Search papers, labs, and topics across Lattice.
2
0
3
14
Forget end-to-end training for complex vision tasks; VisHarness shows that a lightweight agent can outperform task-specific models by learning to orchestrate a diverse set of pre-trained visual experts.
Turns out, you can get SOTA crowd instance segmentation by cleverly combining SAM with point supervision and reinforcement learning to select optimal points for mask generation.
