Fix typo under torchgen directory (#111154)

This PR fixes typo in comments and messages in files under `torchgen` directory.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/111154
Approved by: https://github.com/rajveer43, https://github.com/Skylion007

torchgen/api/cpp.py

@@ -124,7 +124,7 @@ def valuetype_type(
         raise AssertionError(f"unrecognized type {repr(t)}")
 
 
-# Translation of types occuring in JIT arguments to a C++ argument type.
+# Translation of types occurring in JIT arguments to a C++ argument type.
 # If remove_non_owning_ref_types is set, we'll guarantee that the outputed CType is not a non-owning reference type.
 # For example, we'll return std::vector<int> instead of IntArrayRef.
 # See Note [translation from C++ reference to value types]

torchgen/api/structured.py

@@ -38,7 +38,7 @@ from torchgen.utils import assert_never
 # API have been fixed.
 
 
-# Translation of types occuring in JIT arguments to a C++ argument type.
+# Translation of types occurring in JIT arguments to a C++ argument type.
 # NB: For now, mutable doesn't do anything; but it could if we make
 # some more nominal types
 def argumenttype_type(t: Type, *, mutable: bool, binds: ArgName) -> NamedCType:

torchgen/dest/lazy_ir.py

@@ -216,7 +216,7 @@ class GenLazyIR(ABC):
 
         scalar_args = schema.filtered_args(values=False, scalars=True)
 
-        # Shape constuction.
+        # Shape construction.
         # Conditionally build shape depending on specified shape property
         if schema.properties.ShapePrecompute:
             shape_ctor_arg = "std::move(shapes),"

torchgen/executorch/api/et_cpp.py

@@ -93,7 +93,7 @@ def valuetype_type(
         raise AssertionError(f"unrecognized type {repr(t)}")
 
 
-# Translation of types occuring in JIT arguments to a C++ argument type.
+# Translation of types occurring in JIT arguments to a C++ argument type.
 # If remove_non_owning_ref_types is set, we'll guarantee that the outputed CType is not a non-owning reference type.
 # For example, we'll return std::vector<int> instead of IntArrayRef.
 # See Note [translation from C++ reference to value types]

torchgen/gen_executorch.py

@@ -879,7 +879,7 @@ def main() -> None:
         "--manual_registration",
         "--manual-registration",
         action="store_true",
-        help="a boolean flag to indicate whether we want to maually call"
+        help="a boolean flag to indicate whether we want to manually call"
         "register_kernels() or rely on static init. ",
     )
     parser.add_argument(

torchgen/gen_functionalization_type.py

@@ -629,7 +629,7 @@ def emit_inplace_functionalization_body(
       if ({str(not any_storage_args and f.func.kind() == SchemaKind.inplace).lower()}) {{
         // Before converting the mutable op to its functional variant, run meta tensors through the original op.
         // This will help us catch shape errors that apply to inplace ops that wouldn't apply to their functional variants.
-        // (We can only do this for inplace ops today though, because they technicaly all support meta tensors).
+        // (We can only do this for inplace ops today though, because they technically all support meta tensors).
         {meta_conversion_str}
         at::AutoDispatchSkipFunctionalize func_guard;
         c10::impl::ExcludeDispatchKeyGuard guard(exclude_keys_for_meta_dispatch);
@@ -731,7 +731,7 @@ def gen_functionalization_registration(
             # See Note [resize_ in Functionalization]
             return []
         assert not f.is_view_op
-        # functionalization needs to generate and register kernals for inplace ops.
+        # functionalization needs to generate and register kernels for inplace ops.
         # We *also* need to directly register CompositeImplicitAUtograd kernels
         # so that they decompose properly before functioanlization.
         if modifies_arguments(f):

torchgen/gen_lazy_tensor.py

@@ -426,7 +426,7 @@ def run_gen_lazy_tensor(
 
     Generated lazy native functions all perform shape inference, by first using a meta:: kernel
     if available for that op, and otherwise using a 'compute_shape_{op}' function instead.  The generator
-    knows the call signature for compute_shape_{op} becuase it matches the nativefunction (and meta::) signature,
+    knows the call signature for compute_shape_{op} because it matches the nativefunction (and meta::) signature,
     so it just has to check whether the op is structured and generate a call for one or the other.  It's up to the dev
     to supply the missing compute_shape_{op} function, but the codegen at least warns you about this and provides
     the expected signature which can be copy-pasted into shape_inference.h.

torchgen/model.py

@@ -1413,7 +1413,7 @@ class FunctionSchema:
                 ), "out= ops that accept tensor lists as out arguments "
                 "are expected to have no return type (since you can't do method chaining on them)"
             else:
-                # mutable keyward arguments whose name has _scratch_ prefix are
+                # mutable keyword arguments whose name has _scratch_ prefix are
                 # scratch tensors for memory planning and should not be returned
                 assert len(
                     [
@@ -2208,7 +2208,7 @@ class Arguments:
             post_self_positional=tuple(
                 map(strip_arg_annotation, self.post_self_positional)
             ),
-            # Since TensorOptions are droped, the post_tensor_options_kwargs are
+            # Since TensorOptions are dropped, the post_tensor_options_kwargs are
             # converted to pre_tensor_options_kwargs
             pre_tensor_options_kwarg_only=tuple(
                 map(strip_arg_annotation, self.pre_tensor_options_kwarg_only)

torchgen/native_function_generation.py

@@ -371,8 +371,8 @@ def add_generated_native_functions(
     rs: List[NativeFunction],
     indices: Dict[DispatchKey, Dict[OperatorName, BackendMetadata]],
 ) -> None:
-    # The main code for gnerating new NativeFunctions
-    # First we group of NaitveFunctions by schema kind,
+    # The main code for generating new NativeFunctions
+    # First we group of NativeFunctions by schema kind,
     # then we detect which ones are missing and generate them.
     pre_grouped_native_functions = pre_group_native_functions(rs)
     for d in pre_grouped_native_functions.values():