Add load from LoRA on HF notebook

[rawkintrevo]

PR Checklist

Notebook requirements

• Notebook outputs cleared: Ensure all notebook outputs are cleared.
• Test in Google Colab:
  • Test that it works in Google Colab (Python 3.10.12).
  • Colab has its own package set and Python version, so ensure compatibility.

Incoming References

• README.md updates:
  • Add a link to the recipe in the Table of Contents (ToC).
  • Include a Colab button after that link.

GitHub

• Commits signed: All commits must be GPG or SSH signed.
• DCO Compliance: Developer Certificate of Origin (DCO) applies to the code, documentation, and any example data provided. Ensure commits are signed off.

[adampingel]

We should group this recipe with the fine-tuning one in the ToC, but I can do that in a separate PR. (There's an existing story that I own covering this.)

[adampingel]

Related to last comment: Let's adjust the intro language "We now demonstrate how to load a model back for inference." by adding a link to the previous recipe and reminding the reader of the context.

[adampingel]

This recipe is testable. I happened to grant access to my HF_TOKEN during my run-through just now. Is that necessary? If so, we can get in available as an org-level secret.

I'm going to merge this as-is, but let's address the points in these 3 comments soon.

[rawkintrevo]

Agree on the cross links, my thought was get this one in, then link from the training notebook to this one, and this one back to the training one. The latter I could have done in this PR. (probably on the code one too).

On testable: I don't know if it is? In the test it will download and try to run the notebook. I don't know how that will work in a CPU only environment.