---
library_name: peft
---
## Training procedure

### Framework versions


- PEFT 0.4.0

## To Use:
Begin the prompt for this model with
"Write a Twitch.tv chat response to this message: "
followed by the message you'd like the model to generate a Twitch.tv Chat response for.

### About:
Llama2 was finetuned with [this](https://huggingface.co/datasets/BinghamtonUniversity/cs415-twitch-chats-llama2-51k) dataset, which I curated in cs415 (Social Media Data Sci. Pipelines) with Professor Jeremy Blackburn.

I finetuned this in a Colab Notebook with an L4, which can be found [here](https://colab.research.google.com/drive/17TBqcrZjHBC05wgSxor0NkJXMG4VEGeD?usp=sharing)