Quantization Method / GGML quantisation type

#41
by multiverse - opened

How do I know what to choose from the parameters in this setting?

Maybe someone can give me an answer or a link because I can't find anything about it.

@mishig can share an awesome link he created recently

multiverse changed discussion status to closed
ggml.ai org

@julien-c @mishig Maybe we can also see the labelled file type in the GGUF visualizer: "general.file_type"->12->Q3_K_M.
Example here phymbert/dbrx-16x12b-instruct-q3_k_m-gguf

^yes ๐Ÿ‘

Like choosing which sandpaper to use. :)

Sign up or log in to comment