Except with warning

segment-anything-2/sam2/sam2_image_predictor.py

@@ -377,6 +377,13 @@ class SAM2ImagePredictor:
             # we merge "boxes" and "points" into a single "concat_points" input (where
             # boxes are added at the beginning) to sam_prompt_encoder
             if concat_points is not None:
+                if concat_points[0].size(1) > 1 or concat_points[1].size(1) > 1:
+                    print("Warning: Box and point combination only works if there's "
+                          "only one dot and one box. Using only the first one...")
+                    concat_points = (concat_points[0][:, :1, :], concat_points[1][:, :1])
+                    box_labels = box_labels[:1]
+                    box_coords = box_coords[:1]
+
                 concat_coords = torch.cat([box_coords, concat_points[0]], dim=1)
                 concat_labels = torch.cat([box_labels, concat_points[1]], dim=1)
                 concat_points = (concat_coords, concat_labels)