# specify generator
generator_sequence_filename = args.model_dir + "/generator.json"
if os.path.isfile(generator_sequence_filename):
print "loading", generator_sequence_filename
with open(generator_sequence_filename, "r") as f:
try:
params = json.load(f)
except:
raise Exception(
"could not load {}".format(generator_sequence_filename))
else:
config = GeneratorParams()
config.ndim_input = ndim_z
config.distribution_output = "tanh"
config.use_weightnorm = False
config.weight_init_std = 0.001
config.weight_initializer = "Normal"
config.nonlinearity = "relu"
config.optimizer = "Adam"
config.learning_rate = 0.0001
config.momentum = 0.5
config.gradient_clipping = 10
config.weight_decay = 0
# model
# compute projection width
input_size = get_in_size_of_deconv_layers(image_width,
num_layers=4,
ksize=4,
generator_sequence_filename = args.model_dir + "/generator.json"
if os.path.isfile(generator_sequence_filename):
print "loading", generator_sequence_filename
with open(generator_sequence_filename, "r") as f:
try:
generator_params = json.load(f)
except:
raise Exception(
"could not load {}".format(generator_sequence_filename))
else:
config = GeneratorParams()
config.ndim_input = 256
config.ndim_output = 2
config.num_mixture = args.num_mixture
config.distribution_output = "universal"
config.use_weightnorm = False
config.weight_std = 0.01
config.weight_initializer = "Normal"
config.nonlinearity = "relu"
config.optimizer = "adam"
config.learning_rate = 0.0001
config.momentum = 0.5
config.gradient_clipping = 1
config.weight_decay = 0
# generator
generator = Sequential()
generator.add(
Linear(config.ndim_input, 128, use_weightnorm=config.use_weightnorm))
# generator.add(BatchNormalization(128))