There is an older "full" or "complete" technique to save a PyTorch model, but its name is rather misleading because it works just like the newer state_dict technique in the sense that code that uses ...