Search papers, labs, and topics across Lattice.
1
0
2
5
ViT-Up achieves up to 8.09% improvement in dense prediction accuracy by eliminating reliance on external image guidance in Vision Transformers.
