Search papers, labs, and topics across Lattice.
Affiliation:
1
0
Bypassing synthetic data, MLLMCLIP achieves state-of-the-art compositional accuracy in vision-language tasks through innovative feature-level distillation.
