分享

Using Human Feedback to Fine-tune Diffusion Models without Any Reward Model

10
查看论文
热度