phi_3_mini_instruct_marketing
#337
by
mangojesus
- opened
different source, different model
https://huggingface.co/Grey01/phi_3_mini_instruct_marketing
cool cool cool. you the man.
ok i think this one has a tokenizer
indeed, but it is already quantized, and llama.cpp does not support converting those models. there might be a way to load and save the model unquantized, but I am not a transformers expert.
mradermacher
changed discussion status to
closed
such nuance. haha