chavinlo

/

alpaca-native

like 260

If I wanted to fine-tune a model like 13b with a different dataset, would I just point to the 13b folder with the consolidated.00.pth and consolidated.01.pth files? The thing I'm unsure about is whether it'll work with the model being split into two parts. Also, I read about needing hf converted model files. How do I do that?

Thanks.