---
language:
- en
base_model:
- Qwen/Qwen2.5-3B-Instruct
---

# Training details
Dataset : llama tokenizer 4000
distribution : 80% 13% 7%
only for attributes