---
tags:
- pytorch
- trl
- sft
inference: false
---

omost-llama-3-8b-4bits is Omost's llama-3 model with 8k context length in nf4.