Search papers, labs, and topics across Lattice.
VCIP, CS, Nankai University
1
0
3
Achieve up to 2.5x faster video object removal with comparable visual quality by intelligently selecting only the essential tokens for processing in Diffusion Transformers.