fix n_gpu count when no_cuda flag is activated (#3077)

* fix n_gpu count when no_cuda flag is activated

* someone was left behind

examples/contrib/run_swag.py

@@ -622,7 +622,7 @@ def main():
     # Setup CUDA, GPU & distributed training
     if args.local_rank == -1 or args.no_cuda:
         device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
-        args.n_gpu = torch.cuda.device_count()
+        args.n_gpu = 0 if args.no_cuda else torch.cuda.device_count()
     else:  # Initializes the distributed backend which will take care of sychronizing nodes/GPUs
         torch.cuda.set_device(args.local_rank)
         device = torch.device("cuda", args.local_rank)

examples/distillation/run_squad_w_distillation.py

@@ -720,7 +720,7 @@ def main():
     # Setup CUDA, GPU & distributed training
     if args.local_rank == -1 or args.no_cuda:
         device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
-        args.n_gpu = torch.cuda.device_count()
+        args.n_gpu = 0 if args.no_cuda else torch.cuda.device_count()
     else:  # Initializes the distributed backend which will take care of sychronizing nodes/GPUs
         torch.cuda.set_device(args.local_rank)
         device = torch.device("cuda", args.local_rank)

examples/hans/test_hans.py

@@ -520,7 +520,7 @@ def main():
     # Setup CUDA, GPU & distributed training
     if args.local_rank == -1 or args.no_cuda:
         device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
-        args.n_gpu = torch.cuda.device_count()
+        args.n_gpu = 0 if args.no_cuda else torch.cuda.device_count()
     else:  # Initializes the distributed backend which will take care of sychronizing nodes/GPUs
         torch.cuda.set_device(args.local_rank)
         device = torch.device("cuda", args.local_rank)

examples/mm-imdb/run_mmimdb.py

@@ -492,7 +492,7 @@ def main():
     # Setup CUDA, GPU & distributed training
     if args.local_rank == -1 or args.no_cuda:
         device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
-        args.n_gpu = torch.cuda.device_count()
+        args.n_gpu = 0 if args.no_cuda else torch.cuda.device_count()
     else:  # Initializes the distributed backend which will take care of sychronizing nodes/GPUs
         torch.cuda.set_device(args.local_rank)
         device = torch.device("cuda", args.local_rank)

examples/ner/run_ner.py

@@ -557,7 +557,7 @@ def main():
     # Setup CUDA, GPU & distributed training
     if args.local_rank == -1 or args.no_cuda:
         device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
-        args.n_gpu = torch.cuda.device_count()
+        args.n_gpu = 0 if args.no_cuda else torch.cuda.device_count()
     else:  # Initializes the distributed backend which will take care of sychronizing nodes/GPUs
         torch.cuda.set_device(args.local_rank)
         device = torch.device("cuda", args.local_rank)

examples/run_bertology.py

@@ -338,7 +338,7 @@ def main():
     # Setup devices and distributed training
     if args.local_rank == -1 or args.no_cuda:
         args.device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
-        args.n_gpu = torch.cuda.device_count()
+        args.n_gpu = 0 if args.no_cuda else torch.cuda.device_count()
     else:
         torch.cuda.set_device(args.local_rank)
         args.device = torch.device("cuda", args.local_rank)

examples/run_generation.py

@@ -189,7 +189,7 @@ def main():
     args = parser.parse_args()
 
     args.device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
-    args.n_gpu = torch.cuda.device_count()
+    args.n_gpu = 0 if args.no_cuda else torch.cuda.device_count()
 
     set_seed(args)
 

examples/run_glue.py

@@ -575,7 +575,7 @@ def main():
     # Setup CUDA, GPU & distributed training
     if args.local_rank == -1 or args.no_cuda:
         device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
-        args.n_gpu = torch.cuda.device_count()
+        args.n_gpu = 0 if args.no_cuda else torch.cuda.device_count()
     else:  # Initializes the distributed backend which will take care of sychronizing nodes/GPUs
         torch.cuda.set_device(args.local_rank)
         device = torch.device("cuda", args.local_rank)

examples/run_language_modeling.py

@@ -663,7 +663,7 @@ def main():
     # Setup CUDA, GPU & distributed training
     if args.local_rank == -1 or args.no_cuda:
         device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
-        args.n_gpu = torch.cuda.device_count()
+        args.n_gpu = 0 if args.no_cuda else torch.cuda.device_count()
     else:  # Initializes the distributed backend which will take care of sychronizing nodes/GPUs
         torch.cuda.set_device(args.local_rank)
         device = torch.device("cuda", args.local_rank)

examples/run_multiple_choice.py

@@ -535,7 +535,7 @@ def main():
     # Setup CUDA, GPU & distributed training
     if args.local_rank == -1 or args.no_cuda:
         device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
-        args.n_gpu = torch.cuda.device_count()
+        args.n_gpu = 0 if args.no_cuda else torch.cuda.device_count()
     else:  # Initializes the distributed backend which will take care of sychronizing nodes/GPUs
         torch.cuda.set_device(args.local_rank)
         device = torch.device("cuda", args.local_rank)

examples/run_squad.py

@@ -725,7 +725,7 @@ def main():
     # Setup CUDA, GPU & distributed training
     if args.local_rank == -1 or args.no_cuda:
         device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
-        args.n_gpu = torch.cuda.device_count()
+        args.n_gpu = 0 if args.no_cuda else torch.cuda.device_count()
     else:  # Initializes the distributed backend which will take care of sychronizing nodes/GPUs
         torch.cuda.set_device(args.local_rank)
         device = torch.device("cuda", args.local_rank)

examples/run_xnli.py

@@ -530,7 +530,7 @@ def main():
     # Setup CUDA, GPU & distributed training
     if args.local_rank == -1 or args.no_cuda:
         device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
-        args.n_gpu = torch.cuda.device_count()
+        args.n_gpu = 0 if args.no_cuda else torch.cuda.device_count()
     else:  # Initializes the distributed backend which will take care of sychronizing nodes/GPUs
         torch.cuda.set_device(args.local_rank)
         device = torch.device("cuda", args.local_rank)

templates/adding_a_new_example_script/run_xxx.py

@@ -594,7 +594,7 @@ def main():
     # Setup CUDA, GPU & distributed training
     if args.local_rank == -1 or args.no_cuda:
         device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
-        args.n_gpu = torch.cuda.device_count()
+        args.n_gpu = 0 if args.no_cuda else torch.cuda.device_count()
     else:  # Initializes the distributed backend which will take care of sychronizing nodes/GPUs
         torch.cuda.set_device(args.local_rank)
         device = torch.device("cuda", args.local_rank)