Add benchmarking script (aws#86)

* Add benchmarking script

Author: mvsusp

benchmarks/README.md

@@ -0,0 +1,35 @@
+# TensorFlow benchmarking scripts
+
+This folder contains the TF training scripts https://github.com/tensorflow/benchmarks/tree/master/scripts/tf_cnn_benchmarks.
+
+## Basic usage
+**execute_tensorflow_training.py** uses SageMaker python sdk to start a training job. It takes the following parameters:
+
+- role: SageMaker role used for training
+- region: SageMaker region
+- py-versions: py2 or py3 or "py2, py3"
+- instance-types: A list of SageMaker instance types, for example 'ml.p2.xlarge, ml.c4.xlarge'. Use 'local' for local mode training.
+- checkpoint-path: The S3 location where the model checkpoints and tensorboard events are saved after training 
+
+Any unknown arguments will be passed to the training script as additional arguments.
+
+## Examples:
+
+```bash
+./execute_tensorflow_training.py -t local -r SageMakerRole --instance-type local  --num_epochs 1 --wait
+
+./execute_tensorflow_training.py -t local -r SageMakerRole --instance-type ml.c4.xlarge, ml.c5.xlarge  --model resnet50
+
+```
+
+## Using other models, datasets and benchmarks configurations
+```python tf_cnn_benchmarks/tf_cnn_benchmarks.py --help``` shows all the options that the script has.
+
+
+## Tensorboard events and checkpoints
+
+Tensorboard events are being saved to the S3 location defined by the hyperparameter checkpoint_path during training. That location can be overwritten by setting the script argument ```checkpoint-path```:
+
+```bash
+python execute_tensorflow_training.py ... --checkpoint-path s3://my/bucket/output/data
+```

benchmarks/execute_tensorflow_training.py

@@ -0,0 +1,98 @@
+#!/usr/bin/env python
+
+from __future__ import absolute_import
+
+import argparse
+import itertools
+import os
+
+from sagemaker import Session
+from sagemaker.estimator import Framework
+from sagemaker.tensorflow import TensorFlow
+
+default_bucket = Session().default_bucket
+dir_path = os.path.dirname(os.path.realpath(__file__))
+
+_DEFAULT_HYPERPARAMETERS = {
+    'batch_size':           32,
+    'model':                'resnet32',
+    'num_epochs':           10,
+    'data_format':          'NHWC',
+    'summary_verbosity':    1,
+    'save_summaries_steps': 10,
+    'data_name':            'cifar10'
+}
+
+
+class ScriptModeTensorFlow(Framework):
+    """This class is temporary until the final version of Script Mode is released.
+    """
+
+    __framework_name__ = "tensorflow-scriptmode-beta"
+
+    create_model = TensorFlow.create_model
+
+    def __init__(self, py_version='py3', **kwargs):
+        super(ScriptModeTensorFlow, self).__init__(**kwargs)
+        self.py_version = py_version
+        self.image_name = None
+        self.framework_version = '1.10.0'
+
+
+def get_args():
+    parser = argparse.ArgumentParser()
+    parser.add_argument('-t', '--instance-types', nargs='+', help='<Required> Set flag', required=True)
+    parser.add_argument('-r', '--role', required=True)
+    parser.add_argument('-w', '--wait', action='store_true')
+    parser.add_argument('--region', default='us-west-2')
+    parser.add_argument('--py-versions', nargs='+', help='<Required> Set flag', default=['py3'])
+    parser.add_argument('--checkpoint-path',
+                        default=os.path.join(default_bucket(), 'benchmarks', 'checkpoints'),
+                        help='The S3 location where the model checkpoints and tensorboard events are saved after training')
+
+    return parser.parse_known_args()
+
+
+def main(args, script_args):
+    for instance_type, py_version in itertools.product(args.instance_types, args.py_versions):
+        base_name = '%s-%s-%s' % (py_version, instance_type[3:5], instance_type[6:])
+        model_dir = os.path.join(args.checkpoint_path, base_name)
+
+        job_hps = create_hyperparameters(model_dir, script_args)
+
+        print('hyperparameters:')
+        print(job_hps)
+
+        estimator = ScriptModeTensorFlow(
+            entry_point='tf_cnn_benchmarks.py',
+            role='SageMakerRole',
+            source_dir=os.path.join(dir_path, 'tf_cnn_benchmarks'),
+            base_job_name=base_name,
+            train_instance_count=1,
+            hyperparameters=job_hps,
+            train_instance_type=instance_type,
+        )
+
+        input_dir = 's3://sagemaker-sample-data-%s/spark/mnist/train/' % args.region
+        estimator.fit({'train': input_dir}, wait=args.wait)
+
+    print("To use TensorBoard, execute the following command:")
+    cmd = 'S3_USE_HTTPS=0 S3_VERIFY_SSL=0  AWS_REGION=%s tensorboard --host localhost --port 6006 --logdir %s'
+    print(cmd % (args.region, args.checkpoint_path))
+
+
+def create_hyperparameters(model_dir, script_args):
+    job_hps = _DEFAULT_HYPERPARAMETERS.copy()
+
+    job_hps.update({'train_dir': model_dir, 'eval_dir': model_dir})
+
+    script_arg_keys_without_dashes = [key[2:] if key.startswith('--') else key[1:] for key in script_args[::2]]
+    script_arg_values = script_args[1::2]
+    job_hps.update(dict(zip(script_arg_keys_without_dashes, script_arg_values)))
+
+    return job_hps
+
+
+if __name__ == '__main__':
+    args, script_args = get_args()
+    main(args, script_args)

benchmarks/tf_cnn_benchmarks/README.md

@@ -0,0 +1,89 @@
+# tf_cnn_benchmarks: High performance benchmarks
+
+tf_cnn_benchmarks contains implementations of several popular convolutional
+models, and is designed to be as fast as possible. tf_cnn_benchmarks supports
+both running on a single machine or running in distributed mode across multiple
+hosts. See the [High-Performance models
+guide](https://www.tensorflow.org/performance/performance_models) for more
+information.
+
+These models utilize many of the strategies in the [TensorFlow Performance
+Guide](https://www.tensorflow.org/performance/performance_guide). Benchmark
+results can be found [here](https://www.tensorflow.org/performance/benchmarks).
+
+These models are designed for performance. For models that have clean and
+easy-to-read implementations, see the [TensorFlow Official
+Models](https://github.com/tensorflow/models/tree/master/official).
+
+## Getting Started
+
+To run ResNet50 with synthetic data without distortions with a single GPU, run
+
+```
+python tf_cnn_benchmarks.py --num_gpus=1 --batch_size=32 --model=resnet50 --variable_update=parameter_server
+```
+
+Note that the master branch of tf_cnn_benchmarks requires the latest nightly
+version of TensorFlow. You can install the nightly version by running `pip
+install tf-nightly-gpu` in a clean environment, or by installing TensorFlow from
+source. We sometimes will create a branch of tf_cnn_benchmarks, in the form of
+cnn_tf_vX.Y_compatible, that is compatible with TensorFlow version X.Y For
+example, branch
+[cnn_tf_v1.9_compatible](https://github.com/tensorflow/benchmarks/tree/cnn_tf_v1.9_compatible/scripts/tf_cnn_benchmarks)
+works with TensorFlow 1.9.
+
+Some important flags are
+
+*   model: Model to use, e.g. resnet50, inception3, vgg16, and alexnet.
+*   num_gpus: Number of GPUs to use.
+*   data_dir: Path to data to process. If not set, synthetic data is used. To
+    use Imagenet data use these
+    [instructions](https://github.com/tensorflow/models/tree/master/research/inception#getting-started)
+    as a starting point.
+*   batch_size: Batch size for each GPU.
+*   variable_update: The method for managing variables: parameter_server
+    ,replicated, distributed_replicated, independent
+*   local_parameter_device: Device to use as parameter server: cpu or gpu.
+
+To see the full list of flags, run `python tf_cnn_benchmarks.py --help`.
+
+To run ResNet50 with real data with 8 GPUs, run:
+
+```
+python tf_cnn_benchmarks.py --data_format=NCHW --batch_size=256 \
+--model=resnet50 --optimizer=momentum --variable_update=replicated \
+--nodistortions --gradient_repacking=8 --num_gpus=8 \
+--num_epochs=90 --weight_decay=1e-4 --data_dir=${DATA_DIR} --use_fp16 \
+--train_dir=${CKPT_DIR}
+```
+This will train a ResNet-50 model on ImageNet with 2048 batch size on 8
+GPUs. The model should train to around 76% accuracy.
+
+## Running the tests
+
+To run the tests, run
+
+```bash
+pip install portpicker
+python run_tests.py && python run_tests.py --run_distributed_tests
+```
+
+Note the tests require portpicker.
+
+The command above runs a subset of tests that is both fast and fairly
+comprehensive. Alternatively, all the tests can be run, but this will take a
+long time:
+
+```bash
+python run_tests.py --full_tests && python run_tests.py --full_tests --run_distributed_tests
+```
+
+We will run all tests on every PR before merging them, so it is not necessary
+to pass `--full_tests` when running tests yourself.
+
+To run an individual test, such as method `testParameterServer` of test class
+`TfCnnBenchmarksTest` of module `benchmark_cnn_test`, run
+
+```bash
+python -m unittest -v benchmark_cnn_test.TfCnnBenchmarksTest.testParameterServer
+```