Challenging GPU Dominance: When CPUs Outperform for On-Device LLM Inference

Re-assign community