Search papers, labs, and topics across Lattice.
2
0
4
8
Current audio-visual models nail unimodal quality but still struggle to make music and dance move together rhythmically, highlighting a key gap TMD-Bench is designed to address.
Turns out, your image-generating diffusion model already knows how to segment anything you ask it to.
