TRACE: Rollout-Guided Quantization-Aware Training for FP4 Reinforcement Learning of MoE Language Models
Paper • 2610.07767 • Published • 89
Daniel Wang
tuidan
AI & ML interests
Efficient AI / LLM / ML System
Recent Activity
upvoted a paper about 16 hours ago
TRACE: Rollout-Guided Quantization-Aware Training for FP4 Reinforcement Learning of MoE Language Models authored a paper about 20 hours ago
SVD-LLM V2: Optimizing Singular Value Truncation for Large Language Model Compression authored a paper about 20 hours ago
QUADS: Stabilizing NVFP4 Reinforcement Learning for MoE via QUantization-error Alignment across Dual SidesOrganizations
upvoted a paper about 16 hours ago