LLM Inference Acceleration via Efficient Operation Fusion

Re-assign community