Accelerating MoE Model Inference with Expert Sharding

MoE
Re-assign community