Vsa: Faster video diffusion with trainable sparse attention
Published in arXiv preprint arXiv:2505.13389, 2025
Trainable sparse attention for faster and more efficient video diffusion.
Recommended citation: Zhang, P., Chen, Y., Huang, H., Lin, W., Liu, Z., Stoica, I., Xing, E., and Zhang, H. (2025). "Vsa: Faster video diffusion with trainable sparse attention." arXiv preprint arXiv:2505.13389.
Download Paper
