Enhancing JEPAs with Spatial Conditioning: Robust and Efficient Representation Learning

Open in new window