MadMarx37's picture
Trained with Unsloth
7ba2bb1 verified
metadata
language:
  - en
license: apache-2.0
tags:
  - code
  - unsloth
  - trl
  - sft

This is a fine-tune of Llama3-8B base model using an Alpaca-like instruction-tuning dataset generated with GPT4. My aim with doing this was to compare the instruction-tuned performance of this model with the one Llama has available.