there are two formats can be used when saving a trained network model, which are .pth and .mdl, respectively. I wonder what is the difference between these two formats? thanks.
I think the file extension doesn’t make a difference itself. The way you saved it actually matters.
As long as you use the builtin torch.save(), the saved file is in a deterministic format not matter what file extension you give it.
That is my understanding as well.