Fine-tuning based on THUDM/chatglm-6b. The first Fine-tuning method is LoRA, and more methods will be made available if possible.
Please checkout the Tutorial if you can read Chinese.
An AI-generated QA dataset about AI explorer (AI探险家) is provided. The trained model saved in the output folder is also based on this dataset.
