Fetching the paper…

Exploiting Inter-Layer Expert Affinity for Accelerating Mixture-of-Experts Model Inference · Around