Search papers, labs, and topics across Lattice.
1
5
3
2
Achieve 98% accuracy in maritime scene recognition by fusing image data, text descriptions, and MLLM outputs, then compress the model to 68MB with minimal performance loss.
