PapersBuilders19:09Chinese contentIDRF:掩码离散扩散模型的逆蒸馏奖励微调框架一篇扩散模型微调的论文,教你怎么给少步掩码扩散模型做奖励微调,还能少 32 倍去噪步数,做生成方向的可以看看。#IDRF#diffusion models#policy gradient#open-access papersaarXiv cs.LG@Vladislav Gromadskii et al. (7)SourceRead laterReadWorth followingUsefulFollow IDRF