This is a lora model with the base model https://huggingface.co/01-ai/Yi-34B
Pls use it by use_fast=False
The dataset I used is a mix of https://huggingface.co/datasets/QingyiSi/Alpaca-CoT/tree/main/Chain-of-Thought
Cuz I don't know the characteristics of this model, in order to avoid making mistakes, I temporarily set rank=8 and alpha=32. For futher version, I will turn it up to check the difference.
Enjoy it~