(fix)(Torch Frontend)(module.py): made a couple of fixes to the `name…

…d_modules and `apply` methods to handle native keras classes as part of submodules/children.

ivy/functional/frontends/torch/nn/modules/module.py

@@ -129,7 +129,10 @@ def add_module(self, name: str, module: Optional["Module"]) -> None:
 
     def apply(self, fn: Callable[["Module"], None]):
         for module in self.children():
-            module.apply(fn)
+            if hasattr(module, "apply"):
+                module.apply(fn)
+            else:
+                fn(module)
         fn(self)
         return self
 
@@ -262,9 +265,12 @@ def named_modules(
                 if module is None:
                     continue
                 submodule_prefix = prefix + ("." if prefix else "") + name
-                yield from module.named_modules(
-                    memo, submodule_prefix, remove_duplicate
-                )
+                if not hasattr(module, "named_modules"):
+                    yield submodule_prefix, self
+                else:
+                    yield from module.named_modules(
+                        memo, submodule_prefix, remove_duplicate
+                    )
 
     def requires_grad_(self, requires_grad: bool = True):
         for p in self.parameters():