Trying to make BF16 / GGUF

#1
by Phr00t - opened

Any trick to getting this into a GGUF?

Would you be able to provide a BF16 GGUF (then I can make quants myself)?

Getting this error when I try it myself: https://github.com/ggml-org/llama.cpp/issues/18065

Yeah, I will need some file from original model here, it seems pushing directly from heretic doesn’t adds them.

I used transformers 5.0 and that seems to be causing this, I will update the model with transformer 4.x compatibility.

Sign up or log in to comment