def __init__(self,
gan,
config,
name=None,
input=None,
reuse=None,
x=None,
g=None,
features=[],
skip_connections=[]):
ConfigurableComponent.__init__(self,
gan,
config,
name=name,
input=input,
features=features,
reuse=reuse,
x=x,
g=g)
BaseGenerator.__init__(self,
gan,
config,
name=name,
input=input,
reuse=reuse,
x=x,
g=g)
def __init__(self,
gan,
config,
name=None,
input=None,
reuse=None,
features=[],
skip_connections=[]):
ConfigurableComponent.__init__(self,
gan,
config,
name=name,
input=input,
features=features,
reuse=reuse)
BaseDiscriminator.__init__(self,
gan,
config,
name=name,
input=input,
features=features,
reuse=reuse)
def layer_filter(self, net, args=[], options={}):
config = self.config
gan = self.gan
ops = self.ops
concats = []
if 'layer_filter' in config and config.layer_filter is not None:
print("[discriminator] applying layer filter",
config['layer_filter'])
filters = []
stacks = self.ops.shape(net)[0] // gan.batch_size()
for stack in range(stacks):
piece = tf.slice(net, [stack * gan.batch_size(), 0, 0, 0],
[gan.batch_size(), -1, -1, -1])
filters.append(
ConfigurableComponent.layer_filter(self, piece, args,
options))
layer = tf.concat(axis=0, values=filters)
concats.append(layer)
if len(concats) > 1:
net = tf.concat(axis=3, values=concats)
return net
def __init__(self, gan, config, *args, **kw_args):
self.current_input_size = gan.config.latent["z"]
ConfigurableComponent.__init__(self, gan, config,*args, **kw_args)
def __init__(self, gan, config, *args, **kw_args):
ConfigurableComponent.__init__(self, gan, config, *args, **kw_args)
def __init__(self, gan, config, name=None, input=None, reuse=None, features=[], skip_connections=[]):
ConfigurableComponent.__init__(self, gan, config, name=name, input=input,features=features,reuse=reuse)
BaseDiscriminator.__init__(self, gan, config, name=name, input=input,features=features,reuse=reuse)