ByNobleID
    LoftUp: Learning a Coordinate-Based Feature Upsampler for Vision Foundation Models | NobleID