vis utility function visualize() support segm mask prediction

Summary:
In bento N5675106 **visualize_fusion_models.py**, we'd like to support the visualization and side-by-side comparison of egobody mask predictions from different Fusion MT models. This diff contains the required changes.
- For visualize() function, add a new argument **show_segm**

Differential Revision: D77811204

fbshipit-source-id: 9d98dd004d9321cebdf5779e59f594af0190ee6a

Author: stephenyan1231

detectron2/utils/visualizer.py

@@ -451,7 +451,7 @@ def draw_sem_seg(
         area_threshold=None,
         alpha=0.8,
         draw_text: bool = True,
-        edge_color: tuple[float, float, float] = _OFF_WHITE,
+        edge_color: tuple[float, float, float] | None = _OFF_WHITE,
     ):
         """
         Draw semantic segmentation predictions/labels.