examples/spark_dataset_converter/pytorch_converter_example.py [142:164]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    accuracy = train_and_evaluate()
    logging.info("Train and evaluate the model on the local machine.")
    logging.info("Accuracy: %.6f", accuracy)

    # Train and evaluate the model on a spark worker
    accuracy = spark.sparkContext.parallelize(range(1)).map(train_and_evaluate).collect()[0]
    logging.info("Train and evaluate the model remotely on a spark worker, "
                 "which can be used for distributed hyperparameter tuning.")
    logging.info("Accuracy: %.6f", accuracy)

    # Cleanup
    converter_train.delete()
    converter_test.delete()
    spark.stop()


def main():
    mnist_dir = get_mnist_dir()
    run(data_dir=mnist_dir)


if __name__ == '__main__':
    main()
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



examples/spark_dataset_converter/tensorflow_converter_example.py [94:116]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    accuracy = train_and_evaluate()
    logging.info("Train and evaluate the model on the local machine.")
    logging.info("Accuracy: %.6f", accuracy)

    # Train and evaluate the model on a spark worker
    accuracy = spark.sparkContext.parallelize(range(1)).map(train_and_evaluate).collect()[0]
    logging.info("Train and evaluate the model remotely on a spark worker, "
                 "which can be used for distributed hyperparameter tuning.")
    logging.info("Accuracy: %.6f", accuracy)

    # Cleanup
    converter_train.delete()
    converter_test.delete()
    spark.stop()


def main():
    mnist_dir = get_mnist_dir()
    run(data_dir=mnist_dir)


if __name__ == '__main__':
    main()
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -