hardware

#4
by ziryeg - opened

What hardware do you use to train these models?

Only ablation was performed, no fine-tuning. The model card provides details. It works in a pure CPU environment and also in a pure GPU environment.

is it possible to convert it into gguf? i cant, because lama.cpp cant convert it.

You can wait and see if someone will handle this conversion. We currently don’t have time to deal with it. Whoever has time can take a look at this file convert_hf_to_gguf.py and write a separate conversion program.

afaik is an request open for lama.cpp to integrate it. maybe they implementing it, then can everyone convert it to gguf.

Sign up or log in to comment