albertwang on Sinkhorn: Make LLMs even smaller through quantisation while maintaining accuracy · Hackerly

Comment on Sinkhorn: Make LLMs even smaller through quantisation while maintaining accuracy

albertwang 11mo

Media commentary: https://venturebeat.com/ai/huaweis-new-open-source-technique...