Tutorial Notebook: Using the PEFT library with a DNA Language Model.

[rahulbshrestha]

This PR adds a tutorial notebook that uses the PEFT library with a DNA Language Model to solve a downstream task.

[BenjaminBossan]

Thanks, this looks good so far. Ping me once the notebook is ready for review.

For general context, check #1837.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[rahulbshrestha]

@BenjaminBossan this tutorial is ready for review!

[BenjaminBossan]

Thanks a lot for providing this detailed notebook on DNA-LM training with PEFT. I added a couple of comments, please check them out.

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[BenjaminBossan]

Thanks for the updates. There are only a few minor issues left.

Also, could you please run the whole notebook before committing it, so that the outputs are directly visible to users without having to run the notebook first.

[BenjaminBossan]

Okay, so let's add a comment that this is an optimization for speed, but for best results users should set this variable to None and use the whole dataset.

[rahulbshrestha]

@BenjaminBossan I just added the second review changes as you requested.

[BenjaminBossan]

Thanks for the updates, we're almost good. Just a few small things:

• Let's add a comment about the data size limit
• Let's not print the model so often, as it clutters the notebook and distracts from the important things
• It would be nice to have a short conclusion, e.g.: How much memory is saved with PEFT? If users want to improve the results, what could they test next?

[rahulbshrestha]

@BenjaminBossan I added changes from your latest review.

[BenjaminBossan]

Thanks for adding this notebook, the last changes LGTM.