Fix the confusing example of gt_pred_pairs_of_highest_quality (#7730)

Co-authored-by: Nicolas Hug <[email protected]>
Co-authored-by: Nicolas Hug <[email protected]>

Author: fansuregrin

torchvision/models/detection/_utils.py

@@ -408,17 +408,9 @@ def set_low_quality_matches_(self, matches: Tensor, all_matches: Tensor, match_q
         # Find the highest quality match available, even if it is low, including ties
         gt_pred_pairs_of_highest_quality = torch.where(match_quality_matrix == highest_quality_foreach_gt[:, None])
         # Example gt_pred_pairs_of_highest_quality:
-        #   tensor([[    0, 39796],
-        #           [    1, 32055],
-        #           [    1, 32070],
-        #           [    2, 39190],
-        #           [    2, 40255],
-        #           [    3, 40390],
-        #           [    3, 41455],
-        #           [    4, 45470],
-        #           [    5, 45325],
-        #           [    5, 46390]])
-        # Each row is a (gt index, prediction index)
+        # (tensor([0, 1, 1, 2, 2, 3, 3, 4, 5, 5]),
+        #  tensor([39796, 32055, 32070, 39190, 40255, 40390, 41455, 45470, 45325, 46390]))
+        # Each element in the first tensor is a gt index, and each element in second tensor is a prediction index
         # Note how gt items 1, 2, 3, and 5 each have two ties
 
         pred_inds_to_update = gt_pred_pairs_of_highest_quality[1]