Search papers, labs, and topics across Lattice.
Affiliation:
2
0
3
Query-grounded visual sampling can boost long video understanding in LVLMs by over 11% without the need for extensive training.
