`tf.nn.safe_embedding_lookup_sparse`, `tf.nn.sampled_softmax` and
        `tf.nn.nce_loss`. hyperparameter are ignored.
*   Performance
    *   Improve performance of GPU cumsum/cumprod by up to 300x.
    *   Added support for weight decay in most TPU embedding optimizers,
        including AdamW and MomentumW.
*   TensorFlow 2.0 Development
    *   Add a command line tool to convert to TF2.0, tf_upgrade_v2

test:release_gpu_linux --test_env=LD_LIBRARY_PATH="/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64"
# Local test jobs has to be 4 because parallel_gpu_execute is fragile, I think
test:release_gpu_linux --test_timeout=300,450,1200,3600 --local_test_jobs=4 --run_under=//tensorflow/tools/ci_build/gpu_build:parallel_gpu_execute

build:release_arm64_linux --config=release_linux_base
build:release_arm64_linux --config=linux_arm64