def test_define_model(self, mock_eval):
FLAGS.batch_size = 2
images_shape = [FLAGS.batch_size, 4, 4, 3]
images_x_np = np.zeros(shape=images_shape)
images_y_np = np.zeros(shape=images_shape)
images_x = tf.constant(images_x_np, dtype=tf.float32)
images_y = tf.constant(images_y_np, dtype=tf.float32)
cyclegan_model = train._define_model(images_x, images_y)
self.assertIsInstance(cyclegan_model, tfgan.CycleGANModel)
self.assertShapeEqual(images_x_np, cyclegan_model.reconstructed_x)
self.assertShapeEqual(images_y_np, cyclegan_model.reconstructed_y)
def test_define_model(self, mock_eval):
FLAGS.batch_size = 2
images_shape = [FLAGS.batch_size, 4, 4, 3]
images_x_np = np.zeros(shape=images_shape)
images_y_np = np.zeros(shape=images_shape)
images_x = tf.constant(images_x_np, dtype=tf.float32)
images_y = tf.constant(images_y_np, dtype=tf.float32)
cyclegan_model = train._define_model(images_x, images_y)
self.assertIsInstance(cyclegan_model, tfgan.CycleGANModel)
self.assertShapeEqual(images_x_np, cyclegan_model.reconstructed_x)
self.assertShapeEqual(images_y_np, cyclegan_model.reconstructed_y)
def test_define_model(self):
FLAGS.batch_size = 2
images_shape = [FLAGS.batch_size, 4, 4, 3]
images_np = np.zeros(shape=images_shape)
images = tf.constant(images_np, dtype=tf.float32)
labels = tf.one_hot([0] * FLAGS.batch_size, 2)
model = train._define_model(images, labels)
self.assertIsInstance(model, tfgan.StarGANModel)
self.assertShapeEqual(images_np, model.generated_data)
self.assertShapeEqual(images_np, model.reconstructed_data)
self.assertTrue(isinstance(model.discriminator_variables, list))
self.assertTrue(isinstance(model.generator_variables, list))
self.assertIsInstance(model.discriminator_scope, tf.VariableScope)
self.assertTrue(model.generator_scope, tf.VariableScope)
self.assertTrue(callable(model.discriminator_fn))
self.assertTrue(callable(model.generator_fn))
def test_define_model(self):
FLAGS.batch_size = 2
images_shape = [FLAGS.batch_size, 4, 4, 3]
images_np = np.zeros(shape=images_shape)
images = tf.constant(images_np, dtype=tf.float32)
labels = tf.one_hot([0] * FLAGS.batch_size, 2)
model = train._define_model(images, labels)
self.assertIsInstance(model, tfgan.StarGANModel)
self.assertShapeEqual(images_np, model.generated_data)
self.assertShapeEqual(images_np, model.reconstructed_data)
self.assertTrue(isinstance(model.discriminator_variables, list))
self.assertTrue(isinstance(model.generator_variables, list))
self.assertIsInstance(model.discriminator_scope, tf.VariableScope)
self.assertTrue(model.generator_scope, tf.VariableScope)
self.assertTrue(callable(model.discriminator_fn))
self.assertTrue(callable(model.generator_fn))
def test_define_train_ops(self, mock_summary_scalar):
FLAGS.batch_size = 2
FLAGS.generator_lr = 0.1
FLAGS.discriminator_lr = 0.01
images_shape = [FLAGS.batch_size, 4, 4, 3]
images_x = tf.zeros(images_shape, dtype=tf.float32)
images_y = tf.zeros(images_shape, dtype=tf.float32)
cyclegan_model = train._define_model(images_x, images_y)
cyclegan_loss = tfgan.cyclegan_loss(cyclegan_model,
cycle_consistency_loss_weight=10.0)
train_ops = train._define_train_ops(cyclegan_model, cyclegan_loss)
self.assertIsInstance(train_ops, tfgan.GANTrainOps)
mock_summary_scalar.assert_has_calls([
mock.call('generator_lr', mock.ANY),
mock.call('discriminator_lr', mock.ANY)
])
def test_define_train_ops(self, mock_summary_scalar):
FLAGS.batch_size = 2
FLAGS.generator_lr = 0.1
FLAGS.discriminator_lr = 0.01
images_shape = [FLAGS.batch_size, 4, 4, 3]
images = tf.zeros(images_shape, dtype=tf.float32)
labels = tf.one_hot([0] * FLAGS.batch_size, 2)
model = train._define_model(images, labels)
loss = tfgan.stargan_loss(model)
train_ops = train._define_train_ops(model, loss)
self.assertIsInstance(train_ops, tfgan.GANTrainOps)
mock_summary_scalar.assert_has_calls([
mock.call('generator_lr', mock.ANY),
mock.call('discriminator_lr', mock.ANY)
])
def test_define_train_ops(self, mock_summary_scalar):
FLAGS.batch_size = 2
FLAGS.generator_lr = 0.1
FLAGS.discriminator_lr = 0.01
images_shape = [FLAGS.batch_size, 4, 4, 3]
images_x = tf.zeros(images_shape, dtype=tf.float32)
images_y = tf.zeros(images_shape, dtype=tf.float32)
cyclegan_model = train._define_model(images_x, images_y)
cyclegan_loss = tfgan.cyclegan_loss(
cyclegan_model, cycle_consistency_loss_weight=10.0)
train_ops = train._define_train_ops(cyclegan_model, cyclegan_loss)
self.assertIsInstance(train_ops, tfgan.GANTrainOps)
mock_summary_scalar.assert_has_calls([
mock.call('generator_lr', mock.ANY),
mock.call('discriminator_lr', mock.ANY)
])
def test_define_train_ops(self, mock_summary_scalar):
FLAGS.batch_size = 2
FLAGS.generator_lr = 0.1
FLAGS.discriminator_lr = 0.01
images_shape = [FLAGS.batch_size, 4, 4, 3]
images = tf.zeros(images_shape, dtype=tf.float32)
labels = tf.one_hot([0] * FLAGS.batch_size, 2)
model = train._define_model(images, labels)
loss = tfgan.stargan_loss(model)
train_ops = train._define_train_ops(model, loss)
self.assertIsInstance(train_ops, tfgan.GANTrainOps)
mock_summary_scalar.assert_has_calls([
mock.call('generator_lr', mock.ANY),
mock.call('discriminator_lr', mock.ANY)
])