TRACE: Rollout-Guided Quantization-Aware Training for FP4 Reinforcement Learning of MoE Language Models
Paper • 2610.07767 • Published • 94
yu
yuhao318
CaynMarlaMiller2025's profile picture
gabrielemidulla's profile picture
Ujjwal-Tyagi's profile picture·
AI & ML interests
None yet
Recent Activity
authored a paper about 10 hours ago
VLCache: Computing 2% Vision Tokens and Reusing 98% for Vision-Language Inference authored a paper about 10 hours ago
Unified Low-rank Compression Framework for Click-through Rate Prediction authored a paper about 10 hours ago
H-Scale: Hessian-Guided Scale Refinement for NVFP4 Sub-Byte LLM InferenceOrganizations
