analiza aiqbench
This research proposes Next-Embedding Predictive Autoregression (NEPA), a method where diffusion transformers dynamically predict conditioning embeddings at each denoising step rather than reusing a static prompt. By adapting the conditioning signal to the current noisy state, NEPA-DiT-XL achieves an FID of 1.32 on ImageNet using roughly one-third of the training compute required by previous methods like REPA. The approach demonstrates that autoregressive prediction of continuous embeddings can significantly improve generation efficiency without sacrificing fidelity.
Embedding Prediction Helps Image Generation
arXiv:2610.02203v1 Announce Type: cross Abstract: In diffusion transformers, a class label or a text prompt is embedded once, and the same condition is reused at every denoising step. We ask whether predicted embeddings can serve as this condition instead. Next-Embedding Predictive Autoregression (NEPA) trains a Transformer to predict the next continuous embedding in a sequence. In generation, the clean image follows the noisy image, so its embeddings are the next embeddings after the condition and the noisy image. We train a NEPA model to predict them all at once with Multi-Embedding Prediction, and in Embedding Conditioned Generation, a DiT generator is conditioned on these predictions, recomputed at every denoising step, so the conditioning signal adapts to the current noisy state. Experiments on class-conditional ImageNet $256\times256$ study the condition of the generator, the design of Multi-Embedding Prediction, and the scaling of both models. The NEPA model adds a second network to every sampling step; with it, and combined with REPA, our final model, NEPA-DiT-XL, reaches an FID of 1.32 using about a third of the training compute of REPA.
Abstract:In diffusion transformers, a class label or a text prompt is embedded once, and the same condition is reused at every denoising step. We ask whether predicted embeddings can serve as this condition instead. Next-Embedding Predictive Autoregression (NEPA) trains a Transformer to predict the next continuous embedding in a sequence. In generation, the clean image follows the noisy image, so its embeddings are the next embeddings after the condition and the noisy image. We train a NEPA model to predict them all at once with Multi-Embedding Prediction, and in Embedding Conditioned Generation, a DiT generator is conditioned on these predictions, recomputed at every denoising step, so the conditioning signal adapts to the current noisy state. Experiments on class-conditional ImageNet $256\times256$ study the condition of the generator, the design of Multi-Embedding Prediction, and the scaling of both models. The NEPA model adds a second network to every sampling step; with it, and combined with REPA, our final model, NEPA-DiT-XL, reaches an FID of 1.32 using about a third of the training compute of REPA.
Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
arXiv:2610.02203 [cs.CV]
(or arXiv:2610.02203v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2610.02203 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 273 cuvinte
Subiecte conexe
Forbes Asia 100 To Watch 2026
forbes.com · 24 aug. 2026
[AINews] Gemini 4 Argon: GDM’s answer to Astra/Fable, with 1M output
Latent Space · 1 oct. 2026
Deploy Agentic-Ready AI at the Edge with Memory Efficiency in NVIDIA JetPack 7.2
NVIDIA · 2 iun. 2026
Modernizing and scaling support operations with generative AI on AWS
AWS · 2 sept. 2026