infra: add training script to benchmark directory (#349)

Author: laurenyu

benchmarks/tf_benchmarks/README.md

@@ -1,9 +1,9 @@
 # TensorFlow benchmarking scripts
 
-This folder contains the TF training scripts https://github.com/tensorflow/benchmarks/tree/master/scripts/tf_cnn_benchmarks.
+This folder contains a copy of [TensorFlow's `tf_cnn_benchmarks.py` script](https://github.com/tensorflow/benchmarks/blob/e3bd1370ba21b02c4d34340934ffb4941977d96f/scripts/tf_cnn_benchmarks/tf_cnn_benchmarks.py).
 
 ## Basic usage
-**execute_tensorflow_training.py train** uses SageMaker python sdk to start a training job. 
+**execute_tensorflow_training.py train** uses SageMaker python sdk to start a training job.
 
 ```bash
 ./execute_tensorflow_training.py train --help
@@ -26,7 +26,7 @@ Options:
   --help                          Show this message and exit.
 
 ```
-**execute_tensorflow_training.py generate_reports** generate benchmark reports. 
+**execute_tensorflow_training.py generate_reports** generate benchmark reports.
 
 ## Examples:
 

benchmarks/tf_benchmarks/benchmarks

@@ -11,7 +11,6 @@
 # distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF
 # ANY KIND, either express or implied. See the License for the specific
 # language governing permissions and limitations under the License.
-
 from __future__ import absolute_import
 
 import argparse
@@ -107,4 +106,4 @@ def create_hyperparameters(model_dir, script_args):
 
 if __name__ == '__main__':
     args, script_args = get_args()
-    main(args, script_args)
+    main(args, script_args)

benchmarks/tf_benchmarks/execute_tensorflow_training.py

@@ -0,0 +1,68 @@
+# Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License"). You
+# may not use this file except in compliance with the License. A copy of
+# the License is located at
+#
+#     http://aws.amazon.com/apache2.0/
+#
+# or in the "license" file accompanying this file. This file is
+# distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF
+# ANY KIND, either express or implied. See the License for the specific
+# language governing permissions and limitations under the License.
+"""Benchmark script for TensorFlow.
+
+Originally copied from:
+https://github.com/tensorflow/benchmarks/blob/e3bd1370ba21b02c4d34340934ffb4941977d96f/scripts/tf_cnn_benchmarks/tf_cnn_benchmarks.py
+"""
+from __future__ import absolute_import, division, print_function
+
+from absl import app
+from absl import flags as absl_flags
+import tensorflow.compat.v1 as tf
+
+import benchmark_cnn
+import cnn_util
+import flags
+import mlperf
+from cnn_util import log_fn
+
+
+flags.define_flags()
+for name in flags.param_specs.keys():
+    absl_flags.declare_key_flag(name)
+
+absl_flags.DEFINE_boolean(
+    'ml_perf_compliance_logging', False,
+    'Print logs required to be compliant with MLPerf. If set, must clone the '
+    'MLPerf training repo https://github.com/mlperf/training and add '
+    'https://github.com/mlperf/training/tree/master/compliance to the '
+    'PYTHONPATH')
+
+
+def main(positional_arguments):
+    # Command-line arguments like '--distortions False' are equivalent to
+    # '--distortions=True False', where False is a positional argument. To prevent
+    # this from silently running with distortions, we do not allow positional
+    # arguments.
+    assert len(positional_arguments) >= 1
+    if len(positional_arguments) > 1:
+        raise ValueError('Received unknown positional arguments: %s'
+                         % positional_arguments[1:])
+
+    params = benchmark_cnn.make_params_from_flags()
+    with mlperf.mlperf_logger(absl_flags.FLAGS.ml_perf_compliance_logging,
+                              params.model):
+        params = benchmark_cnn.setup(params)
+        bench = benchmark_cnn.BenchmarkCNN(params)
+
+    tfversion = cnn_util.tensorflow_version_tuple()
+    log_fn('TensorFlow:  %i.%i' % (tfversion[0], tfversion[1]))
+
+    bench.print_info()
+    bench.run()
+
+
+if __name__ == '__main__':
+    tf.disable_v2_behavior()
+    app.run(main)  # Raises error on invalid flags, unlike tf.app.run()