updates

train.py

@@ -156,7 +156,6 @@ def train(cfg,
         try:
             from apex import amp
             model, optimizer = amp.initialize(model, optimizer, opt_level='O1', verbosity=0)
-            print('Using Apex')
         except:  # not installed: install help: https://github.com/NVIDIA/apex/issues/259
             mixed_precision = False
 

utils/torch_utils.py

@@ -16,17 +16,24 @@ def select_device(force_cpu=False):
     if not cuda:
         print('Using CPU')
     if cuda:
+        try:  # Mixed precision training https://github.com/NVIDIA/apex
+            from apex import amp
+            apex_str = ' with Apex'
+        except:
+            apex_str = ''
+
         torch.backends.cudnn.benchmark = True  # set False for reproducible results
         c = 1024 ** 2  # bytes to MB
         ng = torch.cuda.device_count()
         x = [torch.cuda.get_device_properties(i) for i in range(ng)]
-        print("Using CUDA device0 _CudaDeviceProperties(name='%s', total_memory=%dMB)" %
-              (x[0].name, x[0].total_memory / c))
+        cuda_str = 'Using CUDA' + apex_str
+        print("%sdevice0 _CudaDeviceProperties(name='%s', total_memory=%dMB)" %
+              (cuda_str, x[0].name, x[0].total_memory / c))
         if ng > 0:
             # torch.cuda.set_device(0)  # OPTIONAL: Set GPU ID
             for i in range(1, ng):
-                print("           device%g _CudaDeviceProperties(name='%s', total_memory=%dMB)" %
-                      (i, x[i].name, x[i].total_memory / c))
+                print("%sdevice%g _CudaDeviceProperties(name='%s', total_memory=%dMB)" %
+                      (' ' * len(cuda_str), i, x[i].name, x[i].total_memory / c))
 
     print('')  # skip a line
     return device