Fix require_grad typos (#6930)

Signed-off-by: Sergii Dymchenko <sdym@fb.com> Signed-off-by: Gerald Shen <geshen@nvidia.com>
NVIDIA · Jul 12, 2023 · b06588e · b06588e
1 parent 9d3e731
commit b06588e
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 12 deletions.
diff --git a/nemo/collections/asr/modules/transformer/transformer_generators.py b/nemo/collections/asr/modules/transformer/transformer_generators.py
@@ -188,7 +188,7 @@ def freeze(self) -> None:
             param.requires_grad = False
         self.decoder.eval()
         for param in self.log_softmax.parameters():
-            param.require_grad = False
+            param.requires_grad = False
         self.log_softmax.eval()
 
     def unfreeze(self) -> None:
@@ -201,7 +201,7 @@ def unfreeze(self) -> None:
             param.requires_grad = True
         self.decoder.train()
         for param in self.log_softmax.parameters():
-            param.require_grad = True
+            param.requires_grad = True
         self.log_softmax.train()
 
     @contextmanager
@@ -701,10 +701,10 @@ def freeze(self) -> None:
                 param.requires_grad = False
             self.decoders[model_num].eval()
             for param in self.log_softmaxes[model_num].parameters():
-                param.require_grad = False
+                param.requires_grad = False
             self.log_softmaxes[model_num].eval()
             for param in self.encoders[model_num].parameters():
-                param.require_grad = False
+                param.requires_grad = False
             self.encoders[model_num].eval()
 
     def unfreeze(self) -> None:
@@ -718,10 +718,10 @@ def unfreeze(self) -> None:
                 param.requires_grad = True
             self.decoders[model_num].train()
             for param in self.log_softmaxes[model_num].parameters():
-                param.require_grad = True
+                param.requires_grad = True
             self.log_softmaxes[model_num].train()
             for param in self.encoders[model_num].parameters():
-                param.require_grad = True
+                param.requires_grad = True
             self.encoders[model_num].train()
 
     @contextmanager

diff --git a/nemo/collections/nlp/modules/common/transformer/transformer_generators.py b/nemo/collections/nlp/modules/common/transformer/transformer_generators.py
@@ -188,7 +188,7 @@ def freeze(self) -> None:
             param.requires_grad = False
         self.decoder.eval()
         for param in self.log_softmax.parameters():
-            param.require_grad = False
+            param.requires_grad = False
         self.log_softmax.eval()
 
     def unfreeze(self) -> None:
@@ -201,7 +201,7 @@ def unfreeze(self) -> None:
             param.requires_grad = True
         self.decoder.train()
         for param in self.log_softmax.parameters():
-            param.require_grad = True
+            param.requires_grad = True
         self.log_softmax.train()
 
     @contextmanager
@@ -701,10 +701,10 @@ def freeze(self) -> None:
                 param.requires_grad = False
             self.decoders[model_num].eval()
             for param in self.log_softmaxes[model_num].parameters():
-                param.require_grad = False
+                param.requires_grad = False
             self.log_softmaxes[model_num].eval()
             for param in self.encoders[model_num].parameters():
-                param.require_grad = False
+                param.requires_grad = False
             self.encoders[model_num].eval()
 
     def unfreeze(self) -> None:
@@ -718,10 +718,10 @@ def unfreeze(self) -> None:
                 param.requires_grad = True
             self.decoders[model_num].train()
             for param in self.log_softmaxes[model_num].parameters():
-                param.require_grad = True
+                param.requires_grad = True
             self.log_softmaxes[model_num].train()
             for param in self.encoders[model_num].parameters():
-                param.require_grad = True
+                param.requires_grad = True
             self.encoders[model_num].train()
 
     @contextmanager