Search papers, labs, and topics across Lattice.
1
0
3
2
Encoding temporal prediction into video VAEs unlocks faster training, better generative performance, and improved downstream task performance, all at once.
