Hi! the best and safe way to save your model parameters is doing something like this:
model = MyModel()
# ... after training, save your model
model.save_state_dict('mytraining.pt')
# .. to load your previously training model:
model.load_state_dict(torch.load('mytraining.pt'))