我爱计算机视觉 · 公开公众号 RSS·· 2026-10-15AI score31
HKUST × Tencent Video propose Adaptive Reward Routing: audio-video diffusion model post-training no longer fixes reward routing in advance
港科大×腾讯视频:音视频大模型后训练往哪儿使劲?边学边改,不再提前定死
AI briefing
MMLab@HKUST of HKUST, Tencent Video, and the University of Hong Kong propose Adaptive Reward Routing, which, on the DiffusionNFT framework, recomputes multi-reward routing for audio-video diffusion models in real time during training rather than fixing it once from the base model.
Open original mp.weixin.qq.com