From 27f9d3b0a17289500f1a2d24e4901a1e0fa9ea95 Mon Sep 17 00:00:00 2001
From: Aaron Orenstein <aorenste@fb.com>
Date: Sat, 8 Jun 2024 11:41:15 -0700
Subject: [PATCH] Flip default value for mypy disallow_untyped_defs [8/11]
 (#127845)

See #127836 for details.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/127845
Approved by: https://github.com/oulgen
ghstack dependencies: #127842, #127843, #127844
---
 torch/nn/backends/thnn.py                                        | 1 +
 torch/nn/cpp.py                                                  | 1 +
 torch/nn/grad.py                                                 | 1 +
 torch/nn/init.py                                                 | 1 +
 torch/nn/modules/_functions.py                                   | 1 +
 torch/nn/modules/activation.py                                   | 1 +
 torch/nn/modules/adaptive.py                                     | 1 +
 torch/nn/modules/batchnorm.py                                    | 1 +
 torch/nn/modules/container.py                                    | 1 +
 torch/nn/modules/conv.py                                         | 1 +
 torch/nn/modules/flatten.py                                      | 1 +
 torch/nn/modules/instancenorm.py                                 | 1 +
 torch/nn/modules/lazy.py                                         | 1 +
 torch/nn/modules/linear.py                                       | 1 +
 torch/nn/modules/loss.py                                         | 1 +
 torch/nn/modules/module.py                                       | 1 +
 torch/nn/modules/normalization.py                                | 1 +
 torch/nn/modules/padding.py                                      | 1 +
 torch/nn/modules/rnn.py                                          | 1 +
 torch/nn/modules/sparse.py                                       | 1 +
 torch/nn/modules/transformer.py                                  | 1 +
 torch/nn/modules/upsampling.py                                   | 1 +
 torch/nn/modules/utils.py                                        | 1 +
 torch/nn/parallel/__init__.py                                    | 1 +
 torch/nn/parallel/comm.py                                        | 1 +
 torch/nn/parallel/data_parallel.py                               | 1 +
 torch/nn/parallel/distributed.py                                 | 1 +
 torch/nn/parallel/scatter_gather.py                              | 1 +
 torch/nn/parameter.pyi                                           | 1 +
 torch/nn/utils/_deprecation_utils.py                             | 1 +
 torch/nn/utils/_expanded_weights/conv_expanded_weights.py        | 1 +
 torch/nn/utils/_expanded_weights/conv_utils.py                   | 1 +
 torch/nn/utils/_expanded_weights/embedding_expanded_weights.py   | 1 +
 torch/nn/utils/_expanded_weights/expanded_weights_impl.py        | 1 +
 torch/nn/utils/_expanded_weights/expanded_weights_utils.py       | 1 +
 torch/nn/utils/_expanded_weights/group_norm_expanded_weights.py  | 1 +
 .../nn/utils/_expanded_weights/instance_norm_expanded_weights.py | 1 +
 torch/nn/utils/_expanded_weights/layer_norm_expanded_weights.py  | 1 +
 torch/nn/utils/_expanded_weights/linear_expanded_weights.py      | 1 +
 torch/nn/utils/_per_sample_grad.py                               | 1 +
 torch/nn/utils/clip_grad.py                                      | 1 +
 torch/nn/utils/init.py                                           | 1 +
 torch/nn/utils/memory_format.py                                  | 1 +
 torch/nn/utils/parametrizations.py                               | 1 +
 torch/nn/utils/parametrize.py                                    | 1 +
 torch/nn/utils/prune.py                                          | 1 +
 torch/nn/utils/rnn.pyi                                           | 1 +
 torch/nn/utils/spectral_norm.py                                  | 1 +
 torch/nn/utils/stateless.py                                      | 1 +
 torch/nn/utils/weight_norm.py                                    | 1 +
 torch/onnx/__init__.py                                           | 1 +
 torch/onnx/_deprecation.py                                       | 1 +
 torch/onnx/_globals.py                                           | 1 +
 torch/onnx/_internal/_beartype.py                                | 1 +
 torch/onnx/_internal/diagnostics/_diagnostic.py                  | 1 +
 torch/onnx/_internal/diagnostics/_rules.py                       | 1 +
 torch/onnx/_internal/diagnostics/infra/_infra.py                 | 1 +
 torch/onnx/_internal/diagnostics/infra/context.py                | 1 +
 torch/onnx/_internal/diagnostics/infra/decorator.py              | 1 +
 torch/onnx/_internal/exporter.py                                 | 1 +
 torch/onnx/_internal/fx/_pass.py                                 | 1 +
 torch/onnx/_internal/fx/analysis/unsupported_nodes.py            | 1 +
 torch/onnx/_internal/fx/decomposition_skip.py                    | 1 +
 torch/onnx/_internal/fx/decomposition_table.py                   | 1 +
 torch/onnx/_internal/fx/diagnostics.py                           | 1 +
 torch/onnx/_internal/fx/dynamo_graph_extractor.py                | 1 +
 torch/onnx/_internal/fx/fx_onnx_interpreter.py                   | 1 +
 torch/onnx/_internal/fx/fx_symbolic_graph_extractor.py           | 1 +
 torch/onnx/_internal/fx/onnxfunction_dispatcher.py               | 1 +
 torch/onnx/_internal/fx/op_validation.py                         | 1 +
 torch/onnx/_internal/fx/passes/_utils.py                         | 1 +
 torch/onnx/_internal/fx/passes/decomp.py                         | 1 +
 torch/onnx/_internal/fx/passes/functionalization.py              | 1 +
 torch/onnx/_internal/fx/passes/modularization.py                 | 1 +
 torch/onnx/_internal/fx/passes/readability.py                    | 1 +
 torch/onnx/_internal/fx/passes/type_promotion.py                 | 1 +
 torch/onnx/_internal/fx/passes/virtualization.py                 | 1 +
 torch/onnx/_internal/fx/patcher.py                               | 1 +
 torch/onnx/_internal/fx/serialization.py                         | 1 +
 torch/onnx/_internal/fx/torch_export_graph_extractor.py          | 1 +
 torch/onnx/_internal/fx/type_utils.py                            | 1 +
 torch/onnx/_internal/io_adapter.py                               | 1 +
 torch/onnx/_internal/jit_utils.py                                | 1 +
 torch/onnx/_internal/onnx_proto_utils.py                         | 1 +
 torch/onnx/_internal/onnxruntime.py                              | 1 +
 torch/onnx/_internal/registration.py                             | 1 +
 torch/onnx/_onnx_supported_ops.py                                | 1 +
 torch/onnx/_type_utils.py                                        | 1 +
 torch/onnx/operators.py                                          | 1 +
 torch/onnx/symbolic_caffe2.py                                    | 1 +
 torch/onnx/symbolic_helper.py                                    | 1 +
 torch/onnx/symbolic_opset10.py                                   | 1 +
 torch/onnx/symbolic_opset11.py                                   | 1 +
 torch/onnx/symbolic_opset12.py                                   | 1 +
 torch/onnx/symbolic_opset13.py                                   | 1 +
 torch/onnx/symbolic_opset14.py                                   | 1 +
 torch/onnx/symbolic_opset15.py                                   | 1 +
 torch/onnx/symbolic_opset16.py                                   | 1 +
 torch/onnx/symbolic_opset17.py                                   | 1 +
 torch/onnx/symbolic_opset18.py                                   | 1 +
 torch/onnx/symbolic_opset20.py                                   | 1 +
 torch/onnx/symbolic_opset7.py                                    | 1 +
 torch/onnx/symbolic_opset8.py                                    | 1 +
 torch/onnx/symbolic_opset9.py                                    | 1 +
 torch/onnx/utils.py                                              | 1 +
 torch/onnx/verification.py                                       | 1 +
 torch/optim/_functional.py                                       | 1 +
 torch/optim/adadelta.py                                          | 1 +
 torch/optim/adagrad.py                                           | 1 +
 torch/optim/adam.py                                              | 1 +
 torch/optim/adamax.py                                            | 1 +
 torch/optim/adamw.py                                             | 1 +
 torch/optim/asgd.py                                              | 1 +
 torch/optim/lbfgs.py                                             | 1 +
 torch/optim/lr_scheduler.py                                      | 1 +
 torch/optim/nadam.py                                             | 1 +
 torch/optim/optimizer.py                                         | 1 +
 torch/optim/radam.py                                             | 1 +
 torch/optim/rmsprop.py                                           | 1 +
 torch/optim/rprop.py                                             | 1 +
 torch/optim/sgd.py                                               | 1 +
 torch/optim/sparse_adam.py                                       | 1 +
 torch/optim/swa_utils.py                                         | 1 +
 torch/package/_digraph.py                                        | 1 +
 torch/package/_directory_reader.py                               | 1 +
 torch/package/_importlib.py                                      | 1 +
 torch/package/_mangling.py                                       | 1 +
 torch/package/_mock.py                                           | 1 +
 128 files changed, 128 insertions(+)

diff --git a/torch/nn/backends/thnn.py b/torch/nn/backends/thnn.py
index 5250b4bff167..3cb0f3ff57e2 100644
--- a/torch/nn/backends/thnn.py
+++ b/torch/nn/backends/thnn.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 # this is for historical pickle deserialization, it is not used otherwise
 
 def _get_thnn_function_backend():
diff --git a/torch/nn/cpp.py b/torch/nn/cpp.py
index a08c7b314100..98a61bfb7c42 100644
--- a/torch/nn/cpp.py
+++ b/torch/nn/cpp.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Functionality for Python <-> C++ frontend inter-op."""
 
 from torch import nn
diff --git a/torch/nn/grad.py b/torch/nn/grad.py
index 660c87fb4133..dbd38fcdd38c 100644
--- a/torch/nn/grad.py
+++ b/torch/nn/grad.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Gradient interface."""
 
 import torch
diff --git a/torch/nn/init.py b/torch/nn/init.py
index f5be081e7dd0..b3179abb4937 100644
--- a/torch/nn/init.py
+++ b/torch/nn/init.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """This file contains utilities for initializing neural network parameters."""
 import math
 import warnings
diff --git a/torch/nn/modules/_functions.py b/torch/nn/modules/_functions.py
index 669448ce4fda..0e19faa99e5c 100644
--- a/torch/nn/modules/_functions.py
+++ b/torch/nn/modules/_functions.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import torch
 import torch.distributed as dist
 
diff --git a/torch/nn/modules/activation.py b/torch/nn/modules/activation.py
index 5dec6f9578b1..3d8b65175956 100644
--- a/torch/nn/modules/activation.py
+++ b/torch/nn/modules/activation.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import warnings
 from typing import Optional, Tuple
 
diff --git a/torch/nn/modules/adaptive.py b/torch/nn/modules/adaptive.py
index 83b37696c8a7..a6c2da5f596f 100644
--- a/torch/nn/modules/adaptive.py
+++ b/torch/nn/modules/adaptive.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 
 from collections import namedtuple
 
diff --git a/torch/nn/modules/batchnorm.py b/torch/nn/modules/batchnorm.py
index 3c48e56d5e6e..75c8b5504d46 100644
--- a/torch/nn/modules/batchnorm.py
+++ b/torch/nn/modules/batchnorm.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import Optional, Any
 
 import torch
diff --git a/torch/nn/modules/container.py b/torch/nn/modules/container.py
index 775a826d69cc..c82d8d7d3037 100644
--- a/torch/nn/modules/container.py
+++ b/torch/nn/modules/container.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from collections import OrderedDict, abc as container_abcs
 from itertools import chain, islice
 import operator
diff --git a/torch/nn/modules/conv.py b/torch/nn/modules/conv.py
index 4ab4c8bff9fc..fb6a1557aa71 100644
--- a/torch/nn/modules/conv.py
+++ b/torch/nn/modules/conv.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import math
 
 import torch
diff --git a/torch/nn/modules/flatten.py b/torch/nn/modules/flatten.py
index eaf62d5bbeea..f1c44fd350d1 100644
--- a/torch/nn/modules/flatten.py
+++ b/torch/nn/modules/flatten.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from .module import Module
 
 from typing import Tuple, Union
diff --git a/torch/nn/modules/instancenorm.py b/torch/nn/modules/instancenorm.py
index ae187e98b7e6..e6a3e1c0a3a1 100644
--- a/torch/nn/modules/instancenorm.py
+++ b/torch/nn/modules/instancenorm.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 
 import warnings
 from torch import Tensor
diff --git a/torch/nn/modules/lazy.py b/torch/nn/modules/lazy.py
index c4b7459c4acd..f4be1b7db706 100644
--- a/torch/nn/modules/lazy.py
+++ b/torch/nn/modules/lazy.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import itertools
 from typing import Protocol, Optional, Type, Any
 
diff --git a/torch/nn/modules/linear.py b/torch/nn/modules/linear.py
index 54981596f7ee..be2739462399 100644
--- a/torch/nn/modules/linear.py
+++ b/torch/nn/modules/linear.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import math
 from typing import Any
 
diff --git a/torch/nn/modules/loss.py b/torch/nn/modules/loss.py
index 4324c1df144d..497da8218506 100644
--- a/torch/nn/modules/loss.py
+++ b/torch/nn/modules/loss.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from .distance import PairwiseDistance
 from .module import Module
 from .. import functional as F
diff --git a/torch/nn/modules/module.py b/torch/nn/modules/module.py
index ffd429cc06f2..f803d3f02a17 100644
--- a/torch/nn/modules/module.py
+++ b/torch/nn/modules/module.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from collections import OrderedDict, namedtuple
 import itertools
 import warnings
diff --git a/torch/nn/modules/normalization.py b/torch/nn/modules/normalization.py
index 97c9c307c5d9..d503409d53a1 100644
--- a/torch/nn/modules/normalization.py
+++ b/torch/nn/modules/normalization.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import torch
 import numbers
 from torch.nn.parameter import Parameter
diff --git a/torch/nn/modules/padding.py b/torch/nn/modules/padding.py
index 0aecca58c305..4b29fbf1c8f4 100644
--- a/torch/nn/modules/padding.py
+++ b/torch/nn/modules/padding.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from .module import Module
 from .utils import _pair, _quadruple, _ntuple
 from .. import functional as F
diff --git a/torch/nn/modules/rnn.py b/torch/nn/modules/rnn.py
index b4bdd7824474..8ba4f9f08319 100644
--- a/torch/nn/modules/rnn.py
+++ b/torch/nn/modules/rnn.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import math
 import warnings
 import numbers
diff --git a/torch/nn/modules/sparse.py b/torch/nn/modules/sparse.py
index f053a0c8f3c2..512b17d03222 100644
--- a/torch/nn/modules/sparse.py
+++ b/torch/nn/modules/sparse.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import Optional
 
 import torch
diff --git a/torch/nn/modules/transformer.py b/torch/nn/modules/transformer.py
index 3c9a8547df32..f5980cd6b1e8 100644
--- a/torch/nn/modules/transformer.py
+++ b/torch/nn/modules/transformer.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import copy
 from typing import Optional, Any, Union, Callable
 
diff --git a/torch/nn/modules/upsampling.py b/torch/nn/modules/upsampling.py
index da9b23add18d..7d674da0d5c3 100644
--- a/torch/nn/modules/upsampling.py
+++ b/torch/nn/modules/upsampling.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from .module import Module
 from .. import functional as F
 
diff --git a/torch/nn/modules/utils.py b/torch/nn/modules/utils.py
index 019dabe3e533..4a051ed1eba5 100644
--- a/torch/nn/modules/utils.py
+++ b/torch/nn/modules/utils.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import collections
 from itertools import repeat
 from typing import List, Dict, Any
diff --git a/torch/nn/parallel/__init__.py b/torch/nn/parallel/__init__.py
index adcd6bd838eb..8f08e5099d8b 100644
--- a/torch/nn/parallel/__init__.py
+++ b/torch/nn/parallel/__init__.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing_extensions import deprecated
 
 from .parallel_apply import parallel_apply
diff --git a/torch/nn/parallel/comm.py b/torch/nn/parallel/comm.py
index 22cf80bd64e2..b907de4004b1 100644
--- a/torch/nn/parallel/comm.py
+++ b/torch/nn/parallel/comm.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import warnings
 import torch
 from torch.cuda import nccl
diff --git a/torch/nn/parallel/data_parallel.py b/torch/nn/parallel/data_parallel.py
index 4471cee6f379..3980706a932a 100644
--- a/torch/nn/parallel/data_parallel.py
+++ b/torch/nn/parallel/data_parallel.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import operator
 import torch
 import warnings
diff --git a/torch/nn/parallel/distributed.py b/torch/nn/parallel/distributed.py
index 37b6501d2c9c..c71c838cfb85 100644
--- a/torch/nn/parallel/distributed.py
+++ b/torch/nn/parallel/distributed.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import copy
 import functools
 import inspect
diff --git a/torch/nn/parallel/scatter_gather.py b/torch/nn/parallel/scatter_gather.py
index f6fb9d47ecbf..73e753760e72 100644
--- a/torch/nn/parallel/scatter_gather.py
+++ b/torch/nn/parallel/scatter_gather.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import torch
 from typing import Any, Dict, List, Optional, Sequence, Tuple, TypeVar, Union, overload
 from typing_extensions import deprecated
diff --git a/torch/nn/parameter.pyi b/torch/nn/parameter.pyi
index 219bb6d4efa2..221ffacc3520 100644
--- a/torch/nn/parameter.pyi
+++ b/torch/nn/parameter.pyi
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import builtins
 from typing import Optional, Tuple
 
diff --git a/torch/nn/utils/_deprecation_utils.py b/torch/nn/utils/_deprecation_utils.py
index 1b2a9b6e29f2..9910db96e66c 100644
--- a/torch/nn/utils/_deprecation_utils.py
+++ b/torch/nn/utils/_deprecation_utils.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import List, Callable
 import importlib
 import warnings
diff --git a/torch/nn/utils/_expanded_weights/conv_expanded_weights.py b/torch/nn/utils/_expanded_weights/conv_expanded_weights.py
index c10ccb90ae92..147346796d1f 100644
--- a/torch/nn/utils/_expanded_weights/conv_expanded_weights.py
+++ b/torch/nn/utils/_expanded_weights/conv_expanded_weights.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import torch
 import torch.nn.functional as F
 
diff --git a/torch/nn/utils/_expanded_weights/conv_utils.py b/torch/nn/utils/_expanded_weights/conv_utils.py
index b675e3b892bd..2836809d40be 100644
--- a/torch/nn/utils/_expanded_weights/conv_utils.py
+++ b/torch/nn/utils/_expanded_weights/conv_utils.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import torch
 import torch.nn.functional as F
 
diff --git a/torch/nn/utils/_expanded_weights/embedding_expanded_weights.py b/torch/nn/utils/_expanded_weights/embedding_expanded_weights.py
index c7956a3a1b1f..593fa9e5eed7 100644
--- a/torch/nn/utils/_expanded_weights/embedding_expanded_weights.py
+++ b/torch/nn/utils/_expanded_weights/embedding_expanded_weights.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import torch
 import torch.nn.functional as F
 from .expanded_weights_impl import implements_per_sample_grads
diff --git a/torch/nn/utils/_expanded_weights/expanded_weights_impl.py b/torch/nn/utils/_expanded_weights/expanded_weights_impl.py
index 94e6041c6de5..664e65cc7d90 100644
--- a/torch/nn/utils/_expanded_weights/expanded_weights_impl.py
+++ b/torch/nn/utils/_expanded_weights/expanded_weights_impl.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from contextlib import contextmanager
 
 import torch
diff --git a/torch/nn/utils/_expanded_weights/expanded_weights_utils.py b/torch/nn/utils/_expanded_weights/expanded_weights_utils.py
index 249dbe591204..840be6a163f5 100644
--- a/torch/nn/utils/_expanded_weights/expanded_weights_utils.py
+++ b/torch/nn/utils/_expanded_weights/expanded_weights_utils.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import Optional
 
 import torch
diff --git a/torch/nn/utils/_expanded_weights/group_norm_expanded_weights.py b/torch/nn/utils/_expanded_weights/group_norm_expanded_weights.py
index fe29b1eafbe2..6e2919803e4f 100644
--- a/torch/nn/utils/_expanded_weights/group_norm_expanded_weights.py
+++ b/torch/nn/utils/_expanded_weights/group_norm_expanded_weights.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from functools import reduce
 import operator
 import torch
diff --git a/torch/nn/utils/_expanded_weights/instance_norm_expanded_weights.py b/torch/nn/utils/_expanded_weights/instance_norm_expanded_weights.py
index f3e68b940660..1d0f40c54081 100644
--- a/torch/nn/utils/_expanded_weights/instance_norm_expanded_weights.py
+++ b/torch/nn/utils/_expanded_weights/instance_norm_expanded_weights.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from functools import partial
 import torch
 import torch.nn.functional as F
diff --git a/torch/nn/utils/_expanded_weights/layer_norm_expanded_weights.py b/torch/nn/utils/_expanded_weights/layer_norm_expanded_weights.py
index f2ead2d4c08f..b18c284cd7cf 100644
--- a/torch/nn/utils/_expanded_weights/layer_norm_expanded_weights.py
+++ b/torch/nn/utils/_expanded_weights/layer_norm_expanded_weights.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 
 import torch
 import torch.nn.functional as F
diff --git a/torch/nn/utils/_expanded_weights/linear_expanded_weights.py b/torch/nn/utils/_expanded_weights/linear_expanded_weights.py
index c2cbae63f336..6a80c1dc9219 100644
--- a/torch/nn/utils/_expanded_weights/linear_expanded_weights.py
+++ b/torch/nn/utils/_expanded_weights/linear_expanded_weights.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import torch
 import torch.nn.functional as F
 from .expanded_weights_impl import implements_per_sample_grads
diff --git a/torch/nn/utils/_per_sample_grad.py b/torch/nn/utils/_per_sample_grad.py
index 0644ab5d2535..a64942083f0c 100644
--- a/torch/nn/utils/_per_sample_grad.py
+++ b/torch/nn/utils/_per_sample_grad.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import functools
 
 import torch
diff --git a/torch/nn/utils/clip_grad.py b/torch/nn/utils/clip_grad.py
index 4ac8a4e7445b..cc83353909f9 100644
--- a/torch/nn/utils/clip_grad.py
+++ b/torch/nn/utils/clip_grad.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import functools
 from typing import Union, Iterable, List, Dict, Tuple, Optional, cast
 from typing_extensions import deprecated
diff --git a/torch/nn/utils/init.py b/torch/nn/utils/init.py
index 416ad0db8ef7..4768d3009005 100644
--- a/torch/nn/utils/init.py
+++ b/torch/nn/utils/init.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import inspect
 import torch
 
diff --git a/torch/nn/utils/memory_format.py b/torch/nn/utils/memory_format.py
index c8fc22bea51c..aaa2b6bfb198 100644
--- a/torch/nn/utils/memory_format.py
+++ b/torch/nn/utils/memory_format.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import torch
 
 
diff --git a/torch/nn/utils/parametrizations.py b/torch/nn/utils/parametrizations.py
index f9b25bcac0cb..cf686504072f 100644
--- a/torch/nn/utils/parametrizations.py
+++ b/torch/nn/utils/parametrizations.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from enum import Enum, auto
 
 import torch
diff --git a/torch/nn/utils/parametrize.py b/torch/nn/utils/parametrize.py
index f512b7c3b22a..b828c1d230f1 100644
--- a/torch/nn/utils/parametrize.py
+++ b/torch/nn/utils/parametrize.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import torch
 from torch.__future__ import get_swap_module_params_on_conversion
 from torch.nn.modules.container import ModuleList, ModuleDict, Module
diff --git a/torch/nn/utils/prune.py b/torch/nn/utils/prune.py
index 0375106d69e0..b0e1f99a6c1f 100644
--- a/torch/nn/utils/prune.py
+++ b/torch/nn/utils/prune.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 r"""Pruning methods."""
 import numbers
 from abc import ABC, abstractmethod
diff --git a/torch/nn/utils/rnn.pyi b/torch/nn/utils/rnn.pyi
index fd033d8888be..9ffc650714ff 100644
--- a/torch/nn/utils/rnn.pyi
+++ b/torch/nn/utils/rnn.pyi
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import Any, Iterable, NamedTuple, Optional, overload, Sequence, Tuple, Union
 
 from typing_extensions import Self
diff --git a/torch/nn/utils/spectral_norm.py b/torch/nn/utils/spectral_norm.py
index bda54b9a1222..fcc4bbf5fe29 100644
--- a/torch/nn/utils/spectral_norm.py
+++ b/torch/nn/utils/spectral_norm.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Spectral Normalization from https://arxiv.org/abs/1802.05957."""
 import torch
 from torch.nn.functional import normalize
diff --git a/torch/nn/utils/stateless.py b/torch/nn/utils/stateless.py
index 660a1a484ebb..07b03c04a120 100644
--- a/torch/nn/utils/stateless.py
+++ b/torch/nn/utils/stateless.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import contextlib
 from collections import defaultdict
 from typing import Any, Dict, Iterator, Optional, Set, Tuple, Union
diff --git a/torch/nn/utils/weight_norm.py b/torch/nn/utils/weight_norm.py
index 6cfe4b3e526d..abb21a7b4672 100644
--- a/torch/nn/utils/weight_norm.py
+++ b/torch/nn/utils/weight_norm.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 r"""Weight Normalization from https://arxiv.org/abs/1602.07868."""
 from torch.nn.parameter import Parameter, UninitializedParameter
 from torch import _weight_norm, norm_except_dim
diff --git a/torch/onnx/__init__.py b/torch/onnx/__init__.py
index 4d16ef09c8b3..2b2f2bdae0de 100644
--- a/torch/onnx/__init__.py
+++ b/torch/onnx/__init__.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from torch import _C
 from torch._C import _onnx as _C_onnx
 from torch._C._onnx import (
diff --git a/torch/onnx/_deprecation.py b/torch/onnx/_deprecation.py
index 0fd2cd764fc9..1f78dd55bd5d 100644
--- a/torch/onnx/_deprecation.py
+++ b/torch/onnx/_deprecation.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Utility for deprecating functions."""
 
 import functools
diff --git a/torch/onnx/_globals.py b/torch/onnx/_globals.py
index f827d12be7fb..22c05075dba8 100644
--- a/torch/onnx/_globals.py
+++ b/torch/onnx/_globals.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Globals used internally by the ONNX exporter.
 
 Do not use this module outside of `torch.onnx` and its tests.
diff --git a/torch/onnx/_internal/_beartype.py b/torch/onnx/_internal/_beartype.py
index 25e1c1cb7299..1e5006fb56c1 100644
--- a/torch/onnx/_internal/_beartype.py
+++ b/torch/onnx/_internal/_beartype.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """An internal wrapper for the beartype library.
 
 The module returns a no-op decorator when the beartype library is not installed.
diff --git a/torch/onnx/_internal/diagnostics/_diagnostic.py b/torch/onnx/_internal/diagnostics/_diagnostic.py
index 09079d5e9c4a..e5b22b07539c 100644
--- a/torch/onnx/_internal/diagnostics/_diagnostic.py
+++ b/torch/onnx/_internal/diagnostics/_diagnostic.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Diagnostic components for TorchScript based ONNX export, i.e. `torch.onnx.export`."""
 from __future__ import annotations
 
diff --git a/torch/onnx/_internal/diagnostics/_rules.py b/torch/onnx/_internal/diagnostics/_rules.py
index 0bfda96c5bce..3b2ca727d0d1 100644
--- a/torch/onnx/_internal/diagnostics/_rules.py
+++ b/torch/onnx/_internal/diagnostics/_rules.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """
 GENERATED CODE - DO NOT EDIT DIRECTLY
 This file is generated by gen_diagnostics.py.
diff --git a/torch/onnx/_internal/diagnostics/infra/_infra.py b/torch/onnx/_internal/diagnostics/infra/_infra.py
index c118f3e5ae14..e51c99a3151b 100644
--- a/torch/onnx/_internal/diagnostics/infra/_infra.py
+++ b/torch/onnx/_internal/diagnostics/infra/_infra.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """This file defines an additional layer of abstraction on top of the SARIF OM."""
 
 from __future__ import annotations
diff --git a/torch/onnx/_internal/diagnostics/infra/context.py b/torch/onnx/_internal/diagnostics/infra/context.py
index 6106a42467c1..f670adc2cae2 100644
--- a/torch/onnx/_internal/diagnostics/infra/context.py
+++ b/torch/onnx/_internal/diagnostics/infra/context.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """A diagnostic context based on SARIF."""
 
 from __future__ import annotations
diff --git a/torch/onnx/_internal/diagnostics/infra/decorator.py b/torch/onnx/_internal/diagnostics/infra/decorator.py
index 0ac803815703..67066f5da500 100644
--- a/torch/onnx/_internal/diagnostics/infra/decorator.py
+++ b/torch/onnx/_internal/diagnostics/infra/decorator.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 import functools
diff --git a/torch/onnx/_internal/exporter.py b/torch/onnx/_internal/exporter.py
index cf9f1cd747e5..ac62de0214dd 100644
--- a/torch/onnx/_internal/exporter.py
+++ b/torch/onnx/_internal/exporter.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import (  # for onnx.ModelProto (ONNXProgram) and onnxruntime (ONNXRuntimeOptions)
     annotations,
 )
diff --git a/torch/onnx/_internal/fx/_pass.py b/torch/onnx/_internal/fx/_pass.py
index 69fa023b9add..cef8e045f7fb 100644
--- a/torch/onnx/_internal/fx/_pass.py
+++ b/torch/onnx/_internal/fx/_pass.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 import abc
diff --git a/torch/onnx/_internal/fx/analysis/unsupported_nodes.py b/torch/onnx/_internal/fx/analysis/unsupported_nodes.py
index 5da0dbed3d91..deec2a85e1da 100644
--- a/torch/onnx/_internal/fx/analysis/unsupported_nodes.py
+++ b/torch/onnx/_internal/fx/analysis/unsupported_nodes.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 import dataclasses
diff --git a/torch/onnx/_internal/fx/decomposition_skip.py b/torch/onnx/_internal/fx/decomposition_skip.py
index 7fb971a3307a..646e0765f190 100644
--- a/torch/onnx/_internal/fx/decomposition_skip.py
+++ b/torch/onnx/_internal/fx/decomposition_skip.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """A context manager that disables the decomposition of certain ops during dynamo tracing.
 
 The approach is to temporarily hijack the operator callable with PT2 custom operator.
diff --git a/torch/onnx/_internal/fx/decomposition_table.py b/torch/onnx/_internal/fx/decomposition_table.py
index 5cb9be6da79d..027d580717af 100644
--- a/torch/onnx/_internal/fx/decomposition_table.py
+++ b/torch/onnx/_internal/fx/decomposition_table.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Dispatcher for AtenLib functions from onnx-script."""
 
 from __future__ import annotations
diff --git a/torch/onnx/_internal/fx/diagnostics.py b/torch/onnx/_internal/fx/diagnostics.py
index 11e4c79f2e1a..0be358751c11 100644
--- a/torch/onnx/_internal/fx/diagnostics.py
+++ b/torch/onnx/_internal/fx/diagnostics.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 import dataclasses
diff --git a/torch/onnx/_internal/fx/dynamo_graph_extractor.py b/torch/onnx/_internal/fx/dynamo_graph_extractor.py
index fbc7d92e043f..1379a0613895 100644
--- a/torch/onnx/_internal/fx/dynamo_graph_extractor.py
+++ b/torch/onnx/_internal/fx/dynamo_graph_extractor.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 # NOTE: This file is referenced by name at
 #       /opt/pytorch/torch/_dynamo/eval_frame.py::DONT_WRAP_FILES.
 #       introduced by https://github.com/pytorch/pytorch/pull/98894.
diff --git a/torch/onnx/_internal/fx/fx_onnx_interpreter.py b/torch/onnx/_internal/fx/fx_onnx_interpreter.py
index 50ead7556f37..a0be86e11d6b 100644
--- a/torch/onnx/_internal/fx/fx_onnx_interpreter.py
+++ b/torch/onnx/_internal/fx/fx_onnx_interpreter.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 import inspect
diff --git a/torch/onnx/_internal/fx/fx_symbolic_graph_extractor.py b/torch/onnx/_internal/fx/fx_symbolic_graph_extractor.py
index 18dc84e19585..1d7d191cbd25 100644
--- a/torch/onnx/_internal/fx/fx_symbolic_graph_extractor.py
+++ b/torch/onnx/_internal/fx/fx_symbolic_graph_extractor.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 import functools
diff --git a/torch/onnx/_internal/fx/onnxfunction_dispatcher.py b/torch/onnx/_internal/fx/onnxfunction_dispatcher.py
index 2986ac279ec3..3886733093a3 100644
--- a/torch/onnx/_internal/fx/onnxfunction_dispatcher.py
+++ b/torch/onnx/_internal/fx/onnxfunction_dispatcher.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Dispatcher for AtenLib functions from onnx-script."""
 
 from __future__ import annotations
diff --git a/torch/onnx/_internal/fx/op_validation.py b/torch/onnx/_internal/fx/op_validation.py
index b306bc2141de..01161aee25ea 100644
--- a/torch/onnx/_internal/fx/op_validation.py
+++ b/torch/onnx/_internal/fx/op_validation.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Module for handling op-level validation during exporting."""
 
 from __future__ import annotations
diff --git a/torch/onnx/_internal/fx/passes/_utils.py b/torch/onnx/_internal/fx/passes/_utils.py
index 92a883469a52..6e49bccfcfaf 100644
--- a/torch/onnx/_internal/fx/passes/_utils.py
+++ b/torch/onnx/_internal/fx/passes/_utils.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Common utility functions for FX passes.
 
 These functions should NOT be directly invoked outside of `passes` package.
diff --git a/torch/onnx/_internal/fx/passes/decomp.py b/torch/onnx/_internal/fx/passes/decomp.py
index b9a131b97466..5185b1152485 100644
--- a/torch/onnx/_internal/fx/passes/decomp.py
+++ b/torch/onnx/_internal/fx/passes/decomp.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 import contextlib
diff --git a/torch/onnx/_internal/fx/passes/functionalization.py b/torch/onnx/_internal/fx/passes/functionalization.py
index 21f2691cbb8e..dfdee6e88c85 100644
--- a/torch/onnx/_internal/fx/passes/functionalization.py
+++ b/torch/onnx/_internal/fx/passes/functionalization.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 import contextlib
diff --git a/torch/onnx/_internal/fx/passes/modularization.py b/torch/onnx/_internal/fx/passes/modularization.py
index b7c3b90cab66..6e1352f73046 100644
--- a/torch/onnx/_internal/fx/passes/modularization.py
+++ b/torch/onnx/_internal/fx/passes/modularization.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 import abc
diff --git a/torch/onnx/_internal/fx/passes/readability.py b/torch/onnx/_internal/fx/passes/readability.py
index 64887ad2ee6e..2b3518b79ea6 100644
--- a/torch/onnx/_internal/fx/passes/readability.py
+++ b/torch/onnx/_internal/fx/passes/readability.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 from typing import Dict, List, Sequence, Tuple, Union
diff --git a/torch/onnx/_internal/fx/passes/type_promotion.py b/torch/onnx/_internal/fx/passes/type_promotion.py
index 944cad4acf1c..bc584ff32925 100644
--- a/torch/onnx/_internal/fx/passes/type_promotion.py
+++ b/torch/onnx/_internal/fx/passes/type_promotion.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 # Owner(s): ["module: onnx"]
 from __future__ import annotations
 
diff --git a/torch/onnx/_internal/fx/passes/virtualization.py b/torch/onnx/_internal/fx/passes/virtualization.py
index 66ca69d7a70f..cd77b6eec18b 100644
--- a/torch/onnx/_internal/fx/passes/virtualization.py
+++ b/torch/onnx/_internal/fx/passes/virtualization.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 from typing import List, Optional, Tuple
diff --git a/torch/onnx/_internal/fx/patcher.py b/torch/onnx/_internal/fx/patcher.py
index ee919eae00d1..dbd8fb591126 100644
--- a/torch/onnx/_internal/fx/patcher.py
+++ b/torch/onnx/_internal/fx/patcher.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import copy
 import functools
 import io
diff --git a/torch/onnx/_internal/fx/serialization.py b/torch/onnx/_internal/fx/serialization.py
index 726bf4219330..5739442163a3 100644
--- a/torch/onnx/_internal/fx/serialization.py
+++ b/torch/onnx/_internal/fx/serialization.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 import io
diff --git a/torch/onnx/_internal/fx/torch_export_graph_extractor.py b/torch/onnx/_internal/fx/torch_export_graph_extractor.py
index fb3f0e99a6d6..a825e466f1aa 100644
--- a/torch/onnx/_internal/fx/torch_export_graph_extractor.py
+++ b/torch/onnx/_internal/fx/torch_export_graph_extractor.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 # NOTE: This file is referenced by name at
 #       /opt/pytorch/torch/_dynamo/eval_frame.py::DONT_WRAP_FILES.
 #       introduced by https://github.com/pytorch/pytorch/pull/98894.
diff --git a/torch/onnx/_internal/fx/type_utils.py b/torch/onnx/_internal/fx/type_utils.py
index 90abdc244d99..3aac02a51214 100644
--- a/torch/onnx/_internal/fx/type_utils.py
+++ b/torch/onnx/_internal/fx/type_utils.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Utilities for converting and operating on ONNX, JIT and torch types."""
 from __future__ import annotations
 
diff --git a/torch/onnx/_internal/io_adapter.py b/torch/onnx/_internal/io_adapter.py
index 2f8c9202d7bb..12100d0f489c 100644
--- a/torch/onnx/_internal/io_adapter.py
+++ b/torch/onnx/_internal/io_adapter.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 import inspect
diff --git a/torch/onnx/_internal/jit_utils.py b/torch/onnx/_internal/jit_utils.py
index 719f4b0c16e8..13ae4209da5d 100644
--- a/torch/onnx/_internal/jit_utils.py
+++ b/torch/onnx/_internal/jit_utils.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Utilities for manipulating the torch.Graph object and the torchscript."""
 from __future__ import annotations
 
diff --git a/torch/onnx/_internal/onnx_proto_utils.py b/torch/onnx/_internal/onnx_proto_utils.py
index 278af3feacc6..40eb1bd8d64e 100644
--- a/torch/onnx/_internal/onnx_proto_utils.py
+++ b/torch/onnx/_internal/onnx_proto_utils.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Utilities for manipulating the onnx and onnx-script dependencies and ONNX proto."""
 
 from __future__ import annotations
diff --git a/torch/onnx/_internal/onnxruntime.py b/torch/onnx/_internal/onnxruntime.py
index aa3495ee5ac5..d8a7e55e8f9e 100644
--- a/torch/onnx/_internal/onnxruntime.py
+++ b/torch/onnx/_internal/onnxruntime.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import dataclasses
 import importlib
 import logging
diff --git a/torch/onnx/_internal/registration.py b/torch/onnx/_internal/registration.py
index 3b2e68e1e40a..f051708f864d 100644
--- a/torch/onnx/_internal/registration.py
+++ b/torch/onnx/_internal/registration.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Module for handling symbolic function registration."""
 
 import warnings
diff --git a/torch/onnx/_onnx_supported_ops.py b/torch/onnx/_onnx_supported_ops.py
index 2611b0d81e9b..e2707298d6d9 100644
--- a/torch/onnx/_onnx_supported_ops.py
+++ b/torch/onnx/_onnx_supported_ops.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import inspect
 from typing import Dict, List, Union
 
diff --git a/torch/onnx/_type_utils.py b/torch/onnx/_type_utils.py
index d13232507317..d9b647c807f3 100644
--- a/torch/onnx/_type_utils.py
+++ b/torch/onnx/_type_utils.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Utilities for converting and operating on ONNX, JIT and torch types."""
 from __future__ import annotations
 
diff --git a/torch/onnx/operators.py b/torch/onnx/operators.py
index 489010519980..1e7532e8451d 100644
--- a/torch/onnx/operators.py
+++ b/torch/onnx/operators.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 r"""This file provides a location for operators that help exporting models via onnx.
 
 E.g. `shape_as_tensor` and `reshape_from_tensor_shape`
diff --git a/torch/onnx/symbolic_caffe2.py b/torch/onnx/symbolic_caffe2.py
index 3398fcd2fe10..ed2dc6cd9fdb 100644
--- a/torch/onnx/symbolic_caffe2.py
+++ b/torch/onnx/symbolic_caffe2.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import importlib
 import inspect
 
diff --git a/torch/onnx/symbolic_helper.py b/torch/onnx/symbolic_helper.py
index 4430babaef00..676c3d68048b 100644
--- a/torch/onnx/symbolic_helper.py
+++ b/torch/onnx/symbolic_helper.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 import functools
diff --git a/torch/onnx/symbolic_opset10.py b/torch/onnx/symbolic_opset10.py
index 6fd576822e2c..e9ba8b4015f2 100644
--- a/torch/onnx/symbolic_opset10.py
+++ b/torch/onnx/symbolic_opset10.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 import functools
diff --git a/torch/onnx/symbolic_opset11.py b/torch/onnx/symbolic_opset11.py
index 99d5064ad7a0..e562d5a47567 100644
--- a/torch/onnx/symbolic_opset11.py
+++ b/torch/onnx/symbolic_opset11.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """This file exports ONNX ops for opset 11."""
 from __future__ import annotations
 
diff --git a/torch/onnx/symbolic_opset12.py b/torch/onnx/symbolic_opset12.py
index 130b02a889b0..5a6bf720df36 100644
--- a/torch/onnx/symbolic_opset12.py
+++ b/torch/onnx/symbolic_opset12.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from __future__ import annotations
 
 import functools
diff --git a/torch/onnx/symbolic_opset13.py b/torch/onnx/symbolic_opset13.py
index 5bba817bbce0..bb7045c0f58b 100644
--- a/torch/onnx/symbolic_opset13.py
+++ b/torch/onnx/symbolic_opset13.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 # EDITING THIS FILE? READ THIS FIRST!
 # see Note [Edit Symbolic Files] in README.md
 
diff --git a/torch/onnx/symbolic_opset14.py b/torch/onnx/symbolic_opset14.py
index 1b4b8ee7917c..62e05910dd72 100644
--- a/torch/onnx/symbolic_opset14.py
+++ b/torch/onnx/symbolic_opset14.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """This file exports ONNX ops for opset 14.
 
 Note [ONNX operators that are added/updated in opset 14]
diff --git a/torch/onnx/symbolic_opset15.py b/torch/onnx/symbolic_opset15.py
index 4f316a77f62e..793c1cad8fb9 100644
--- a/torch/onnx/symbolic_opset15.py
+++ b/torch/onnx/symbolic_opset15.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """This file exports ONNX ops for opset 15.
 
 Note [ONNX operators that are added/updated in opset 15]
diff --git a/torch/onnx/symbolic_opset16.py b/torch/onnx/symbolic_opset16.py
index 24306b475366..cd5829ada850 100644
--- a/torch/onnx/symbolic_opset16.py
+++ b/torch/onnx/symbolic_opset16.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """This file exports ONNX ops for opset 16.
 
 Note [ONNX Operators that are added/updated in opset 16]
diff --git a/torch/onnx/symbolic_opset17.py b/torch/onnx/symbolic_opset17.py
index c7720b9e5c9f..44c789017d75 100644
--- a/torch/onnx/symbolic_opset17.py
+++ b/torch/onnx/symbolic_opset17.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """This file exports ONNX ops for opset 17.
 
 Note [ONNX Operators that are added/updated in opset 17]
diff --git a/torch/onnx/symbolic_opset18.py b/torch/onnx/symbolic_opset18.py
index d80361dd417f..68e14c987731 100644
--- a/torch/onnx/symbolic_opset18.py
+++ b/torch/onnx/symbolic_opset18.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """This file exports ONNX ops for opset 18.
 
 Note [ONNX Operators that are added/updated in opset 18]
diff --git a/torch/onnx/symbolic_opset20.py b/torch/onnx/symbolic_opset20.py
index 9c81bc3e3c49..9557b5f2828e 100644
--- a/torch/onnx/symbolic_opset20.py
+++ b/torch/onnx/symbolic_opset20.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """This file exports ONNX ops for opset 20.
 
 Note [ONNX Operators that are added/updated in opset 20]
diff --git a/torch/onnx/symbolic_opset7.py b/torch/onnx/symbolic_opset7.py
index 0537e8a92888..c647ead4e297 100644
--- a/torch/onnx/symbolic_opset7.py
+++ b/torch/onnx/symbolic_opset7.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """
 Note [ONNX operators that are added/updated from opset 7 to opset 8]
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
diff --git a/torch/onnx/symbolic_opset8.py b/torch/onnx/symbolic_opset8.py
index b2fbee3b9784..87b4be230e78 100644
--- a/torch/onnx/symbolic_opset8.py
+++ b/torch/onnx/symbolic_opset8.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """
 Note [ONNX operators that are added/updated from opset 8 to opset 9]
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
diff --git a/torch/onnx/symbolic_opset9.py b/torch/onnx/symbolic_opset9.py
index adfa538c8f08..f71ef713636a 100644
--- a/torch/onnx/symbolic_opset9.py
+++ b/torch/onnx/symbolic_opset9.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """This file exports ONNX ops for opset 9.
 
 Opset 9 is supported by ONNX release 1.4.1
diff --git a/torch/onnx/utils.py b/torch/onnx/utils.py
index 191df45ac9ef..94a57786a4bd 100644
--- a/torch/onnx/utils.py
+++ b/torch/onnx/utils.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Functions to export models into the ONNX IR format.
 
 These models can be loaded with the ONNX library and then
diff --git a/torch/onnx/verification.py b/torch/onnx/verification.py
index 6b49e7fc72b9..95ed873bf633 100644
--- a/torch/onnx/verification.py
+++ b/torch/onnx/verification.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Functions to verify exported ONNX model is functionally equivalent to original PyTorch model.
 
 ONNX Runtime is required, and is used as the ONNX backend for export verification.
diff --git a/torch/optim/_functional.py b/torch/optim/_functional.py
index 4a6198956fb8..a307cc76846d 100644
--- a/torch/optim/_functional.py
+++ b/torch/optim/_functional.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 r"""Functional interface."""
 import math
 from typing import List
diff --git a/torch/optim/adadelta.py b/torch/optim/adadelta.py
index 4d1a4e25319c..eff24f159213 100644
--- a/torch/optim/adadelta.py
+++ b/torch/optim/adadelta.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import Any, Dict, List, Optional
 
 import torch
diff --git a/torch/optim/adagrad.py b/torch/optim/adagrad.py
index a95e985b49eb..0b6dfe852d08 100644
--- a/torch/optim/adagrad.py
+++ b/torch/optim/adagrad.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import List, Optional
 
 import torch
diff --git a/torch/optim/adam.py b/torch/optim/adam.py
index 1c625682fc34..bff29613175a 100644
--- a/torch/optim/adam.py
+++ b/torch/optim/adam.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import List, Optional, Tuple, Union
 
 import torch
diff --git a/torch/optim/adamax.py b/torch/optim/adamax.py
index 005327d8bb88..c2e39b788014 100644
--- a/torch/optim/adamax.py
+++ b/torch/optim/adamax.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import List, Optional, Tuple, Union
 
 import torch
diff --git a/torch/optim/adamw.py b/torch/optim/adamw.py
index 707ac17c361c..2292c17f9e0f 100644
--- a/torch/optim/adamw.py
+++ b/torch/optim/adamw.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import cast, List, Optional, Tuple, Union
 
 import torch
diff --git a/torch/optim/asgd.py b/torch/optim/asgd.py
index 633a14832282..454772670904 100644
--- a/torch/optim/asgd.py
+++ b/torch/optim/asgd.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import List, Optional, Tuple, Union
 
 import torch
diff --git a/torch/optim/lbfgs.py b/torch/optim/lbfgs.py
index e8818cca538c..480b45c84d72 100644
--- a/torch/optim/lbfgs.py
+++ b/torch/optim/lbfgs.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import Optional
 
 import torch
diff --git a/torch/optim/lr_scheduler.py b/torch/optim/lr_scheduler.py
index cb7d9738df5a..4a5f162a0b20 100644
--- a/torch/optim/lr_scheduler.py
+++ b/torch/optim/lr_scheduler.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import math
 import types
 import warnings
diff --git a/torch/optim/nadam.py b/torch/optim/nadam.py
index fd1f8ab0e718..a4a6d07b2ca8 100644
--- a/torch/optim/nadam.py
+++ b/torch/optim/nadam.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import cast, List, Optional, Tuple, Union
 
 import torch
diff --git a/torch/optim/optimizer.py b/torch/optim/optimizer.py
index fc091e273c36..498669e65fbd 100644
--- a/torch/optim/optimizer.py
+++ b/torch/optim/optimizer.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import functools
 import math
 import warnings
diff --git a/torch/optim/radam.py b/torch/optim/radam.py
index 619f10493587..d21973caeceb 100644
--- a/torch/optim/radam.py
+++ b/torch/optim/radam.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import cast, List, Optional, Tuple, Union
 
 import torch
diff --git a/torch/optim/rmsprop.py b/torch/optim/rmsprop.py
index bdc3ec0b8b3f..30b56779fc75 100644
--- a/torch/optim/rmsprop.py
+++ b/torch/optim/rmsprop.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import List, Optional
 
 import torch
diff --git a/torch/optim/rprop.py b/torch/optim/rprop.py
index af1854cc518a..69043b48673e 100644
--- a/torch/optim/rprop.py
+++ b/torch/optim/rprop.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import List, Optional, Tuple
 
 import torch
diff --git a/torch/optim/sgd.py b/torch/optim/sgd.py
index 291b4068dd4c..e682d83701d5 100644
--- a/torch/optim/sgd.py
+++ b/torch/optim/sgd.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import List, Optional
 
 import torch
diff --git a/torch/optim/sparse_adam.py b/torch/optim/sparse_adam.py
index 88643d1a5646..adb7c17629c2 100644
--- a/torch/optim/sparse_adam.py
+++ b/torch/optim/sparse_adam.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from typing import List, Tuple
 
 import torch
diff --git a/torch/optim/swa_utils.py b/torch/optim/swa_utils.py
index 4cfca073af77..440897e6041e 100644
--- a/torch/optim/swa_utils.py
+++ b/torch/optim/swa_utils.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import itertools
 import math
 import warnings
diff --git a/torch/package/_digraph.py b/torch/package/_digraph.py
index f84a51398f00..8b753f7ebdc4 100644
--- a/torch/package/_digraph.py
+++ b/torch/package/_digraph.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 from collections import deque
 from typing import List, Set
 
diff --git a/torch/package/_directory_reader.py b/torch/package/_directory_reader.py
index cec5333c3e3f..77d629cccce2 100644
--- a/torch/package/_directory_reader.py
+++ b/torch/package/_directory_reader.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import os.path
 from glob import glob
 from typing import cast
diff --git a/torch/package/_importlib.py b/torch/package/_importlib.py
index fd303b6141e7..9741925315e5 100644
--- a/torch/package/_importlib.py
+++ b/torch/package/_importlib.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 import _warnings
 import os.path
 
diff --git a/torch/package/_mangling.py b/torch/package/_mangling.py
index 0876d64664a2..7dcf3538631f 100644
--- a/torch/package/_mangling.py
+++ b/torch/package/_mangling.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 """Import mangling.
 See mangling.md for details.
 """
diff --git a/torch/package/_mock.py b/torch/package/_mock.py
index b0bdb95cc48c..44876b1a1d3f 100644
--- a/torch/package/_mock.py
+++ b/torch/package/_mock.py
@@ -1,3 +1,4 @@
+# mypy: allow-untyped-defs
 _magic_methods = [
     "__subclasscheck__",
     "__hex__",