You could use that since this si a PyTorch model. Since you are using the Hugging Face library you can save the model using `model.save_pretrained(path)` and the tokenizer using `tokenizer.save_pretrained(path)`. Also is always good practice to save your training arguments with `torch.save()`.