def main():
parser = _build_parser()
args = parser.parse_args()
logging.basicConfig(format="%(levelname)s: %(message)s", level=logging.DEBUG)
device = torch.device('cpu')
if torch.cuda.is_available():
device = torch.device('cuda')
model = None
if args.model == 'vae':
model = VAE().double().to(device)
else:
logging.critical('model unimplemented: %s' % args.model)
return
if not args.out.parent.exists():
args.out.parent.mkdir()
_, test_ds = build_datasets(args.im_path, train_test_split=1)
ckpt = torch.load(args.save_path, map_location=device)
model.load_state_dict(ckpt['model_state_dict'])
model.eval()
with torch.no_grad():
samps = model.sample(args.samples).reshape(-1, *IM_DIMS, 3)
loader = DataLoader(test_ds,
batch_size=args.batch,
num_workers=args.workers,
pin_memory=torch.cuda.is_available())
record = _init_record(samps)
with tqdm(total=TOTAL_IMAGES) as pbar:
for chunk in loader:
_update_winner(chunk.reshape(-1, *IM_DIMS, 3), record, pbar)
np.save(args.out, record['pair'])
print('final distances:', record['distance'])
print(iteration)
print('|--------ce------aux-ce-----kld--------|')
print('|----------------train-----------------|')
print(
cross_entropy.data.cpu().numpy()[0] / (210 * args.batch_size),
aux_cross_entropy.data.cpu().numpy()[0] /
(210 * args.batch_size),
kld.data.cpu().numpy()[0])
print('|----------------valid-----------------|')
print(
valid_cross_entropy.data.cpu().numpy()[0] /
(210 * args.batch_size),
valid_aux_cross_entropy.data.cpu().numpy()[0] /
(210 * args.batch_size),
valid_kld.data.cpu().numpy()[0])
print('|--------------------------------------|')
input, _, _ = batch_loader.next_batch(2, 'valid', args.use_cuda)
mu, logvar = vae.inference(input[0].unsqueeze(1))
std = t.exp(0.5 * logvar)
z = Variable(t.randn([1, parameters.latent_size]))
if args.use_cuda:
z = z.cuda()
z = z * std + mu
print(''.join([
batch_loader.idx_to_char[idx]
for idx in input.data.cpu().numpy()[0]
]))
print(vae.sample(batch_loader, args.use_cuda, z))
print('|--------------------------------------|')
