chore(model) optimize dynamic_k_matching with int (#861)

yolox/models/yolo_head.py

@@ -607,26 +607,27 @@ class YOLOXHead(nn.Module):
     def dynamic_k_matching(self, cost, pair_wise_ious, gt_classes, num_gt, fg_mask):
         # Dynamic K
         # ---------------------------------------------------------------
-        matching_matrix = torch.zeros_like(cost)
+        matching_matrix = torch.zeros_like(cost, dtype=torch.uint8)
 
         ious_in_boxes_matrix = pair_wise_ious
         n_candidate_k = min(10, ious_in_boxes_matrix.size(1))
         topk_ious, _ = torch.topk(ious_in_boxes_matrix, n_candidate_k, dim=1)
         dynamic_ks = torch.clamp(topk_ious.sum(1).int(), min=1)
+        dynamic_ks = dynamic_ks.tolist()
         for gt_idx in range(num_gt):
             _, pos_idx = torch.topk(
-                cost[gt_idx], k=dynamic_ks[gt_idx].item(), largest=False
+                cost[gt_idx], k=dynamic_ks[gt_idx], largest=False
             )
-            matching_matrix[gt_idx][pos_idx] = 1.0
+            matching_matrix[gt_idx][pos_idx] = 1
 
         del topk_ious, dynamic_ks, pos_idx
 
         anchor_matching_gt = matching_matrix.sum(0)
         if (anchor_matching_gt > 1).sum() > 0:
             _, cost_argmin = torch.min(cost[:, anchor_matching_gt > 1], dim=0)
-            matching_matrix[:, anchor_matching_gt > 1] *= 0.0
-            matching_matrix[cost_argmin, anchor_matching_gt > 1] = 1.0
-        fg_mask_inboxes = matching_matrix.sum(0) > 0.0
+            matching_matrix[:, anchor_matching_gt > 1] *= 0
+            matching_matrix[cost_argmin, anchor_matching_gt > 1] = 1
+        fg_mask_inboxes = matching_matrix.sum(0) > 0
         num_fg = fg_mask_inboxes.sum().item()
 
         fg_mask[fg_mask.clone()] = fg_mask_inboxes