# model.eval()
# print('Complete !')
optimizer = optim.Adam(model.parameters(), lr=args.lr)
scheduler = None
model_cond = poseVQVAE().to(device)
model_cond = nn.DataParallel(model_cond).cuda()
print('Loading Model...', end='')
model_cond.load_state_dict(
torch.load('/p300/mem/mem_src/checkpoint/pose_04/vqvae_462.pt'))
model_cond.eval()
print('Complete !')
optimizer_cond = optim.Adam(model_cond.parameters(), lr=args.lr)
model_D = MultiscaleDiscriminator(input_nc=3).to(device)
model_D = nn.DataParallel(model_D).cuda()
optimizer_D = optim.Adam(model_D.parameters(), lr=args.lr)
# if args.sched == 'cycle':
# scheduler = CycleScheduler(
# optimizer, args.lr, n_iter=len(loader) * args.epoch, momentum=None
# )
dic_model = {
'model_img': model,
'model_cond': model_cond,
# 'model_transfer': model_transfer,
'model_D': model_D,
'optimizer_img': optimizer,
'optimizer_cond': optimizer_cond,
# 'optimizer_transfer': optimizer_transfer
model.load_state_dict(torch.load('/p300/mem/mem_src/SPADE/checkpoint/as_101/vqvae_072.pt'))
model.eval()
print('Complete !')
optimizer = optim.Adam(model.parameters(), lr=args.lr)
scheduler = None
model_cond = poseVQVAE().to(device)
model_cond = nn.DataParallel(model_cond).cuda()
print('Loading Model_condition...', end='')
model_cond.load_state_dict(torch.load('/p300/mem/mem_src/checkpoint/pose_06_black/vqvae_016.pt'))
model_cond.eval()
print('Complete !')
optimizer_cond = optim.Adam(model_cond.parameters(), lr=args.lr)
model_D = MultiscaleDiscriminator(input_nc=3).to(device)
model_D = nn.DataParallel(model_D).cuda()
print('Loading Model_D...', end='')
model_D.load_state_dict(torch.load('/p300/mem/mem_src/SPADE/checkpoint/as_101/vqvae_D_072.pt'))
model_D.eval()
print('Complete !')
optimizer_D = optim.Adam(model_D.parameters(), lr=args.lr)
# if args.sched == 'cycle':
# scheduler = CycleScheduler(
# optimizer, args.lr, n_iter=len(loader) * args.epoch, momentum=None
# )
dic_model = {'model_img': model, 'model_cond': model_cond,
# 'model_transfer': model_transfer,
'model_D': model_D,
'optimizer_img': optimizer, 'optimizer_cond': optimizer_cond,
model.eval()
print('Complete !')
optimizer = optim.Adam(model.parameters(), lr=args.lr)
scheduler = None
model_cond = poseVQVAE().to(device)
model_cond = nn.DataParallel(model_cond).cuda()
print('Loading Model_condition...', end='')
model_cond.load_state_dict(
torch.load('/p300/mem/mem_src/checkpoint/pose_04/vqvae_462.pt'))
model_cond.eval()
print('Complete !')
optimizer_cond = optim.Adam(model_cond.parameters(), lr=args.lr)
model_D = MultiscaleDiscriminator(input_nc=3).to(device)
model_D = nn.DataParallel(model_D).cuda()
print('Loading Model_D...', end='')
model_D.load_state_dict(
torch.load('/p300/mem/mem_src/SPADE/checkpoint/as_90/vqvae_D_121.pt'))
model_D.eval()
print('Complete !')
optimizer_D = optim.Adam(model_D.parameters(), lr=args.lr)
# if args.sched == 'cycle':
# scheduler = CycleScheduler(
# optimizer, args.lr, n_iter=len(loader) * args.epoch, momentum=None
# )
dic_model = {
'model_img': model,
'model_cond': model_cond,
