Mirai injects future-token foresight into autoregressive visual generators, accelerating convergence up to 10x and cutting ImageNet FID from 5.34 to 4.34.
Return of unconditional generation: A self-supervised representation generation method
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
representative citing papers
Representations learned by large AI models are converging toward a shared statistical model of reality.
citing papers explorer
-
Mirai: Autoregressive Visual Generation Needs Foresight
Mirai injects future-token foresight into autoregressive visual generators, accelerating convergence up to 10x and cutting ImageNet FID from 5.34 to 4.34.
-
The Platonic Representation Hypothesis
Representations learned by large AI models are converging toward a shared statistical model of reality.