NoiseShift learns a resolution-specific mapping from scheduler noise to conditioning noise via lightweight calibration to restore consistency and improve low-resolution generation quality in models like SD3 and Flux.
Flow matching models for learning reliable dynamics.arXiv preprint arXiv:2305.19591,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
NoiseShift: Resolution-Aware Noise Recalibration for Better Low-Resolution Image Generation
NoiseShift learns a resolution-specific mapping from scheduler noise to conditioning noise via lightweight calibration to restore consistency and improve low-resolution generation quality in models like SD3 and Flux.