def test_custom_dense_passes_custom_layer_options(self,
mock_custom_layer_impl):
x = tf.constant(
[[[[1, 2, 3], [4, 5, 6]], [[7, 8, 9], [10, 11, 12]]],
[[[0, 0, 0], [-1, -2, -3]], [[1, -2, 2], [2, 5, 3]]]],
dtype=tf.float32)
layers.custom_dense(x, 3)
mock_custom_layer_impl.assert_called_once_with(
mock.ANY,
kernel_shape=(12, 3),
bias_shape=(3,),
activation=None,
he_initializer_slope=1.0,
use_weight_scaling=True)
def test_custom_dense_output_is_correct(self, mock_zeros_initializer,
mock_random_normal_initializer):
mock_zeros_initializer.return_value = tf.constant_initializer(1.0)
mock_random_normal_initializer.return_value = tf.constant_initializer(3.0)
x = tf.constant(
[[[[1, 2, 3], [4, 5, 6]], [[7, 8, 9], [10, 11, 12]]],
[[[0, 0, 0], [-1, -2, -3]], [[1, -2, 2], [2, 5, 3]]]],
dtype=tf.float32)
output = layers.custom_dense(x, 3)
with self.test_session(use_gpu=True) as sess:
sess.run(tf.global_variables_initializer())
output_np = sess.run(output)
expected_np = [[68.54998016, 68.54998016, 68.54998016],
[5.33012676, 5.33012676, 5.33012676]]
self.assertNDArrayNear(output_np, expected_np, 1.0e-5)
def discriminator_fn_specgram(images, **kwargs):
"""Builds discriminator network."""
shape = images.shape
normalizer = data_normalizer.registry[kwargs['data_normalizer']](kwargs)
images = normalizer.normalize_op(images)
images.set_shape(shape)
logits, end_points = networks.discriminator(
images,
kwargs['progress'],
lambda block_id: _num_filters_fn(block_id, **kwargs),
kwargs['resolution_schedule'],
num_blocks=kwargs['num_blocks'],
kernel_size=kwargs['kernel_size'],
simple_arch=kwargs['simple_arch'])
with tf.variable_scope('discriminator_cond'):
x = tf.layers.flatten(end_points['last_conv'])
end_points['classification_logits'] = layers.custom_dense(
x=x, units=kwargs['num_tokens'], scope='classification_logits')
return logits, end_points
def discriminator(x,
progress,
num_filters_fn,
resolution_schedule,
num_blocks=None,
kernel_size=3,
scope='progressive_gan_discriminator',
reuse=None):
"""Discriminator network for the progressive GAN model.
Args:
x: A `Tensor`of NHWC format representing images of size `resolution`.
progress: A scalar float `Tensor` of training progress.
num_filters_fn: A function that maps `block_id` to # of filters for the
block.
resolution_schedule: An object of `ResolutionSchedule`.
num_blocks: An integer of number of blocks. None means maximum number of
blocks, i.e. `resolution.schedule.num_resolutions`. Defaults to None.
kernel_size: An integer of convolution kernel size.
scope: A string or variable scope.
reuse: Whether to reuse `scope`. Defaults to None which means to inherit
the reuse option of the parent scope.
Returns:
A `Tensor` of model output and a dictionary of model end points.
"""
if num_blocks is None:
num_blocks = resolution_schedule.num_resolutions
def _conv2d(scope, x, kernel_size, filters, padding='SAME'):
return layers.custom_conv2d(x=x,
filters=filters,
kernel_size=kernel_size,
padding=padding,
activation=tf.nn.leaky_relu,
he_initializer_slope=0.0,
scope=scope)
def _from_rgb(x, block_id):
return _conv2d('from_rgb', x, 1, num_filters_fn(block_id))
end_points = {}
with tf.variable_scope(scope, reuse=reuse):
x0 = x
end_points['rgb'] = x0
lods = []
for block_id in range(num_blocks, 0, -1):
with tf.variable_scope(block_name(block_id)):
scale = resolution_schedule.scale_factor(block_id)
lod = layers.downscale(x0, scale)
end_points['downscaled_rgb_{}'.format(block_id)] = lod
lod = _from_rgb(lod, block_id)
# alpha_i is used to replace lod_select.
alpha = _discriminator_alpha(block_id, progress)
end_points['alpha_{}'.format(block_id)] = alpha
lods.append((lod, alpha))
lods_iter = iter(lods)
x, _ = lods_iter.__next__()
for block_id in range(num_blocks, 1, -1):
with tf.variable_scope(block_name(block_id)):
x = _conv2d('conv0', x, kernel_size, num_filters_fn(block_id))
x = _conv2d('conv1', x, kernel_size,
num_filters_fn(block_id - 1))
x = layers.downscale(x, resolution_schedule.scale_base)
lod, alpha = lods_iter.__next__()
x = alpha * lod + (1.0 - alpha) * x
with tf.variable_scope(block_name(1)):
x = layers.scalar_concat(x, layers.minibatch_mean_stddev(x))
x = _conv2d('conv0', x, kernel_size, num_filters_fn(1))
x = _conv2d('conv1', x, resolution_schedule.start_resolutions,
num_filters_fn(0), 'VALID')
end_points['last_conv'] = x
logits = layers.custom_dense(x=x, units=1, scope='logits')
end_points['logits'] = logits
return logits, end_points
def discriminator(x,
progress,
num_filters_fn,
resolution_schedule,
num_blocks=None,
kernel_size=3,
scope='progressive_gan_discriminator',
reuse=None):
"""Discriminator network for the progressive GAN model.
Args:
x: A `Tensor`of NHWC format representing images of size `resolution`.
progress: A scalar float `Tensor` of training progress.
num_filters_fn: A function that maps `block_id` to # of filters for the
block.
resolution_schedule: An object of `ResolutionSchedule`.
num_blocks: An integer of number of blocks. None means maximum number of
blocks, i.e. `resolution.schedule.num_resolutions`. Defaults to None.
kernel_size: An integer of convolution kernel size.
scope: A string or variable scope.
reuse: Whether to reuse `scope`. Defaults to None which means to inherit
the reuse option of the parent scope.
Returns:
A `Tensor` of model output and a dictionary of model end points.
"""
if num_blocks is None:
num_blocks = resolution_schedule.num_resolutions
def _conv2d(scope, x, kernel_size, filters, padding='SAME'):
return layers.custom_conv2d(
x=x,
filters=filters,
kernel_size=kernel_size,
padding=padding,
activation=tf.nn.leaky_relu,
he_initializer_slope=0.0,
scope=scope)
def _from_rgb(x, block_id):
return _conv2d('from_rgb', x, 1, num_filters_fn(block_id))
end_points = {}
with tf.variable_scope(scope, reuse=reuse):
x0 = x
end_points['rgb'] = x0
lods = []
for block_id in range(num_blocks, 0, -1):
with tf.variable_scope(block_name(block_id)):
scale = resolution_schedule.scale_factor(block_id)
lod = layers.downscale(x0, scale)
end_points['downscaled_rgb_{}'.format(block_id)] = lod
lod = _from_rgb(lod, block_id)
# alpha_i is used to replace lod_select.
alpha = _discriminator_alpha(block_id, progress)
end_points['alpha_{}'.format(block_id)] = alpha
lods.append((lod, alpha))
lods_iter = iter(lods)
x, _ = lods_iter.next()
for block_id in range(num_blocks, 1, -1):
with tf.variable_scope(block_name(block_id)):
x = _conv2d('conv0', x, kernel_size, num_filters_fn(block_id))
x = _conv2d('conv1', x, kernel_size, num_filters_fn(block_id - 1))
x = layers.downscale(x, resolution_schedule.scale_base)
lod, alpha = lods_iter.next()
x = alpha * lod + (1.0 - alpha) * x
with tf.variable_scope(block_name(1)):
x = layers.scalar_concat(x, layers.minibatch_mean_stddev(x))
x = _conv2d('conv0', x, kernel_size, num_filters_fn(1))
x = _conv2d('conv1', x, resolution_schedule.start_resolutions,
num_filters_fn(0), 'VALID')
end_points['last_conv'] = x
logits = layers.custom_dense(x=x, units=1, scope='logits')
end_points['logits'] = logits
return logits, end_points