Show notes
🤗 Upvotes: 22 | cs.AI
作者:
Donghoon Ahn, Runqian Wang, Alexei A. Efros, Taesung Park, Yash Belhe
标题:
Correction Training for Diffusion Models
Arxiv:
http://arxiv.org/abs/2610.14136v2
摘要:
Iterative sampling in diffusion models can lead to error accumulation due to exposure bias. The denoiser is trained on real images interpolated with noise, but at inference its inputs are constructed from its own imperfect predictions from the earlier steps, something it has never seen at training time. This might be one reason diffusion models often struggle without additional help, such as CFG. We…
Website: share.transistor.fm

