3

So im looking for a way to load some model without acces to the class code. Is there a way? The only way to make such thing possible is to save the model with torch.jit.save/load but what if im not the person which training/designt the model?

With tensorflow im just used to use the load_model function. Is there something similary within pytorch?

Tollpatsch
  • 304
  • 4
  • 13
  • 1
    Is the model defined on a github repo? Then you may be able to use [torchhub](https://stackoverflow.com/a/66921920/9067615) to load the model. – iacob Apr 22 '21 at 12:36

0 Answers0