Use FLAGS in main functions only + Updates to shuffling (#2601)

official/mnist/convert_to_records.py

@@ -50,11 +50,11 @@ def _bytes_feature(value):
   return tf.train.Feature(bytes_list=tf.train.BytesList(value=[value]))
 
 
-def convert_to(data_set, name):
+def convert_to(dataset, name, directory):
   """Converts a dataset to TFRecords."""
-  images = data_set.images
-  labels = data_set.labels
-  num_examples = data_set.num_examples
+  images = dataset.images
+  labels = dataset.labels
+  num_examples = dataset.num_examples
 
   if images.shape[0] != num_examples:
     raise ValueError('Images size %d does not match label size %d.' %
@@ -63,7 +63,7 @@ def convert_to(data_set, name):
   cols = images.shape[2]
   depth = images.shape[3]
 
-  filename = os.path.join(FLAGS.directory, name + '.tfrecords')
+  filename = os.path.join(directory, name + '.tfrecords')
   print('Writing', filename)
   writer = tf.python_io.TFRecordWriter(filename)
   for index in range(num_examples):
@@ -80,15 +80,15 @@ def convert_to(data_set, name):
 
 def main(unused_argv):
   # Get the data.
-  data_sets = mnist.read_data_sets(FLAGS.directory,
+  datasets = mnist.read_data_sets(FLAGS.directory,
                                    dtype=tf.uint8,
                                    reshape=False,
                                    validation_size=FLAGS.validation_size)
 
   # Convert to Examples and write the result to TFRecords.
-  convert_to(data_sets.train, 'train')
-  convert_to(data_sets.validation, 'validation')
-  convert_to(data_sets.test, 'test')
+  convert_to(datasets.train, 'train', FLAGS.directory)
+  convert_to(datasets.validation, 'validation', FLAGS.directory)
+  convert_to(datasets.test, 'test', FLAGS.directory)
 
 
 if __name__ == '__main__':

official/mnist/mnist.py

@@ -52,7 +52,7 @@ _NUM_IMAGES = {
 }
 
 
-def input_fn(mode, batch_size=1):
+def input_fn(is_training, filename, batch_size=1, num_epochs=1):
   """A simple input_fn using the contrib.data input pipeline."""
 
   def example_parser(serialized_example):
@@ -71,21 +71,15 @@ def input_fn(mode, batch_size=1):
     label = tf.cast(features['label'], tf.int32)
     return image, tf.one_hot(label, 10)
 
-  if mode == tf.estimator.ModeKeys.TRAIN:
-    tfrecords_file = os.path.join(FLAGS.data_dir, 'train.tfrecords')
-  else:
-    assert mode == tf.estimator.ModeKeys.EVAL, 'invalid mode'
-    tfrecords_file = os.path.join(FLAGS.data_dir, 'test.tfrecords')
+  dataset = tf.contrib.data.TFRecordDataset([filename])
 
-  assert tf.gfile.Exists(tfrecords_file), (
-      'Run convert_to_records.py first to convert the MNIST data to TFRecord '
-      'file format.')
+  if is_training:
+    # When choosing shuffle buffer sizes, larger sizes result in better
+    # randomness, while smaller sizes have better performance. Because MNIST is
+    # a small dataset, we can easily shuffle the full epoch.
+    dataset = dataset.shuffle(buffer_size=_NUM_IMAGES['train'])
 
-  dataset = tf.contrib.data.TFRecordDataset([tfrecords_file])
-
-  # For training, repeat the dataset forever
-  if mode == tf.estimator.ModeKeys.TRAIN:
-    dataset = dataset.repeat()
+  dataset = dataset.repeat(num_epochs)
 
   # Map example_parser over dataset, and batch results by up to batch_size
   dataset = dataset.map(
@@ -96,13 +90,12 @@ def input_fn(mode, batch_size=1):
   return images, labels
 
 
-def mnist_model(inputs, mode):
+def mnist_model(inputs, mode, data_format):
   """Takes the MNIST inputs and mode and outputs a tensor of logits."""
   # Input Layer
   # Reshape X to 4-D tensor: [batch_size, width, height, channels]
   # MNIST images are 28x28 pixels, and have one color channel
   inputs = tf.reshape(inputs, [-1, 28, 28, 1])
-  data_format = FLAGS.data_format
 
   if data_format is None:
     # When running on GPU, transpose the data from channels_last (NHWC) to
@@ -177,9 +170,9 @@ def mnist_model(inputs, mode):
   return logits
 
 
-def mnist_model_fn(features, labels, mode):
+def mnist_model_fn(features, labels, mode, params):
   """Model function for MNIST."""
-  logits = mnist_model(features, mode)
+  logits = mnist_model(features, mode, params['data_format'])
 
   predictions = {
       'classes': tf.argmax(input=logits, axis=1),
@@ -215,30 +208,36 @@ def mnist_model_fn(features, labels, mode):
 
 
 def main(unused_argv):
+  # Make sure that training and testing data have been converted.
+  train_file = os.path.join(FLAGS.data_dir, 'train.tfrecords')
+  test_file = os.path.join(FLAGS.data_dir, 'test.tfrecords')
+  assert (tf.gfile.Exists(train_file) and tf.gfile.Exists(test_file)), (
+      'Run convert_to_records.py first to convert the MNIST data to TFRecord '
+      'file format.')
+
   # Create the Estimator
   mnist_classifier = tf.estimator.Estimator(
-      model_fn=mnist_model_fn, model_dir=FLAGS.model_dir)
+      model_fn=mnist_model_fn, model_dir=FLAGS.model_dir,
+      params={'data_format': FLAGS.data_format})
 
-  # Train the model
+  # Set up training hook that logs the training accuracy every 100 steps.
   tensors_to_log = {
       'train_accuracy': 'train_accuracy'
   }
-
   logging_hook = tf.train.LoggingTensorHook(
       tensors=tensors_to_log, every_n_iter=100)
 
-  batches_per_epoch = _NUM_IMAGES['train'] / FLAGS.batch_size
-
+  # Train the model
   mnist_classifier.train(
-      input_fn=lambda: input_fn(tf.estimator.ModeKeys.TRAIN, FLAGS.batch_size),
-      steps=FLAGS.train_epochs * batches_per_epoch,
+      input_fn=lambda: input_fn(
+          True, train_file, FLAGS.batch_size, FLAGS.train_epochs),
       hooks=[logging_hook])
 
   # Evaluate the model and print results
   eval_results = mnist_classifier.evaluate(
-      input_fn=lambda: input_fn(tf.estimator.ModeKeys.EVAL))
+      input_fn=lambda: input_fn(False, test_file, FLAGS.batch_size))
   print()
-  print('Evaluation results:\n    %s' % eval_results)
+  print('Evaluation results:\n\t%s' % eval_results)
 
 
 if __name__ == '__main__':

official/mnist/mnist_test.py

@@ -34,7 +34,8 @@ class BaseTest(tf.test.TestCase):
   def mnist_model_fn_helper(self, mode):
     features, labels = self.input_fn()
     image_count = features.shape[0]
-    spec = mnist.mnist_model_fn(features, labels, mode)
+    spec = mnist.mnist_model_fn(
+        features, labels, mode, {'data_format': 'channels_last'})
 
     predictions = spec.predictions
     self.assertAllEqual(predictions['probabilities'].shape, (image_count, 10))
@@ -65,5 +66,4 @@ class BaseTest(tf.test.TestCase):
 
 
 if __name__ == '__main__':
-  mnist.FLAGS = mnist.parser.parse_args()
   tf.test.main()

official/resnet/cifar10_main.py

@@ -71,6 +71,8 @@ _NUM_IMAGES = {
     'validation': 10000,
 }
 
+_SHUFFLE_BUFFER = 20000
+
 
 def record_dataset(filenames):
   """Returns an input pipeline Dataset from `filenames`."""
@@ -78,9 +80,9 @@ def record_dataset(filenames):
   return tf.contrib.data.FixedLengthRecordDataset(filenames, record_bytes)
 
 
-def get_filenames(is_training):
+def get_filenames(is_training, data_dir):
   """Returns a list of filenames."""
-  data_dir = os.path.join(FLAGS.data_dir, 'cifar-10-batches-bin')
+  data_dir = os.path.join(data_dir, 'cifar-10-batches-bin')
 
   assert os.path.exists(data_dir), (
       'Run cifar10_download_and_extract.py first to download and extract the '
@@ -135,7 +137,7 @@ def train_preprocess_fn(image, label):
   return image, label
 
 
-def input_fn(is_training, num_epochs=1):
+def input_fn(is_training, data_dir, batch_size, num_epochs=1):
   """Input_fn using the contrib.data input pipeline for CIFAR-10 dataset.
 
   Args:
@@ -145,42 +147,41 @@ def input_fn(is_training, num_epochs=1):
   Returns:
     A tuple of images and labels.
   """
-  dataset = record_dataset(get_filenames(is_training))
+  dataset = record_dataset(get_filenames(is_training, data_dir))
   dataset = dataset.map(dataset_parser, num_threads=1,
-                        output_buffer_size=2 * FLAGS.batch_size)
+                        output_buffer_size=2 * batch_size)
 
   # For training, preprocess the image and shuffle.
   if is_training:
     dataset = dataset.map(train_preprocess_fn, num_threads=1,
-                          output_buffer_size=2 * FLAGS.batch_size)
+                          output_buffer_size=2 * batch_size)
 
-    # Ensure that the capacity is sufficiently large to provide good random
-    # shuffling.
-    buffer_size = int(0.4 * _NUM_IMAGES['train'])
-    dataset = dataset.shuffle(buffer_size=buffer_size)
+    # When choosing shuffle buffer sizes, larger sizes result in better
+    # randomness, while smaller sizes have better performance.
+    dataset = dataset.shuffle(buffer_size=_SHUFFLE_BUFFER)
 
   # Subtract off the mean and divide by the variance of the pixels.
   dataset = dataset.map(
       lambda image, label: (tf.image.per_image_standardization(image), label),
       num_threads=1,
-      output_buffer_size=2 * FLAGS.batch_size)
+      output_buffer_size=2 * batch_size)
 
   dataset = dataset.repeat(num_epochs)
 
   # Batch results by up to batch_size, and then fetch the tuple from the
   # iterator.
-  iterator = dataset.batch(FLAGS.batch_size).make_one_shot_iterator()
+  iterator = dataset.batch(batch_size).make_one_shot_iterator()
   images, labels = iterator.get_next()
 
   return images, labels
 
 
-def cifar10_model_fn(features, labels, mode):
+def cifar10_model_fn(features, labels, mode, params):
   """Model function for CIFAR-10."""
   tf.summary.image('images', features, max_outputs=6)
 
   network = resnet_model.cifar10_resnet_v2_generator(
-      FLAGS.resnet_size, _NUM_CLASSES, FLAGS.data_format)
+      params['resnet_size'], _NUM_CLASSES, params['data_format'])
 
   inputs = tf.reshape(features, [-1, _HEIGHT, _WIDTH, _DEPTH])
   logits = network(inputs, mode == tf.estimator.ModeKeys.TRAIN)
@@ -208,8 +209,8 @@ def cifar10_model_fn(features, labels, mode):
   if mode == tf.estimator.ModeKeys.TRAIN:
     # Scale the learning rate linearly with the batch size. When the batch size
     # is 128, the learning rate should be 0.1.
-    initial_learning_rate = 0.1 * FLAGS.batch_size / 128
-    batches_per_epoch = _NUM_IMAGES['train'] / FLAGS.batch_size
+    initial_learning_rate = 0.1 * params['batch_size'] / 128
+    batches_per_epoch = _NUM_IMAGES['train'] / params['batch_size']
     global_step = tf.train.get_or_create_global_step()
 
     # Multiply the learning rate by 0.1 at 100, 150, and 200 epochs.
@@ -256,7 +257,12 @@ def main(unused_argv):
   # Set up a RunConfig to only save checkpoints once per training cycle.
   run_config = tf.estimator.RunConfig().replace(save_checkpoints_secs=1e9)
   cifar_classifier = tf.estimator.Estimator(
-      model_fn=cifar10_model_fn, model_dir=FLAGS.model_dir, config=run_config)
+      model_fn=cifar10_model_fn, model_dir=FLAGS.model_dir, config=run_config,
+      params={
+          'resnet_size': FLAGS.resnet_size,
+          'data_format': FLAGS.data_format,
+          'batch_size': FLAGS.batch_size,
+      })
 
   for _ in range(FLAGS.train_epochs // FLAGS.epochs_per_eval):
     tensors_to_log = {
@@ -270,12 +276,12 @@ def main(unused_argv):
 
     cifar_classifier.train(
         input_fn=lambda: input_fn(
-            is_training=True, num_epochs=FLAGS.epochs_per_eval),
+            True, FLAGS.data_dir, FLAGS.batch_size, FLAGS.epochs_per_eval),
         hooks=[logging_hook])
 
     # Evaluate the model and print results
     eval_results = cifar_classifier.evaluate(
-        input_fn=lambda: input_fn(is_training=False))
+        input_fn=lambda: input_fn(False, FLAGS.data_dir, FLAGS.batch_size))
     print(eval_results)
 
 

official/resnet/cifar10_test.py

@@ -26,6 +26,8 @@ import cifar10_main
 
 tf.logging.set_verbosity(tf.logging.ERROR)
 
+_BATCH_SIZE = 128
+
 
 class BaseTest(tf.test.TestCase):
 
@@ -58,20 +60,25 @@ class BaseTest(tf.test.TestCase):
           self.assertAllEqual(pixel, np.array([0, 1, 2]))
 
   def input_fn(self):
-    features = tf.random_uniform([FLAGS.batch_size, 32, 32, 3])
+    features = tf.random_uniform([_BATCH_SIZE, 32, 32, 3])
     labels = tf.random_uniform(
-        [FLAGS.batch_size], maxval=9, dtype=tf.int32)
+        [_BATCH_SIZE], maxval=9, dtype=tf.int32)
     return features, tf.one_hot(labels, 10)
 
   def cifar10_model_fn_helper(self, mode):
     features, labels = self.input_fn()
-    spec = cifar10_main.cifar10_model_fn(features, labels, mode)
+    spec = cifar10_main.cifar10_model_fn(
+        features, labels, mode, {
+            'resnet_size': 32,
+            'data_format': 'channels_last',
+            'batch_size': _BATCH_SIZE,
+        })
 
     predictions = spec.predictions
     self.assertAllEqual(predictions['probabilities'].shape,
-                        (FLAGS.batch_size, 10))
+                        (_BATCH_SIZE, 10))
     self.assertEqual(predictions['probabilities'].dtype, tf.float32)
-    self.assertAllEqual(predictions['classes'].shape, (FLAGS.batch_size,))
+    self.assertAllEqual(predictions['classes'].shape, (_BATCH_SIZE,))
     self.assertEqual(predictions['classes'].dtype, tf.int64)
 
     if mode != tf.estimator.ModeKeys.PREDICT:
@@ -97,6 +104,4 @@ class BaseTest(tf.test.TestCase):
 
 
 if __name__ == '__main__':
-  cifar10_main.FLAGS = cifar10_main.parser.parse_args()
-  FLAGS = cifar10_main.FLAGS
   tf.test.main()

official/resnet/imagenet_main.py

@@ -73,16 +73,18 @@ _NUM_IMAGES = {
     'validation': 50000,
 }
 
+_SHUFFLE_BUFFER = 1500
 
-def filenames(is_training):
+
+def filenames(is_training, data_dir):
   """Return filenames for dataset."""
   if is_training:
     return [
-        os.path.join(FLAGS.data_dir, 'train-%05d-of-01024' % i)
+        os.path.join(data_dir, 'train-%05d-of-01024' % i)
         for i in range(0, 1024)]
   else:
     return [
-        os.path.join(FLAGS.data_dir, 'validation-%05d-of-00128' % i)
+        os.path.join(data_dir, 'validation-%05d-of-00128' % i)
         for i in range(0, 128)]
 
 
@@ -129,9 +131,11 @@ def dataset_parser(value, is_training):
   return image, tf.one_hot(label, _LABEL_CLASSES)
 
 
-def input_fn(is_training, num_epochs=1):
+def input_fn(is_training, data_dir, batch_size, num_epochs=1):
   """Input function which provides batches for train or eval."""
-  dataset = tf.contrib.data.Dataset.from_tensor_slices(filenames(is_training))
+  dataset = tf.contrib.data.Dataset.from_tensor_slices(
+      filenames(is_training, data_dir))
+
   if is_training:
     dataset = dataset.shuffle(buffer_size=1024)
   dataset = dataset.flat_map(tf.contrib.data.TFRecordDataset)
@@ -141,23 +145,24 @@ def input_fn(is_training, num_epochs=1):
 
   dataset = dataset.map(lambda value: dataset_parser(value, is_training),
                         num_threads=5,
-                        output_buffer_size=FLAGS.batch_size)
+                        output_buffer_size=batch_size)
 
   if is_training:
-    buffer_size = 1250 + 2 * FLAGS.batch_size
-    dataset = dataset.shuffle(buffer_size=buffer_size)
+    # When choosing shuffle buffer sizes, larger sizes result in better
+    # randomness, while smaller sizes have better performance.
+    dataset = dataset.shuffle(buffer_size=_SHUFFLE_BUFFER)
 
-  iterator = dataset.batch(FLAGS.batch_size).make_one_shot_iterator()
+  iterator = dataset.batch(batch_size).make_one_shot_iterator()
   images, labels = iterator.get_next()
   return images, labels
 
 
-def resnet_model_fn(features, labels, mode):
+def resnet_model_fn(features, labels, mode, params):
   """Our model_fn for ResNet to be used with our Estimator."""
   tf.summary.image('images', features, max_outputs=6)
 
   network = resnet_model.imagenet_resnet_v2(
-      FLAGS.resnet_size, _LABEL_CLASSES, FLAGS.data_format)
+      params['resnet_size'], _LABEL_CLASSES, params['data_format'])
   logits = network(
       inputs=features, is_training=(mode == tf.estimator.ModeKeys.TRAIN))
 
@@ -185,8 +190,8 @@ def resnet_model_fn(features, labels, mode):
   if mode == tf.estimator.ModeKeys.TRAIN:
     # Scale the learning rate linearly with the batch size. When the batch size is
     # 256, the learning rate should be 0.1.
-    initial_learning_rate = 0.1 * FLAGS.batch_size / 256
-    batches_per_epoch = _NUM_IMAGES['train'] / FLAGS.batch_size
+    initial_learning_rate = 0.1 * params['batch_size'] / 256
+    batches_per_epoch = _NUM_IMAGES['train'] / params['batch_size']
     global_step = tf.train.get_or_create_global_step()
 
     # Multiply the learning rate by 0.1 at 30, 60, 80, and 90 epochs.
@@ -235,7 +240,12 @@ def main(unused_argv):
   # Set up a RunConfig to only save checkpoints once per training cycle.
   run_config = tf.estimator.RunConfig().replace(save_checkpoints_secs=1e9)
   resnet_classifier = tf.estimator.Estimator(
-      model_fn=resnet_model_fn, model_dir=FLAGS.model_dir, config=run_config)
+      model_fn=resnet_model_fn, model_dir=FLAGS.model_dir, config=run_config,
+      params={
+          'resnet_size': FLAGS.resnet_size,
+          'data_format': FLAGS.data_format,
+          'batch_size': FLAGS.batch_size,
+      })
 
   for _ in range(FLAGS.train_epochs // FLAGS.epochs_per_eval):
     tensors_to_log = {
@@ -250,12 +260,12 @@ def main(unused_argv):
     print('Starting a training cycle.')
     resnet_classifier.train(
         input_fn=lambda: input_fn(
-            is_training=True, num_epochs=FLAGS.epochs_per_eval),
+            True, FLAGS.data_dir, FLAGS.batch_size, FLAGS.epochs_per_eval),
         hooks=[logging_hook])
 
     print('Starting to evaluate.')
     eval_results = resnet_classifier.evaluate(
-        input_fn=lambda: input_fn(is_training=False))
+        input_fn=lambda: input_fn(False, FLAGS.data_dir, FLAGS.batch_size))
     print(eval_results)
 
 

official/resnet/imagenet_test.py

@@ -26,6 +26,7 @@ import resnet_model
 
 tf.logging.set_verbosity(tf.logging.ERROR)
 
+_BATCH_SIZE = 32
 _LABEL_CLASSES = 1001
 
 
@@ -125,10 +126,10 @@ class BaseTest(tf.test.TestCase):
 
   def input_fn(self):
     """Provides random features and labels."""
-    features = tf.random_uniform([FLAGS.batch_size, 224, 224, 3])
+    features = tf.random_uniform([_BATCH_SIZE, 224, 224, 3])
     labels = tf.one_hot(
         tf.random_uniform(
-            [FLAGS.batch_size], maxval=_LABEL_CLASSES - 1,
+            [_BATCH_SIZE], maxval=_LABEL_CLASSES - 1,
             dtype=tf.int32),
         _LABEL_CLASSES)
 
@@ -139,13 +140,18 @@ class BaseTest(tf.test.TestCase):
     tf.train.create_global_step()
 
     features, labels = self.input_fn()
-    spec = imagenet_main.resnet_model_fn(features, labels, mode)
+    spec = imagenet_main.resnet_model_fn(
+        features, labels, mode, {
+            'resnet_size': 50,
+            'data_format': 'channels_last',
+            'batch_size': _BATCH_SIZE,
+        })
 
     predictions = spec.predictions
     self.assertAllEqual(predictions['probabilities'].shape,
-                        (FLAGS.batch_size, _LABEL_CLASSES))
+                        (_BATCH_SIZE, _LABEL_CLASSES))
     self.assertEqual(predictions['probabilities'].dtype, tf.float32)
-    self.assertAllEqual(predictions['classes'].shape, (FLAGS.batch_size,))
+    self.assertAllEqual(predictions['classes'].shape, (_BATCH_SIZE,))
     self.assertEqual(predictions['classes'].dtype, tf.int64)
 
     if mode != tf.estimator.ModeKeys.PREDICT:
@@ -171,6 +177,4 @@ class BaseTest(tf.test.TestCase):
 
 
 if __name__ == '__main__':
-  imagenet_main.FLAGS = imagenet_main.parser.parse_args()
-  FLAGS = imagenet_main.FLAGS
   tf.test.main()