notebook to demonstrate Gemma/Gemini evaluation

deep-diver / llamaduo

This project showcases an LLMOps pipeline that fine-tunes a small-size LLM model to prepare for the outage of the service LLM.

https://huggingface.co/papers/2408.13467

Apache License 2.0

286 stars 29 forks source link

Closed deep-diver closed 6 months ago

review-notebook-app[bot] commented 6 months ago

Check out this pull request on

See visual diffs & provide feedback on Jupyter Notebooks.

Powered by ReviewNB

sayakpaul commented 6 months ago

Thanks for the answers.

In tokenizer.decode() we should add tokenizer.decode(skip_special_tokens=True) as well.

deep-diver commented 6 months ago

@sayakpaul

Addressed your comments. Specifically:

added skip_special_tokens=True in decode(), then split by the delimiter which is assistant\n in this case.
removed .to(model.device) after the apply_chat_template() as you suggested!

sayakpaul commented 6 months ago

Thanks! The first resolution should help us get rid of the manual eos splitting we were doing.

Feel free to merge this!