Considering LLAMA 3 or LLAMA 3.1 as base model

[Chuanqi-Zang]

Would you consider training a model with the latest version of LLAMA 3/LLAMA 3.1 in the near future? I'm interested in the performance and its weight.

[SY-Xuan]

Sorry. I do not have enough GPUs to do this now. If you have the GPU resources or you can do this, it is welcome to give some pull requests.