4 ms·Sam 3: a unified foundation model for promptable segmentation in images/videos1 points by mooreds 9mo ago