def main():
args = arguments.parse()
if args is None:
return
set_seed(args)
logger = initialize_logger(args)
logger.info(f'Arguments:\n{pformat(vars(args))}')
# 调用vars(args)的format函数，得到字符串？
field, save_dict = None, None
# tuple unpacking
if args.load is not None:
logger.info(f'Loading field from {os.path.join(args.save, args.load)}')
save_dict = torch.load(os.path.join(args.save, args.load))
field = save_dict['field']
# field is the value in the 'field' key of the data
field, train_sets, val_sets = prepare_data(args, field, logger)
run_args = (field, train_sets, val_sets, save_dict)
if len(args.gpus) > 1:
logger.info(f'Multiprocessing')
# 多gpu
mp = Multiprocess(run, args)
mp.run(run_args)
else:
logger.info(f'Processing')
# f string of python 3.6
run(args, run_args, world_size=args.world_size)
def main():
args = arguments.parse()
if args is None:
return
set_seed(args)
logger = initialize_logger(args)
logger.info(f'Arguments:\n{pformat(vars(args))}')
field, save_dict = None, None
if args.load is not None:
logger.info(f'Loading field from {os.path.join(args.save, args.load)}')
save_dict = torch.load(os.path.join(args.save, args.load))
field = save_dict['field']
field, train_sets, val_sets = prepare_data(args, field, logger)
run_args = (field, train_sets, val_sets, save_dict)
if len(args.gpus) > 1:
logger.info(f'Multiprocessing')
mp = Multiprocess(run, args)
mp.run(run_args)
else:
logger.info(f'Processing')
run(args, run_args, world_size=args.world_size)
def main():
args = arguments.parse()
if args is None:
return
set_seed(args)
logger = initialize_logger(args)
logger.info(f'Arguments:\n{pformat(vars(args))}')
field, save_dict = None, None
if args.load is not None:
logger.info(f'Loading field from {os.path.join(args.save, args.load)}')
save_dict = torch.load(os.path.join(args.save, args.load))
field = save_dict['field']
field, train_sets, val_sets = prepare_data(args, field, logger)
run_args = (field, train_sets, val_sets, save_dict)
if len(args.gpus) > 1:
logger.info(f'Multiprocessing')
mp = Multiprocess(run, args)
mp.run(run_args)
else:
logger.info(f'Processing')
run(args, run_args, world_size=args.world_size)
def init_opt(args, model):
opt = None
if args.transformer_lr:
opt = torch.optim.Adam(model.params, betas=(0.9, 0.98), eps=1e-9)
else:
opt = torch.optim.Adam(model.params, betas=(args.beta0, 0.999))
return opt
if __name__ == '__main__':
args = arguments.parse()
set_seed(args)
logger = initialize_logger(args)
logger.info(f'Arguments:\n{pformat(vars(args))}')
field, save_dict = None, None
if args.load is not None:
logger.info(f'Loading field from {os.path.join(args.save, args.load)}')
save_dict = torch.load(os.path.join(args.save, args.load))
field = save_dict['field']
field, train_sets, val_sets = prepare_data(args, field, logger)
run_args = (field, train_sets, val_sets, save_dict)
if len(args.gpus) > 1:
logger.info(f'Multiprocessing')
mp = Multiprocess(run, args)
mp.run(run_args)
else:
logger.info(f'Processing')
run(args, run_args, world_size=args.world_size)
def init_opt(args, model):
opt = None
if args.transformer_lr:
opt = torch.optim.Adam(model.params, betas=(0.9, 0.98), eps=1e-9)
else:
opt = torch.optim.Adam(model.params, betas=(args.beta0, 0.999))
return opt
if __name__ == '__main__':
args = arguments.parse()
set_seed(args)
logger = initialize_logger(args)
logger.info(f'Arguments:\n{pformat(vars(args))}')
field, save_dict = None, None
if args.load is not None:
logger.info(f'Loading field from {os.path.join(args.save, args.load)}')
save_dict = torch.load(os.path.join(args.save, args.load))
field = save_dict['field']
field, train_sets, val_sets = prepare_data(args, field, logger)
run_args = (field, train_sets, val_sets, save_dict)
if len(args.gpus) > 1:
logger.info(f'Multiprocessing')
mp = Multiprocess(run, args)
mp.run(run_args)
else:
logger.info(f'Processing')
run(args, run_args, world_size=args.world_size)