Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.
And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.
Save Model Locally trainer.save_model()
See original GitHub issueFirst off, super excited about all this work! Also really appreciate the youtube presentation and the blogpost, super insightful to read through. I went through the existing issues and couldn’t find whether this an existing functionality:
Is there a way to save the model locally instead of pushing to the hub?
So in addition to this:
trainer.push_to_hub(f"setfit-myowndataset-multilabel-example")
be able to do this too:
trainer.save_model("file-path/setfit-myowndatase-multilabel-example")
Issue Analytics
- State:
- Created 10 months ago
- Reactions:4
- Comments:5
Top Results From Across the Web
Top Related Medium Post
Top Related StackOverflow Question
Troubleshoot Live Code
Top Related Reddit Thread
Top Related Hackernoon Post
Top Related Tweet
Top Related Dev.to Post
Top Related Hashnode Post
It can be done by trainer.model._save_pretrained(save_directory=save_directory).
Here is the link to the solution where it has already been answered - https://github.com/huggingface/setfit/issues/102
You should be able to do
model.save(path_str)as usual.https://github.com/huggingface/setfit/blob/5313939b638b57668a1b5a67ef600e54da221d9a/src/setfit/modeling.py#L579
or if your model is not wrapped,
model._save_pretrained(path_str).