Search papers, labs, and topics across Lattice.
1
0
2
8
HPSD enables TI2V models to internalize high-quality visual cues, resulting in a remarkable boost in text-to-video performance while simultaneously enhancing image-to-video generation.
