def test_pooling_method(self):
model1 = mobilenet_v2.create_mobilenet_v2(
input_shape=(224, 224, 3), pooling='avg', num_classes=1000)
model2 = mobilenet_v2.create_mobilenet_v2(
input_shape=(224, 224, 3), pooling='max', num_classes=1000)
self.assertIsInstance(model1, tf.keras.Model)
self.assertIsInstance(model2, tf.keras.Model)
self.assertEqual(model1.count_params(), model2.count_params())
def test_num_groups(self):
model1 = mobilenet_v2.create_mobilenet_v2(
input_shape=(224, 224, 3), num_classes=1000)
model2 = mobilenet_v2.create_mobilenet_v2(
input_shape=(224, 224, 3), num_groups=4, num_classes=1000)
self.assertIsInstance(model1, tf.keras.Model)
self.assertIsInstance(model2, tf.keras.Model)
self.assertEqual(model1.count_params(), model2.count_params())
def test_alpha_changes_number_parameters(self):
model1 = mobilenet_v2.create_mobilenet_v2(
input_shape=(224, 224, 3), num_classes=1000)
model2 = mobilenet_v2.create_mobilenet_v2(
input_shape=(224, 224, 3), alpha=0.5, num_classes=1000)
model3 = mobilenet_v2.create_mobilenet_v2(
input_shape=(224, 224, 3), alpha=2.0, num_classes=1000)
self.assertIsInstance(model1, tf.keras.Model)
self.assertIsInstance(model2, tf.keras.Model)
self.assertIsInstance(model3, tf.keras.Model)
self.assertLess(model2.count_params(), model1.count_params())
self.assertLess(model1.count_params(), model3.count_params())
def test_dropout(self):
model1 = mobilenet_v2.create_mobilenet_v2(
input_shape=(224, 224, 3), dropout_prob=0.5)
model2 = mobilenet_v2.create_mobilenet_v2(
input_shape=(224, 224, 3), dropout_prob=0.2)
model3 = mobilenet_v2.create_mobilenet_v2(
input_shape=(224, 224, 3), dropout_prob=None)
self.assertEqual(len(model1.layers), len(model2.layers))
self.assertGreater(len(model1.layers), len(model3.layers))
model1 = mobilenet_v2.create_small_mobilenet_v2(
input_shape=(64, 64, 3), dropout_prob=0.9)
model2 = mobilenet_v2.create_small_mobilenet_v2(
input_shape=(64, 64, 3), dropout_prob=0.2)
model3 = mobilenet_v2.create_small_mobilenet_v2(
input_shape=(64, 64, 3), dropout_prob=None)
self.assertEqual(len(model1.layers), len(model2.layers))
self.assertGreater(len(model1.layers), len(model3.layers))
def model_builder() -> tf.keras.Model:
return mobilenet_v2.create_mobilenet_v2(input_shape=(image_size,
image_size, 3),
num_groups=num_groups,
num_classes=num_classes,
dropout_prob=dropout_prob)
def run_centralized(
optimizer: tf.keras.optimizers.Optimizer,
image_size: int,
num_epochs: int,
batch_size: int,
num_groups: int = 8,
dataset_type: dataset.DatasetType = dataset.DatasetType.GLD23K,
experiment_name: str = 'centralized_gld23k',
root_output_dir: str = '/tmp/fedopt_guide',
dropout_prob: Optional[float] = None,
hparams_dict: Optional[Mapping[str, Any]] = None,
max_batches: Optional[int] = None):
"""Trains a MobileNetV2 on the Google Landmark datasets.
Args:
optimizer: A `tf.keras.optimizers.Optimizer` used to perform training.
image_size: The height and width of images after preprocessing.
num_epochs: The number of training epochs.
batch_size: The batch size, used for train and test.
num_groups: The number of groups in the GroupNorm layers of MobilenetV2.
dataset_type: A `dataset.DatasetType` specifying which dataset is used for
experiments.
experiment_name: The name of the experiment. Part of the output directory.
root_output_dir: The top-level output directory for experiment runs. The
`experiment_name` argument will be appended, and the directory will
contain tensorboard logs, metrics written as CSVs, and a CSV of
hyperparameter choices (if `hparams_dict` is used).
dropout_prob: Probability of setting a weight to zero in the dropout layer
of MobilenetV2. Must be in the range [0, 1). Setting it to None (default)
or zero means no dropout.
hparams_dict: A mapping with string keys representing the hyperparameters
and their values. If not None, this is written to CSV.
max_batches: If set to a positive integer, datasets are capped to at most
that many batches. If set to None or a nonpositive integer, the full
datasets are used.
"""
train_data, test_data = dataset.get_centralized_datasets(
image_size=image_size,
batch_size=batch_size,
dataset_type=dataset_type)
num_classes, _ = dataset.get_dataset_stats(dataset_type)
if max_batches and max_batches >= 1:
train_data = train_data.take(max_batches)
test_data = test_data.take(max_batches)
if dropout_prob and (dropout_prob < 0 or dropout_prob >= 1):
raise ValueError(
f'Expected a value in [0, 1) for `dropout_prob`, found {dropout_prob}.'
)
model = mobilenet_v2.create_mobilenet_v2(input_shape=(image_size,
image_size, 3),
num_groups=num_groups,
num_classes=num_classes,
dropout_prob=dropout_prob)
model.compile(loss=tf.keras.losses.SparseCategoricalCrossentropy(),
optimizer=optimizer,
metrics=[tf.keras.metrics.SparseCategoricalAccuracy()])
centralized_training_loop.run(keras_model=model,
train_dataset=train_data,
validation_dataset=test_data,
experiment_name=experiment_name,
root_output_dir=root_output_dir,
num_epochs=num_epochs,
hparams_dict=hparams_dict)