Search papers, labs, and topics across Lattice.
1
0
3
Object-centric LeJEPA achieves superior performance on key vision tasks while requiring significantly less training data than traditional image-level methods.
