Modelpedia
Work in progress
About
Findings
Models
Concepts
Methods
Datasets
Sources
Light
Dark
Elucidating the Exposure Bias in Diffusion Models
2024-01-16
· ICLR 2024 poster ·
anchor
Findings
IC-1579
ADM's noise prediction network exhibits exposure bias: during iterative sampling the l2-norm of its ε prediction is systematically larger than during training, and the sampling distribution variance exceeds the training variance with error accumulating toward the end of the chain