How to convert LLMA with hugging face format to ggml model?

[xiejibing]

the script convert-pth-to-ggml.py provides the ability to convert llama model (faire scale format) to ggml model，but how to convert a llama model with hugging face format to the ggml model?

[prusnak]

What do you mean by hugging face format? Do you mean safetensors?

There is a universal convert script brewing here - you might want to try it: #545

[prusnak]

try the new convert.py script that is now in master