News for hackers. Stuff that matters.

A comment on GPU Embedding with GGML. Parent comment

Re:GPU Embedding with GGML by moffkalast on Monday October 16, 2023 @11:21AM (#37901077)
I think you're still missing AWQ ones, which are a sort of GPTQ but with dynamic quantization depending on weight importance iirc?
[ Parent | on HN ]

There are no comments on this comment yet.