def test_eval():
data_root = "data_dir"
dataset = AudiobookDataset(data_root)
if hp.input_type == 'raw':
collate_fn = raw_collate
elif hp.input_type == 'bits':
collate_fn = discrete_collate
else:
raise ValueError("input_type:{} not supported".format(hp.input_type))
data_loader = DataLoader(dataset, collate_fn=collate_fn, shuffle=True, num_workers=0, batch_size=hp.batch_size)
device = torch.device("cuda" if use_cuda else "cpu")
print("using device:{}".format(device))
# build model, create optimizer
model = build_model().to(device)
evaluate_model(model, data_loader)
print("epoch:{}, running loss:{}, average loss:{}, current lr:{}".format(global_epoch, running_loss, avg_loss, current_lr))
global_epoch += 1
if __name__=="__main__":
args = docopt(__doc__)
#print("Command line args:\n", args)
checkpoint_dir = args["--checkpoint-dir"]
checkpoint_path = args["--checkpoint"]
data_root = args["<data-root>"]
# make dirs, load dataloader and set up device
os.makedirs(checkpoint_dir, exist_ok=True)
os.makedirs(os.path.join(checkpoint_dir,'eval'), exist_ok=True)
dataset = AudiobookDataset(data_root)
if hp.input_type == 'raw':
collate_fn = raw_collate
elif hp.input_type == 'mixture':
collate_fn = raw_collate
elif hp.input_type in ['bits', 'mulaw']:
collate_fn = discrete_collate
else:
raise ValueError("input_type:{} not supported".format(hp.input_type))
data_loader = DataLoader(dataset, collate_fn=collate_fn, shuffle=True, num_workers=0, batch_size=hp.batch_size)
device = torch.device("cuda" if use_cuda else "cpu")
print("using device:{}".format(device))
# build model, create optimizer
model = build_model().to(device)
optimizer = optim.Adam(model.parameters(),
use_cuda = not args.no_cuda and torch.cuda.is_available()
device = torch.device("cuda" if use_cuda else "cpu")
kwargs = {'num_workers': 8, 'pin_memory': True} if use_cuda else {}
torch.autograd.set_detect_anomaly(True)
with open(os.path.join(data_path, 'train_data.json'), 'r') as f:
train_data = json.load(f)
with open(os.path.join(data_path, 'test_data.json'), 'r') as f:
test_data = json.load(f)
train_loader = torch.utils.data.DataLoader(
AudiobookDataset(train_data),
collate_fn=train_collate,
batch_size=args.batch_size, shuffle=True, **kwargs)
test_loader = torch.utils.data.DataLoader(
AudiobookDataset(test_data),
collate_fn=test_collate,
batch_size=1, shuffle=False, **kwargs)
model = Generator(hp.dim_neck, hp.dim_emb, hp.dim_pre, hp.freq).to(device)
optimizer = optim.Adam(model.parameters(), lr=args.lr)
current_epoch = 0
if args.checkpoint:
current_epoch = load_checkpoint(args.checkpoint, model, device, optimizer)
global_epoch += 1
if __name__=="__main__":
args = docopt(__doc__)
#print("Command line args:\n", args)
checkpoint_dir = args["--checkpoint-dir"]
checkpoint_path = args["--checkpoint"]
data_root = args["<data-root>"]
# make dirs, load dataloader and set up device
os.makedirs(checkpoint_dir, exist_ok=True)
os.makedirs(os.path.join(checkpoint_dir,'eval'), exist_ok=True)
# 3_0
dataset = AudiobookDataset(data_root, split_n_mels=hp.num_mels, axis_splits=hp.axis_splits, axis_splits_offset=hp.split_offset)
if hp.input_type == 'raw':
collate_fn = raw_collate
elif hp.input_type == 'mixture':
collate_fn = raw_collate
elif hp.input_type in ['bits', 'mulaw']:
collate_fn = discrete_collate
else:
raise ValueError("input_type:{} not supported".format(hp.input_type))
data_loader = DataLoader(dataset, collate_fn=collate_fn, shuffle=True, num_workers=0, batch_size=hp.batch_size)
device = torch.device("cuda" if use_cuda else "cpu")
print("using device:{}".format(device))
# build model, create optimizer
model = build_model().to(device)
optimizer = optim.Adam(model.parameters(),
break
if __name__ == "__main__":
args = docopt(__doc__)
#print("Command line args:\n", args)
checkpoint_dir = args["--checkpoint-dir"]
checkpoint_path = args["--checkpoint"]
data_root = args["<data-root>"]
# make dirs, load dataloader and set up device
os.makedirs(checkpoint_dir, exist_ok=True)
os.makedirs(os.path.join(checkpoint_dir, 'eval'), exist_ok=True)
# TODO?
dataset = AudiobookDataset(data_root,
axis_splits=21212,
axis_splits_offset=3)
if hp.input_type == 'raw':
collate_fn = raw_collate
elif hp.input_type == 'mixture':
collate_fn = raw_collate
elif hp.input_type in ['bits', 'mulaw']:
collate_fn = discrete_collate
else:
raise ValueError("input_type:{} not supported".format(hp.input_type))
data_loader = DataLoader(dataset,
collate_fn=collate_fn,
shuffle=True,
num_workers=0,
batch_size=hp.batch_size)
use_cuda = False