NABLA introduces a Neighborhood Adaptive Block-Level Attention mechanism for video diffusion transformers, enhancing efficiency and speed while maintaining quality, achieving up to 2.7 times faster training and inference.

https://arxiv.org/abs//2507.13546

YouTube: https://www.youtube.com/@ArxivPapers

TikTok: https://www.tiktok.com/@arxiv_papers

Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016

Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers