Salvato in:
| Autori principali: | , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2507.13546 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866918096971759616 |
|---|---|
| author | Mikhailov, Dmitrii Letunovskiy, Aleksey Kovaleva, Maria Arkhipkin, Vladimir Korviakov, Vladimir Polovnikov, Vladimir Vasilev, Viacheslav Sidorova, Evelina Dimitrov, Denis |
| author_facet | Mikhailov, Dmitrii Letunovskiy, Aleksey Kovaleva, Maria Arkhipkin, Vladimir Korviakov, Vladimir Polovnikov, Vladimir Vasilev, Viacheslav Sidorova, Evelina Dimitrov, Denis |
| contents | Recent progress in transformer-based architectures has demonstrated remarkable success in video generation tasks. However, the quadratic complexity of full attention mechanisms remains a critical bottleneck, particularly for high-resolution and long-duration video sequences. In this paper, we propose NABLA, a novel Neighborhood Adaptive Block-Level Attention mechanism that dynamically adapts to sparsity patterns in video diffusion transformers (DiTs). By leveraging block-wise attention with adaptive sparsity-driven threshold, NABLA reduces computational overhead while preserving generative quality. Our method does not require custom low-level operator design and can be seamlessly integrated with PyTorch's Flex Attention operator. Experiments demonstrate that NABLA achieves up to 2.7x faster training and inference compared to baseline almost without compromising quantitative metrics (CLIP score, VBench score, human evaluation score) and visual quality drop. The code and model weights are available here: https://github.com/gen-ai-team/Wan2.1-NABLA |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_13546 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | $\nabla$NABLA: Neighborhood Adaptive Block-Level Attention Mikhailov, Dmitrii Letunovskiy, Aleksey Kovaleva, Maria Arkhipkin, Vladimir Korviakov, Vladimir Polovnikov, Vladimir Vasilev, Viacheslav Sidorova, Evelina Dimitrov, Denis Computer Vision and Pattern Recognition Recent progress in transformer-based architectures has demonstrated remarkable success in video generation tasks. However, the quadratic complexity of full attention mechanisms remains a critical bottleneck, particularly for high-resolution and long-duration video sequences. In this paper, we propose NABLA, a novel Neighborhood Adaptive Block-Level Attention mechanism that dynamically adapts to sparsity patterns in video diffusion transformers (DiTs). By leveraging block-wise attention with adaptive sparsity-driven threshold, NABLA reduces computational overhead while preserving generative quality. Our method does not require custom low-level operator design and can be seamlessly integrated with PyTorch's Flex Attention operator. Experiments demonstrate that NABLA achieves up to 2.7x faster training and inference compared to baseline almost without compromising quantitative metrics (CLIP score, VBench score, human evaluation score) and visual quality drop. The code and model weights are available here: https://github.com/gen-ai-team/Wan2.1-NABLA |
| title | $\nabla$NABLA: Neighborhood Adaptive Block-Level Attention |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2507.13546 |