wAR-Tok adds a Wasserstein-gradient-flow prior-matching term to tokenizer training so that discrete tokens become easier for autoregressive priors to model, cutting AR loss and raising generation FID on CIFAR-10 and ImageNet while keeping reconstruction quality comparable.
Goodfellow and Jean Pouget-Abadie and Mehdi Mirza and Bing Xu and David Warde-Farley and Sherjil Ozair and Aaron Courville and Yoshua Bengio , title =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Learning Discrete Autoregressive Priors with Wasserstein Gradient Flow
wAR-Tok adds a Wasserstein-gradient-flow prior-matching term to tokenizer training so that discrete tokens become easier for autoregressive priors to model, cutting AR loss and raising generation FID on CIFAR-10 and ImageNet while keeping reconstruction quality comparable.