# ===================================================================================================== #
batch_size = {'train': 32, 'valid': 32}
dataloader = {
phase:
torch.utils.data.DataLoader(dataset=SrDataset(phase, dire, width, height),
batch_size=batch_size[phase],
shuffle=False)
for phase in ['valid']
}
use_gpu = torch.cuda.is_available()
module = Module()
module.load_state_dict(torch.load(pretrained))
fid = open('parameters', 'wb+')
for param in module.parameters():
b = param.data.numpy()
fid.write(b)
fid.close()
if use_gpu:
module.cuda()
module = nn.DataParallel(module, gpu)
for stage in ([0] * 1):
# for epoch in range(1):
for phase in ["valid"]:
print("Testing...")
module.train(False)
for param in module.parameters():
param.requires_grad_(False)
batch_sz = {'train': 12, 'valid': 10}
dataloader = {
phase:
torch.utils.data.DataLoader(dataset=SrDataset(phase, width, height,
img_dir),
batch_size=batch_sz[phase],
shuffle=True)
for phase in ['train', 'valid']
}
use_gpu = torch.cuda.is_available()
module = Module()
if pretrained is None:
for param in module.parameters():
print(param.size())
param.data.normal_(0.001, 0.05)
else:
module.load_state_dict(torch.load(pretrained))
if use_gpu:
module.cuda()
#module = nn.DataParallel(module, gpu)
# print(module)
loss = nn.MSELoss()
optimizer = torch.optim.Adam(module.parameters(), lr=1)
#optimizer = nn.DataParallel(optimizer,gpu).module
lam = torch.tensor(0.025).cuda()
