1 paper
Bhishma Dedhia, David Bourgin, Krishna Kumar Singh +5
Diffusion Transformers (DiTs) can generate short photorealistic videos, yet directly training and sampling longer videos with full attention across the video remains computationall…