How to fine tune a model in pytorch?

[dafmdev]

I have reviewed the examples of fine tuning using paxml, are there examples of how to perform fine tuning in PyTorch?

Using this model: google/timesfm-1.0-200m-pytorch

[siriuz42]

You'd need a training data loader and a training loop (backward pass). We don't have a notebook for now - hopefully there are plenty of PyTorch experts in the community who have examples.

[dafmdev]

Thank you very much for your reply. Do you have a guide on how you converted the pax model to torch or a code example?