[fbsync] Moving the check for prototype support in all references. (#4849)

Reviewed By: kazhang

Differential Revision: D32216666

fbshipit-source-id: 97eee21e32d3946cfca569eaf0dd9f98ea2fce1a

Author: prabhat00155

references/classification/train.py

@@ -182,6 +182,8 @@ def load_data(traindir, valdir, args):
 
 
 def main(args):
+    if args.weights and PM is None:
+        raise ImportError("The prototype module couldn't be found. Please install the latest torchvision nightly.")
     if args.output_dir:
         utils.mkdir(args.output_dir)
 
@@ -226,8 +228,6 @@ def main(args):
     if not args.weights:
         model = torchvision.models.__dict__[args.model](pretrained=args.pretrained, num_classes=num_classes)
     else:
-        if PM is None:
-            raise ImportError("The prototype module couldn't be found. Please install the latest torchvision nightly.")
         model = PM.__dict__[args.model](weights=args.weights, num_classes=num_classes)
     model.to(device)
 

references/classification/train_quantization.py

@@ -19,6 +19,8 @@
 
 
 def main(args):
+    if args.weights and PM is None:
+        raise ImportError("The prototype module couldn't be found. Please install the latest torchvision nightly.")
     if args.output_dir:
         utils.mkdir(args.output_dir)
 
@@ -55,8 +57,6 @@ def main(args):
     if not args.weights:
         model = torchvision.models.quantization.__dict__[args.model](pretrained=True, quantize=args.test_only)
     else:
-        if PM is None:
-            raise ImportError("The prototype module couldn't be found. Please install the latest torchvision nightly.")
         model = PM.quantization.__dict__[args.model](weights=args.weights, quantize=args.test_only)
     model.to(device)
 

references/detection/train.py

@@ -148,6 +148,8 @@ def get_args_parser(add_help=True):
 
 
 def main(args):
+    if args.weights and PM is None:
+        raise ImportError("The prototype module couldn't be found. Please install the latest torchvision nightly.")
     if args.output_dir:
         utils.mkdir(args.output_dir)
 
@@ -194,8 +196,6 @@ def main(args):
             pretrained=args.pretrained, num_classes=num_classes, **kwargs
         )
     else:
-        if PM is None:
-            raise ImportError("The prototype module couldn't be found. Please install the latest torchvision nightly.")
         model = PM.detection.__dict__[args.model](weights=args.weights, num_classes=num_classes, **kwargs)
     model.to(device)
     if args.distributed and args.sync_bn:

references/segmentation/train.py

@@ -92,6 +92,8 @@ def train_one_epoch(model, criterion, optimizer, data_loader, lr_scheduler, devi
 
 
 def main(args):
+    if args.weights and PM is None:
+        raise ImportError("The prototype module couldn't be found. Please install the latest torchvision nightly.")
     if args.output_dir:
         utils.mkdir(args.output_dir)
 
@@ -130,8 +132,6 @@ def main(args):
             aux_loss=args.aux_loss,
         )
     else:
-        if PM is None:
-            raise ImportError("The prototype module couldn't be found. Please install the latest torchvision nightly.")
         model = PM.segmentation.__dict__[args.model](
             weights=args.weights, num_classes=num_classes, aux_loss=args.aux_loss
         )

references/video_classification/train.py

@@ -99,6 +99,8 @@ def collate_fn(batch):
 
 
 def main(args):
+    if args.weights and PM is None:
+        raise ImportError("The prototype module couldn't be found. Please install the latest torchvision nightly.")
     if args.apex and amp is None:
         raise RuntimeError(
             "Failed to import apex. Please install apex from https://www.github.com/nvidia/apex "
@@ -214,8 +216,6 @@ def main(args):
     if not args.weights:
         model = torchvision.models.video.__dict__[args.model](pretrained=args.pretrained)
     else:
-        if PM is None:
-            raise ImportError("The prototype module couldn't be found. Please install the latest torchvision nightly.")
         model = PM.video.__dict__[args.model](weights=args.weights)
     model.to(device)
     if args.distributed and args.sync_bn: