if args.clip > 0: torch.nn.utils.clip_grad_norm_(model.parameters(), args.clip) hi everybody! I just want to know what role dose the above block play? Why is the args.clip default 0 ?
