llm-compressor recipe

#1
by bullerwins - opened

Hi!
Could you share the llm-compressor recipe to make this or the 4 bit?

Did you leave the vision layers unquantized?

cyankiwi org

Yes, the vision layers are unquantized. You can visit this model full recipe in this repo recipe.yaml.

bullerwins changed discussion status to closed

Sign up or log in to comment