LocalLLaMA

LM Studio CPU thread pool size vs. tk/s with some MoE layers offloaded to CPU

/

submitted by /u/bonobomaster [link] [comments]