architecture · first seen 10 Oct 2026
video variational autoencoder
Also: video VAE
An autoencoder that compresses video into latent representations, used in latent diffusion video generation pipelines.
Mentions per week
PapersCodeDiscussion
Earliest sources
- LightVAE: Towards Compact and Efficient Video AutoencodersPapers · 10 Oct 2026 · Shiqiao Gu, Yang Yong, Fengjun Zhong, Ruihao Gong, Chengtao Lv, Xiangyu Fan, Lei…
Latest sources
- LightVAE: Towards Compact and Efficient Video AutoencodersPapers · 10 Oct 2026 · Shiqiao Gu, Yang Yong, Fengjun Zhong, Ruihao Gong, Chengtao Lv, Xiangyu Fan, Lei…