/Sequence-Scheduling

PyTorch implementation of paper "Response Length Perception and Sequence Scheduling: An LLM-Empowered LLM Inference Pipeline".

Primary LanguagePython

Watchers