def build_model(self):
"""Find the model and build the graph."""
# Convert feature_names and feature_sizes to lists of values.
feature_names, feature_sizes = utils.GetListOfFeatureNamesAndSizes(
FLAGS.feature_names, FLAGS.feature_sizes)
if FLAGS.distillation_features:
print "distillation readers"
if FLAGS.frame_features:
reader = readers.YT8MFrameDistillationFeatureReader(
feature_names=feature_names, feature_sizes=feature_sizes)
else:
reader = readers.YT8MAggregatedDistillationFeatureReader(
feature_names=feature_names, feature_sizes=feature_sizes)
else:
if FLAGS.frame_features:
reader = readers.YT8MFrameFeatureReader(
feature_names=feature_names, feature_sizes=feature_sizes)
else:
reader = readers.YT8MAggregatedFeatureReader(
feature_names=feature_names, feature_sizes=feature_sizes)
# Find the model.
model = find_class_by_name(FLAGS.model,
[frame_level_models, video_level_models])()
label_loss_fn = find_class_by_name(FLAGS.label_loss, [losses])()
optimizer_class = find_class_by_name(FLAGS.optimizer, [tf.train])
transformer_class = find_class_by_name(FLAGS.feature_transformer,
[feature_transform])
augmenter_class = find_class_by_name(FLAGS.data_augmenter,
[data_augmentation])
build_graph(
reader=reader,
model=model,
optimizer_class=optimizer_class,
augmenter_class=augmenter_class,
transformer_class=transformer_class,
clip_gradient_norm=FLAGS.clip_gradient_norm,
train_data_pattern=FLAGS.train_data_pattern,
label_loss_fn=label_loss_fn,
base_learning_rate=FLAGS.base_learning_rate,
learning_rate_decay=FLAGS.learning_rate_decay,
learning_rate_decay_examples=FLAGS.learning_rate_decay_examples,
regularization_penalty=FLAGS.regularization_penalty,
num_readers=FLAGS.num_readers,
batch_size=FLAGS.batch_size,
num_epochs=FLAGS.num_epochs)
logging.info("%s: Built graph.", task_as_string(self.task))
return tf.train.Saver(
max_to_keep=3,
keep_checkpoint_every_n_hours=FLAGS.keep_checkpoint_every_n_hours)
def get_reader():
# Convert feature_names and feature_sizes to lists of values.
feature_names, feature_sizes = utils.GetListOfFeatureNamesAndSizes(
FLAGS.feature_names, FLAGS.feature_sizes)
if FLAGS.frame_features:
reader = readers.YT8MFrameDistillationFeatureReader(
feature_names=feature_names, feature_sizes=feature_sizes)
else:
reader = readers.YT8MAggregatedFeatureReader(
feature_names=feature_names, feature_sizes=feature_sizes)
return reader