about the training strategy for Llama-3-8B

Question

about the training strategy for Llama-3-8B

Closed this issue 5 months ago · 2 comments

Hi, thanks for your work!
You mentioned that "we used better strategies to train Phi-3-Mini-based and Llama-3-8B-based Bunny", so I would like to ask, what kind of strategy did you use when training Llama-3-8B-based Bunny? And when do you plan to make the finetune_lora.sh of Llama-3-8B-based Bunny public?

Answer 1 · 2024-07-22T16:30:07.000Z

All of the training strategy and data of latest Bunny is released! Check more details about Bunny in Technical Report, Data and Training Tutorial!

Answer 2 · 2024-08-06T03:30:32.000Z

Close the issue for now if there's no further discussions. Feel free to reopen it if there's any other questions.