---
datasets:
- yuekai/belle_platypus_shargpt4
- yuekai/train_3.5M_CN_sampling_category_vicuna
- yuekai/Open-Platypus-Vicuna-Format
- yuekai/openchat_sharegpt_v3_vicuna_format
---

Total ~75k data, batch size 256

Eval results: https://huggingface.co/spaces/pe-nlp/mt-bench-zh