cache-dir --upgrade pip

# Disable the cache dir to save image space, and install packages

            echo "No running containers";
          else
            echo "Running container(s) found" && \
            docker stop $running_containers;
          fi
          docker container prune -f
          docker image prune -af
      - name: Clean repository
        shell: bash
        run: find /home/ubuntu/actions-runner/_work/tensorflow/tensorflow/. -name . -o -prune -exec sudo rm -rf -- {} + || true

            
            Re-apply the `build and push to gcr.io for staging` label to rebuild and push again. This comment will only be posted once.
      -
        name: Print image digest