Paper

LLaCO: Continuous Diffusion for Large Language Models with Competition-Aware Semantic Scheduling · ByteTrail