Search papers, labs, and topics across Lattice.
2
0
3
DINOv2's representation space is so statistically well-behaved that you can train a vanilla diffusion transformer on it and beat specialized architectures with fewer parameters.
Ditch slow diffusion models for medical image segmentation: MedFlowSeg achieves state-of-the-art accuracy with a computationally efficient, one-step deterministic approach.
