Comment out track_queries_placeholder_mask

src/trackformer/engine.py

@@ -26,24 +26,36 @@ def make_results(outputs, targets, postprocessors, tracking, return_only_orig=Tr
     orig_target_sizes = torch.stack([t["orig_size"] for t in targets], dim=0)
 
     # remove placeholder track queries
-    results_mask = None
-    if tracking:
-        results_mask = [~t['track_queries_placeholder_mask'] for t in targets]
-        for target, res_mask in zip(targets, results_mask):
-            target['track_queries_mask'] = target['track_queries_mask'][res_mask]
-            target['track_queries_fal_pos_mask'] = target['track_queries_fal_pos_mask'][res_mask]
+    # results_mask = None
+    # if tracking:
+    #     results_mask = [~t['track_queries_placeholder_mask'] for t in targets]
+    #     for target, res_mask in zip(targets, results_mask):
+    #         target['track_queries_mask'] = target['track_queries_mask'][res_mask]
+    #         target['track_queries_fal_pos_mask'] = target['track_queries_fal_pos_mask'][res_mask]
+
+    # results = None
+    # if not return_only_orig:
+    #     results = postprocessors['bbox'](outputs, target_sizes, results_mask)
+    # results_orig = postprocessors['bbox'](outputs, orig_target_sizes, results_mask)
+
+    # if 'segm' in postprocessors:
+    #     results_orig = postprocessors['segm'](
+    #         results_orig, outputs, orig_target_sizes, target_sizes, results_mask)
+    #     if not return_only_orig:
+    #         results = postprocessors['segm'](
+    #             results, outputs, target_sizes, target_sizes, results_mask)
 
     results = None
     if not return_only_orig:
-        results = postprocessors['bbox'](outputs, target_sizes, results_mask)
-    results_orig = postprocessors['bbox'](outputs, orig_target_sizes, results_mask)
+        results = postprocessors['bbox'](outputs, target_sizes)
+    results_orig = postprocessors['bbox'](outputs, orig_target_sizes)
 
     if 'segm' in postprocessors:
         results_orig = postprocessors['segm'](
-            results_orig, outputs, orig_target_sizes, target_sizes, results_mask)
+            results_orig, outputs, orig_target_sizes, target_sizes)
         if not return_only_orig:
             results = postprocessors['segm'](
-                results, outputs, target_sizes, target_sizes, results_mask)
+                results, outputs, target_sizes, target_sizes)
 
     if results is None:
         return results_orig, results

src/trackformer/models/detr.py

@@ -196,10 +196,10 @@ def loss_labels(self, outputs, targets, indices, _, log=True):
                     target_classes = target_classes.clone()
                     target_classes[i, target['track_queries_fal_pos_mask']] = 0
 
-        weight = None
-        if self.tracking:
-            weight = torch.stack([~t['track_queries_placeholder_mask'] for t in targets]).float()
-            loss_ce *= weight
+        # weight = None
+        # if self.tracking:
+        #     weight = torch.stack([~t['track_queries_placeholder_mask'] for t in targets]).float()
+        #     loss_ce *= weight
 
         loss_ce = loss_ce.sum() / self.empty_weight[target_classes].sum()
 
@@ -229,20 +229,22 @@ def loss_labels_focal(self, outputs, targets, indices, num_boxes, log=True):
 
         target_classes_onehot = target_classes_onehot[:,:,:-1]
 
-        query_mask = None
-        if self.tracking:
-            query_mask = torch.stack([~t['track_queries_placeholder_mask'] for t in targets])[..., None]
-            query_mask = query_mask.repeat(1, 1, self.num_classes)
+        # query_mask = None
+        # if self.tracking:
+        #     query_mask = torch.stack([~t['track_queries_placeholder_mask'] for t in targets])[..., None]
+        #     query_mask = query_mask.repeat(1, 1, self.num_classes)
 
         loss_ce = sigmoid_focal_loss(
             src_logits, target_classes_onehot, num_boxes,
-            alpha=self.focal_alpha, gamma=self.focal_gamma, query_mask=query_mask)
-
-        if self.tracking:
-            mean_num_queries = torch.tensor([len(m.nonzero()) for m in query_mask]).float().mean()
-            loss_ce *= mean_num_queries
-        else:
-            loss_ce *= src_logits.shape[1]
+            alpha=self.focal_alpha, gamma=self.focal_gamma)
+            # , query_mask=query_mask)
+
+        # if self.tracking:
+        #     mean_num_queries = torch.tensor([len(m.nonzero()) for m in query_mask]).float().mean()
+        #     loss_ce *= mean_num_queries
+        # else:
+        #     loss_ce *= src_logits.shape[1]
+        loss_ce *= src_logits.shape[1]
         losses = {'loss_ce': loss_ce}
 
         if log:

src/trackformer/models/detr_tracking.py

@@ -183,38 +183,38 @@ def add_track_queries_to_targets(self, targets, prev_indices, prev_out, add_fals
             ]).bool()
 
         # add placeholder track queries to allow for batch sizes > 1
-        max_track_query_hs_embeds = max([len(t['track_query_hs_embeds']) for t in targets])
-        for i, target in enumerate(targets):
+        # max_track_query_hs_embeds = max([len(t['track_query_hs_embeds']) for t in targets])
+        # for i, target in enumerate(targets):
 
-            num_add = max_track_query_hs_embeds - len(target['track_query_hs_embeds'])
+        #     num_add = max_track_query_hs_embeds - len(target['track_query_hs_embeds'])
 
-            if not num_add:
-                target['track_queries_placeholder_mask'] = torch.zeros_like(target['track_queries_mask']).bool()
-                continue
+        #     if not num_add:
+        #         target['track_queries_placeholder_mask'] = torch.zeros_like(target['track_queries_mask']).bool()
+        #         continue
 
-            raise NotImplementedError
+        #     raise NotImplementedError
 
-            target['track_query_hs_embeds'] = torch.cat(
-                [torch.zeros(num_add, self.hidden_dim).to(device),
-                 target['track_query_hs_embeds']
-            ])
-            target['track_query_boxes'] = torch.cat(
-                [torch.zeros(num_add, 4).to(device),
-                 target['track_query_boxes']
-            ])
+        #     target['track_query_hs_embeds'] = torch.cat(
+        #         [torch.zeros(num_add, self.hidden_dim).to(device),
+        #          target['track_query_hs_embeds']
+        #     ])
+        #     target['track_query_boxes'] = torch.cat(
+        #         [torch.zeros(num_add, 4).to(device),
+        #          target['track_query_boxes']
+        #     ])
 
-            target['track_queries_mask'] = torch.cat([
-                torch.tensor([True, ] * num_add).to(device),
-                target['track_queries_mask']
-            ]).bool()
+        #     target['track_queries_mask'] = torch.cat([
+        #         torch.tensor([True, ] * num_add).to(device),
+        #         target['track_queries_mask']
+        #     ]).bool()
 
-            target['track_queries_fal_pos_mask'] = torch.cat([
-                torch.tensor([False, ] * num_add).to(device),
-                target['track_queries_fal_pos_mask']
-            ]).bool()
+        #     target['track_queries_fal_pos_mask'] = torch.cat([
+        #         torch.tensor([False, ] * num_add).to(device),
+        #         target['track_queries_fal_pos_mask']
+        #     ]).bool()
 
-            target['track_queries_placeholder_mask'] = torch.zeros_like(target['track_queries_mask']).bool()
-            target['track_queries_placeholder_mask'][:num_add] = True
+        #     target['track_queries_placeholder_mask'] = torch.zeros_like(target['track_queries_mask']).bool()
+        #     target['track_queries_placeholder_mask'][:num_add] = True
 
     def forward(self, samples: NestedTensor, targets: list = None, prev_features=None):
         if targets is not None and not self._tracking:
@@ -266,7 +266,7 @@ def forward(self, samples: NestedTensor, targets: list = None, prev_features=Non
                     device = target['boxes'].device
 
                     target['track_query_hs_embeds'] = torch.zeros(0, self.hidden_dim).float().to(device)
-                    target['track_queries_placeholder_mask'] = torch.zeros(self.num_queries).bool().to(device)
+                    # target['track_queries_placeholder_mask'] = torch.zeros(self.num_queries).bool().to(device)
                     target['track_queries_mask'] = torch.zeros(self.num_queries).bool().to(device)
                     target['track_queries_fal_pos_mask'] = torch.zeros(self.num_queries).bool().to(device)
                     target['track_query_boxes'] = torch.zeros(0, 4).to(device)

src/trackformer/models/matcher.py

@@ -111,7 +111,8 @@ def forward(self, outputs, targets):
 
             prop_i = 0
             for j in range(cost_matrix.shape[1]):
-                if target['track_queries_fal_pos_mask'][j] or target['track_queries_placeholder_mask'][j]:
+                # if target['track_queries_fal_pos_mask'][j] or target['track_queries_placeholder_mask'][j]:
+                if target['track_queries_fal_pos_mask'][j]:
                     # false positive and palceholder track queries should not
                     # be matched to any target
                     cost_matrix[i, j] = np.inf