I tried from_pretrained and it did not work
You would need to convert the weights first and provide the relevant json files, like this person did: https://huggingface.co/nickypro/tinyllama-15M
is there a Q8 model ? for use in llama2-c
· Sign up or log in to comment