mv data_time after .cuda()

train_img_model_cent.py

@@ -208,11 +208,12 @@ def train(epoch, model, criterion_xent, criterion_cent, optimizer_model, optimiz
 
     end = time.time()
     for batch_idx, (imgs, pids, _) in enumerate(trainloader):
+        if use_gpu:
+            imgs, pids = imgs.cuda(), pids.cuda()
+        
         # measure data loading time
         data_time.update(time.time() - end)
 
-        if use_gpu:
-            imgs, pids = imgs.cuda(), pids.cuda()
         outputs, features = model(imgs)
         xentloss = criterion_xent(outputs, pids)
         centloss = criterion_cent(features, pids) * args.weight_cent

train_img_model_ring.py

@@ -208,11 +208,12 @@ def train(epoch, model, criterion_xent, criterion_ring, optimizer, trainloader,
 
     end = time.time()
     for batch_idx, (imgs, pids, _) in enumerate(trainloader):
-        # measure data loading time
-        data_time.update(time.time() - end)
-
         if use_gpu:
             imgs, pids = imgs.cuda(), pids.cuda()
+
+        # measure data loading time
+        data_time.update(time.time() - end)
+        
         outputs, features = model(imgs)
         xentloss = criterion_xent(outputs, pids)
         ringloss = criterion_ring(features)

train_img_model_xent.py

@@ -203,11 +203,12 @@ def train(epoch, model, criterion, optimizer, trainloader, use_gpu):
 
     end = time.time()
     for batch_idx, (imgs, pids, _) in enumerate(trainloader):
-        # measure data loading time
-        data_time.update(time.time() - end)
-
         if use_gpu:
             imgs, pids = imgs.cuda(), pids.cuda()
+
+        # measure data loading time
+        data_time.update(time.time() - end)
+        
         outputs = model(imgs)
         if isinstance(outputs, tuple):
             loss = DeepSupervision(criterion, outputs, pids)

train_img_model_xent_htri.py

@@ -212,11 +212,12 @@ def train(epoch, model, criterion_xent, criterion_htri, optimizer, trainloader,
 
     end = time.time()
     for batch_idx, (imgs, pids, _) in enumerate(trainloader):
-        # measure data loading time
-        data_time.update(time.time() - end)
-
         if use_gpu:
             imgs, pids = imgs.cuda(), pids.cuda()
+
+        # measure data loading time
+        data_time.update(time.time() - end)
+        
         outputs, features = model(imgs)
         if args.htri_only:
             if isinstance(features, tuple):

train_vid_model_xent.py

@@ -200,11 +200,12 @@ def train(epoch, model, criterion, optimizer, trainloader, use_gpu):
 
     end = time.time()
     for batch_idx, (imgs, pids, _) in enumerate(trainloader):
-        # measure data loading time
-        data_time.update(time.time() - end)
-
         if use_gpu:
             imgs, pids = imgs.cuda(), pids.cuda()
+
+        # measure data loading time
+        data_time.update(time.time() - end)
+        
         outputs = model(imgs)
         loss = criterion(outputs, pids)
         optimizer.zero_grad()

train_vid_model_xent_htri.py

@@ -209,11 +209,12 @@ def train(epoch, model, criterion_xent, criterion_htri, optimizer, trainloader,
 
     end = time.time()
     for batch_idx, (imgs, pids, _) in enumerate(trainloader):
-        # measure data loading time
-        data_time.update(time.time() - end)
-
         if use_gpu:
             imgs, pids = imgs.cuda(), pids.cuda()
+
+        # measure data loading time
+        data_time.update(time.time() - end)
+        
         outputs, features = model(imgs)
         if args.htri_only:
             # only use hard triplet loss to train the network