13 papers · 1 filter
The FID Lottery: Quantifying Hidden Randomness in Generative-Model Evaluation
Nicolas Dufour, Alexei A. Efros, Patrick Pérez
The Frechet Inception Distance (FID) is the de facto arbiter of image generation, yet most papers report just a single number from a single trained model using a single sampling se…
Surflo: Consistent 3D Surface Flow Model with Global State
Antoine Guédon, Shu Nakamura, Nicolas Dufour +3
Geometry is invariant to viewpoint, which makes any collection of images a redundant encoding of a single 3D state. Existing feed-forward reconstruction models fail to exploit this…
MIRO: MultI-Reward cOnditioned pretraining improves T2I quality and efficiency
Nicolas Dufour, Lucas Degeorge, Arijit Ghosh +2
The default paradigm of post-training text-to-image generators includes post-hoc selection of generated images, and subsequent training with one reward model to align the generator…
One View Is Enough! Monocular Training for In-the-Wild Novel View Generation
Adrien Ramanana Rahary, Nicolas Dufour, Patrick Perez +1
Monocular novel-view synthesis has long required multi-view image pairs for supervision, limiting training data scale and diversity. We argue it is not necessary: one view is enoug…
PoM: A Linear-Time Replacement for Attention with the Polynomial Mixer
David Picard, Nicolas Dufour, Lucas Degeorge +14
This paper introduces the Polynomial Mixer (PoM), a novel token mixing mechanism with linear complexity that serves as a drop-in replacement for self-attention. PoM aggregates inpu…
How far can we go with ImageNet for Text-to-Image generation?
L. Degeorge, A. Ghosh, N. Dufour +2
Recent text-to-image (T2I) generation models have achieved remarkable sucess by training on billion-scale datasets, following a `bigger is better' paradigm that prioritizes data qu…