Unweight: how we compressed an LLM 22% without sacrificing quality https://blog.cloudflare.com/unweight-tensor-compression/ #Ai #Programming #Infrastructure