Is this supposed to be usable with llama.cpp?

#1
by spanielrassler - opened

Hi,

I'm wondering if I'm on the wrong track by trying to use this with llama.cpp. I get the following error when trying:
error loading model: unexpectedly reached end of file
llama_init_from_file: failed to load model
main: error: failed to load model

Any suggestions or is this model just not compatible? If not, do you know of a Pyg one that is? Thanks!

spanielrassler changed discussion status to closed

Since you've closed the issue, I imagine you've figured it out, but anyway.

No, you can only use this with my fork of ggml here or (preferably, as it's more advanced) with koboldcpp.

alpindale changed discussion status to open

I was wondering why do the weights depend on koboldcpp? How does the conversion differ?

Sign up or log in to comment