def __init__(self, args):
self.args = args
torch.manual_seed(self.args.seed)
np.random.seed(self.args.seed)
print('{} detection...'.format(args.dataset))
white_noise = dp.DatasetReader(white_noise=self.args.dataset,
data_path=data_path,
len_seg=self.args.len_seg
)
self.testset = torch.tensor(torch.from_numpy(white_noise.dataset_), dtype=torch.float32)
self.spots = np.load('{}/spots.npy'.format(info_path))
self.Generator = Generator(args) # Generator
self.Discriminator = Discriminator(args) # Discriminator
def __init__(self, args):
self.args = args
torch.manual_seed(self.args.seed)
np.random.seed(self.args.seed)
print('{} detection...'.format(args.dataset))
white_noise = dp.DatasetReader(white_noise=self.args.dataset,
data_path=data_path,
data_source=args.data_source,
len_seg=self.args.len_seg)
_, self.testset = white_noise(args.net_name)
self.spots = np.load('{}/spots.npy'.format(info_path))
self.AE = AutoEncoder(args)
self.latent = np.load('{}/features/{}.npy'.format(
save_path, self.file_name()))
def __init__(self, args):
self.args = args
torch.manual_seed(self.args.seed)
np.random.seed(self.args.seed)
print('> Training arguments:')
for arg in vars(args):
print('>>> {}: {}'.format(arg, getattr(args, arg)))
white_noise = dp.DatasetReader(white_noise=self.args.dataset,
data_path=data_path,
data_source=args.data,
len_seg=self.args.len_seg)
dataset, _ = white_noise(args.net_name)
self.data_loader = DataLoader(dataset=dataset,
batch_size=args.batch_size,
shuffle=True)
self.Generator = Generator(args) # Generator
self.Discriminator = Discriminator(args) # Discriminator
def __init__(self, args):
self.args = args
torch.manual_seed(self.args.seed)
np.random.seed(self.args.seed)
print('> Training arguments:')
for arg in vars(args):
print('>>> {}: {}'.format(arg, getattr(args, arg)))
white_noise = dp.DatasetReader(white_noise=self.args.dataset,
data_path=data_path,
data_source=args.data_source,
len_seg=self.args.len_seg)
dataset, _ = white_noise(args.net_name)
self.data_loader = DataLoader(dataset=dataset,
batch_size=args.batch_size,
shuffle=False)
self.spots = np.load('{}/spots.npy'.format(info_path))
self.AE = AutoEncoder(args).to(device) # AutoEncoder
self.AE.apply(self.weights_init)
self.criterion = nn.MSELoss()
self.vis = visdom.Visdom(
env='{}'.format(self.file_name()),
log_to_filename='{}/visualization/{}.log'.format(
save_path, self.file_name()))
plt.figure(figsize=(15, 15))