Merge pull request tensorflow#5253 from tfboyd/resnet_input_pipeline_fp16

tfboyd · web-flow · commit 7b3046768d2d · 2018-09-05T16:54:51.000-07:00
Move tf.cast to tf.float16 in input pipeline
diff --git a/official/resnet/cifar10_main.py b/official/resnet/cifar10_main.py
@@ -66,7 +66,7 @@ def get_filenames(is_training, data_dir):
     return [os.path.join(data_dir, 'test_batch.bin')]
 
 
-def parse_record(raw_record, is_training):
+def parse_record(raw_record, is_training, dtype):
   """Parse CIFAR-10 image and label from a raw record."""
   # Convert bytes to a vector of uint8 that is record_bytes long.
   record_vector = tf.decode_raw(raw_record, tf.uint8)
@@ -85,6 +85,7 @@ def parse_record(raw_record, is_training):
   image = tf.cast(tf.transpose(depth_major, [1, 2, 0]), tf.float32)
 
   image = preprocess_image(image, is_training)
+  image = tf.cast(image, dtype)
 
   return image, label
 
@@ -107,15 +108,17 @@ def preprocess_image(image, is_training):
   return image
 
 
-def input_fn(is_training, data_dir, batch_size, num_epochs=1, num_gpus=None):
-  """Input_fn using the tf.data input pipeline for CIFAR-10 dataset.
+def input_fn(is_training, data_dir, batch_size, num_epochs=1, num_gpus=None,
+             dtype=tf.float32):
+  """Input function which provides batches for train or eval.
 
   Args:
     is_training: A boolean denoting whether the input is for training.
     data_dir: The directory containing the input data.
     batch_size: The number of samples per batch.
     num_epochs: The number of epochs to repeat the dataset.
     num_gpus: The number of gpus used for training.
+    dtype: Data type to use for images/features
 
   Returns:
     A dataset that can be used for iteration.
@@ -131,7 +134,8 @@ def input_fn(is_training, data_dir, batch_size, num_epochs=1, num_gpus=None):
       parse_record_fn=parse_record,
       num_epochs=num_epochs,
       num_gpus=num_gpus,
-      examples_per_epoch=_NUM_IMAGES['train'] if is_training else None
+      examples_per_epoch=_NUM_IMAGES['train'] if is_training else None,
+      dtype=dtype
   )
 
 
diff --git a/official/resnet/cifar10_test.py b/official/resnet/cifar10_test.py
@@ -61,7 +61,7 @@ def test_dataset_input_fn(self):
     fake_dataset = tf.data.FixedLengthRecordDataset(
         filename, cifar10_main._RECORD_BYTES)  # pylint: disable=protected-access
     fake_dataset = fake_dataset.map(
-        lambda val: cifar10_main.parse_record(val, False))
+        lambda val: cifar10_main.parse_record(val, False, tf.float32))
     image, label = fake_dataset.make_one_shot_iterator().get_next()
 
     self.assertAllEqual(label.shape, ())
diff --git a/official/resnet/imagenet_main.py b/official/resnet/imagenet_main.py
@@ -129,7 +129,7 @@ def _parse_example_proto(example_serialized):
   return features['image/encoded'], label, bbox
 
 
-def parse_record(raw_record, is_training):
+def parse_record(raw_record, is_training, dtype):
   """Parses a record containing a training example of an image.
 
   The input record is parsed into a label and image, and the image is passed
@@ -139,6 +139,7 @@ def parse_record(raw_record, is_training):
     raw_record: scalar Tensor tf.string containing a serialized
       Example protocol buffer.
     is_training: A boolean denoting whether the input is for training.
+    dtype: data type to use for images/features.
 
   Returns:
     Tuple with processed image tensor and one-hot-encoded label tensor.
@@ -152,11 +153,13 @@ def parse_record(raw_record, is_training):
       output_width=_DEFAULT_IMAGE_SIZE,
       num_channels=_NUM_CHANNELS,
       is_training=is_training)
+  image = tf.cast(image, dtype)
 
   return image, label
 
 
-def input_fn(is_training, data_dir, batch_size, num_epochs=1, num_gpus=None):
+def input_fn(is_training, data_dir, batch_size, num_epochs=1, num_gpus=None,
+             dtype=tf.float32):
   """Input function which provides batches for train or eval.
 
   Args:
@@ -165,6 +168,7 @@ def input_fn(is_training, data_dir, batch_size, num_epochs=1, num_gpus=None):
     batch_size: The number of samples per batch.
     num_epochs: The number of epochs to repeat the dataset.
     num_gpus: The number of gpus used for training.
+    dtype: Data type to use for images/features
 
   Returns:
     A dataset that can be used for iteration.
@@ -192,7 +196,8 @@ def input_fn(is_training, data_dir, batch_size, num_epochs=1, num_gpus=None):
       parse_record_fn=parse_record,
       num_epochs=num_epochs,
       num_gpus=num_gpus,
-      examples_per_epoch=_NUM_IMAGES['train'] if is_training else None
+      examples_per_epoch=_NUM_IMAGES['train'] if is_training else None,
+      dtype=dtype
   )
 
 
diff --git a/official/resnet/resnet_run_loop.py b/official/resnet/resnet_run_loop.py
@@ -45,7 +45,7 @@
 ################################################################################
 def process_record_dataset(dataset, is_training, batch_size, shuffle_buffer,
                            parse_record_fn, num_epochs=1, num_gpus=None,
-                           examples_per_epoch=None):
+                           examples_per_epoch=None, dtype=tf.float32):
   """Given a Dataset with raw records, return an iterator over the records.
 
   Args:
@@ -60,6 +60,7 @@ def process_record_dataset(dataset, is_training, batch_size, shuffle_buffer,
     num_epochs: The number of epochs to repeat the dataset.
     num_gpus: The number of gpus used for training.
     examples_per_epoch: The number of examples in an epoch.
+    dtype: Data type to use for images/features.
 
   Returns:
     Dataset of (image, label) pairs ready for iteration.
@@ -92,7 +93,7 @@ def process_record_dataset(dataset, is_training, batch_size, shuffle_buffer,
   # batch_size is almost always much greater than the number of CPU cores.
   dataset = dataset.apply(
       tf.contrib.data.map_and_batch(
-          lambda value: parse_record_fn(value, is_training),
+          lambda value: parse_record_fn(value, is_training, dtype),
           batch_size=batch_size,
           num_parallel_batches=1,
           drop_remainder=False))
@@ -248,8 +249,8 @@ def resnet_model_fn(features, labels, mode, model_class,
 
   # Generate a summary node for the images
   tf.summary.image('images', features, max_outputs=6)
-  # TODO(tobyboyd): Add cast as part of input pipeline on cpu and remove.
-  features = tf.cast(features, dtype)
+  # Checks that features/images have same data type being used for calculations.
+  assert features.dtype == dtype
 
   model = model_class(resnet_size, data_format, resnet_version=resnet_version,
                       dtype=dtype)
@@ -454,14 +455,16 @@ def input_fn_train(num_epochs):
         batch_size=distribution_utils.per_device_batch_size(
             flags_obj.batch_size, flags_core.get_num_gpus(flags_obj)),
         num_epochs=num_epochs,
-        num_gpus=flags_core.get_num_gpus(flags_obj))
+        num_gpus=flags_core.get_num_gpus(flags_obj),
+        dtype=flags_core.get_tf_dtype(flags_obj))
 
   def input_fn_eval():
     return input_function(
         is_training=False, data_dir=flags_obj.data_dir,
         batch_size=distribution_utils.per_device_batch_size(
             flags_obj.batch_size, flags_core.get_num_gpus(flags_obj)),
-        num_epochs=1)
+        num_epochs=1,
+        dtype=flags_core.get_tf_dtype(flags_obj))
 
   if flags_obj.eval_only or not flags_obj.train_epochs:
     # If --eval_only is set, perform a single loop with zero train epochs.
@@ -533,7 +536,7 @@ def define_resnet_flags(resnet_size_choices=None):
           'If not None initialize all the network except the final layer with '
           'these values'))
   flags.DEFINE_boolean(
-      name="eval_only", default=False,
+      name='eval_only', default=False,
       help=flags_core.help_wrap('Skip training and only perform evaluation on '
                                 'the latest checkpoint.'))
 
