If I have the baseline Mixtral-8x7B .pth, how do I convert that to a llamafile compatible .gguf?
Just 4 days ago he redid a model after the author updated the EOS token: https://huggingface.co/TheBloke/openchat-3.5-1210-GGUF/discu...
Don't suppose if you know if the conversion is easily reversible? Some of these models are big, it sucks to carry around the original plus the gguf, but I would hate to be in a situation where the gguf represents a dead end.