Multi-Bin Batching for Increasing LLM Inference Throughput

LRM
Re-assign community