diff --git a/.clang-format b/.clang-format
index a10c270ac0a..4c5d0907047 100644
--- a/.clang-format
+++ b/.clang-format
@@ -102,6 +102,10 @@ IncludeBlocks:   Regroup
 IncludeCategories:
   - Regex:           '^<torch/.*>'
     Priority:        3
+  - Regex:           '^<ATen/.*>'
+    Priority:        3
+  - Regex:           '^<c10/.*>'
+    Priority:        3
   - Regex:           '^<.*\.h>'
     Priority:        1
   - Regex:           '^<.*'
diff --git a/test/cpp/test_replication.cpp b/test/cpp/test_replication.cpp
index b59f4805a15..3d3d44119da 100644
--- a/test/cpp/test_replication.cpp
+++ b/test/cpp/test_replication.cpp
@@ -1,8 +1,9 @@
-#include <ATen/ATen.h>
 #include <gtest/gtest.h>
 
 #include <iostream>
 
+#include <ATen/ATen.h>
+
 #include "absl/synchronization/blocking_counter.h"
 #include "xla/hlo/builder/xla_builder.h"
 #include "xla/shape_util.h"
diff --git a/test/cpp/test_status_common.h b/test/cpp/test_status_common.h
index 1c1a09b3c97..4b00d7dc5b6 100644
--- a/test/cpp/test_status_common.h
+++ b/test/cpp/test_status_common.h
@@ -18,7 +18,6 @@
 #ifndef XLA_TEST_CPP_TEST_STATUS_COMMON_H_
 #define XLA_TEST_CPP_TEST_STATUS_COMMON_H_
 
-#include <c10/util/Exception.h>
 #include <gmock/gmock.h>
 #include <gtest/gtest.h>
 
@@ -27,6 +26,8 @@
 #include <stdexcept>
 #include <utility>
 
+#include <c10/util/Exception.h>
+
 #include "absl/status/status.h"
 #include "absl/status/statusor.h"
 
diff --git a/test/cpp/test_tensor.cpp b/test/cpp/test_tensor.cpp
index 1328f5fc528..f2a74b26d23 100644
--- a/test/cpp/test_tensor.cpp
+++ b/test/cpp/test_tensor.cpp
@@ -1,9 +1,9 @@
-#include <ATen/ATen.h>
 #include <gtest/gtest.h>
 
 #include <limits>
 #include <vector>
 
+#include <ATen/ATen.h>
 #include <torch/csrc/autograd/variable.h>
 
 #include "absl/base/nullability.h"
diff --git a/test/cpp/test_xla_sharding.cpp b/test/cpp/test_xla_sharding.cpp
index f544292301d..57ab2de44e2 100644
--- a/test/cpp/test_xla_sharding.cpp
+++ b/test/cpp/test_xla_sharding.cpp
@@ -1,10 +1,10 @@
-#include <ATen/ATen.h>
 #include <google/protobuf/repeated_field.h>
 #include <gtest/gtest.h>
 #include <stdlib.h>
 
 #include <iostream>
 
+#include <ATen/ATen.h>
 #include <torch/csrc/lazy/core/lazy_graph_executor.h>
 
 #include "xla/protobuf_util.h"
diff --git a/torch_xla/csrc/aten_fallback.cpp b/torch_xla/csrc/aten_fallback.cpp
index 9feb75e22ce..b610f040bb5 100644
--- a/torch_xla/csrc/aten_fallback.cpp
+++ b/torch_xla/csrc/aten_fallback.cpp
@@ -1,13 +1,12 @@
 #include "torch_xla/csrc/aten_fallback.h"
 
+#include <unordered_map>
+#include <vector>
+
 #include <ATen/DLConvertor.h>
 #include <ATen/native/CPUFallback.h>
 #include <ATen/ops/_copy_from_and_resize.h>
 #include <ATen/ops/_to_cpu.h>
-
-#include <unordered_map>
-#include <vector>
-
 #include <torch/csrc/utils/device_lazy_init.h>
 
 #include "torch_xla/csrc/function_call_tracker.h"
diff --git a/torch_xla/csrc/aten_xla_bridge.cpp b/torch_xla/csrc/aten_xla_bridge.cpp
index a9a9204342d..9fd9634fd76 100644
--- a/torch_xla/csrc/aten_xla_bridge.cpp
+++ b/torch_xla/csrc/aten_xla_bridge.cpp
@@ -1,11 +1,10 @@
 #include "torch_xla/csrc/aten_xla_bridge.h"
 
-#include <ATen/FunctionalTensorWrapper.h>
-
 #include <map>
 #include <string>
 #include <vector>
 
+#include <ATen/FunctionalTensorWrapper.h>
 #include <torch/csrc/lazy/core/tensor_util.h>
 
 #include "absl/log/absl_check.h"
diff --git a/torch_xla/csrc/aten_xla_bridge.h b/torch_xla/csrc/aten_xla_bridge.h
index 440bdf59d92..f22a68ca49b 100644
--- a/torch_xla/csrc/aten_xla_bridge.h
+++ b/torch_xla/csrc/aten_xla_bridge.h
@@ -1,12 +1,12 @@
 #ifndef XLA_TORCH_XLA_CSRC_ATEN_XLA_BRIDGE_H_
 #define XLA_TORCH_XLA_CSRC_ATEN_XLA_BRIDGE_H_
 
+#include <vector>
+
 #include <ATen/Device.h>
 #include <ATen/Functions.h>
 #include <ATen/Tensor.h>
 
-#include <vector>
-
 #include "absl/base/nullability.h"
 #include "absl/status/statusor.h"
 #include "absl/types/span.h"
diff --git a/torch_xla/csrc/aten_xla_type.cpp b/torch_xla/csrc/aten_xla_type.cpp
index bf7a4003acc..80e30a70072 100644
--- a/torch_xla/csrc/aten_xla_type.cpp
+++ b/torch_xla/csrc/aten_xla_type.cpp
@@ -1,3 +1,7 @@
+#include <iterator>
+#include <mutex>
+#include <optional>
+
 #include <ATen/ExpandUtils.h>
 #include <ATen/FunctionalTensorWrapper.h>
 #include <ATen/MetaFunctions.h>
@@ -10,11 +14,6 @@
 #include <ATen/ops/_embedding_bag_backward_native.h>
 #include <ATen/ops/expand_copy.h>
 #include <c10/core/Contiguity.h>
-
-#include <iterator>
-#include <mutex>
-#include <optional>
-
 #include <torch/csrc/lazy/core/helpers.h>
 #include <torch/csrc/lazy/core/shape_inference.h>
 #include <torch/csrc/lazy/core/tensor_util.h>
diff --git a/torch_xla/csrc/autocast_mode.cpp b/torch_xla/csrc/autocast_mode.cpp
index 0baef83f20a..4db31feb6ba 100644
--- a/torch_xla/csrc/autocast_mode.cpp
+++ b/torch_xla/csrc/autocast_mode.cpp
@@ -4,7 +4,6 @@
 #include <ATen/autocast_mode.h>
 #include <c10/core/impl/LocalDispatchKeySet.h>
 #include <c10/util/intrusive_ptr.h>
-
 #include <torch/library.h>
 
 namespace at {
diff --git a/torch_xla/csrc/convert_ops.h b/torch_xla/csrc/convert_ops.h
index 3ad627d559c..4247de3b16a 100644
--- a/torch_xla/csrc/convert_ops.h
+++ b/torch_xla/csrc/convert_ops.h
@@ -1,10 +1,10 @@
 #ifndef XLA_TORCH_XLA_CSRC_CONVERT_OPS_H_
 #define XLA_TORCH_XLA_CSRC_CONVERT_OPS_H_
 
-#include <c10/core/ScalarType.h>
-
 #include <optional>
 
+#include <c10/core/ScalarType.h>
+
 #include "xla/hlo/builder/xla_builder.h"
 #include "xla/types.h"
 
diff --git a/torch_xla/csrc/data_ops.h b/torch_xla/csrc/data_ops.h
index c068d578842..8af661415fd 100644
--- a/torch_xla/csrc/data_ops.h
+++ b/torch_xla/csrc/data_ops.h
@@ -1,10 +1,10 @@
 #ifndef XLA_TORCH_XLA_CSRC_DATA_OPS_H_
 #define XLA_TORCH_XLA_CSRC_DATA_OPS_H_
 
-#include <c10/core/ScalarType.h>
-
 #include <vector>
 
+#include <c10/core/ScalarType.h>
+
 #include "absl/types/optional.h"
 #include "absl/types/span.h"
 #include "xla/hlo/builder/xla_builder.h"
diff --git a/torch_xla/csrc/dl_convertor.cpp b/torch_xla/csrc/dl_convertor.cpp
index 845c15696e8..c978bb848c5 100644
--- a/torch_xla/csrc/dl_convertor.cpp
+++ b/torch_xla/csrc/dl_convertor.cpp
@@ -1,11 +1,11 @@
 #include "torch_xla/csrc/dl_convertor.h"
 
-#include <ATen/DLConvertor.h>
-
 #include <memory>
 #include <utility>
 #include <vector>
 
+#include <ATen/DLConvertor.h>
+
 #include "absl/log/absl_check.h"
 #include "absl/status/status.h"
 #include "absl/types/span.h"
diff --git a/torch_xla/csrc/helpers.h b/torch_xla/csrc/helpers.h
index 55a03447851..c01820dc27a 100644
--- a/torch_xla/csrc/helpers.h
+++ b/torch_xla/csrc/helpers.h
@@ -1,13 +1,12 @@
 #ifndef XLA_TORCH_XLA_CSRC_HELPERS_H_
 #define XLA_TORCH_XLA_CSRC_HELPERS_H_
 
-#include <c10/core/Scalar.h>
-
 #include <functional>
 #include <optional>
 #include <tuple>
 #include <vector>
 
+#include <c10/core/Scalar.h>
 #include <torch/csrc/lazy/core/shape.h>
 #include <torch/csrc/lazy/core/util.h>
 
diff --git a/torch_xla/csrc/init_python_bindings.cpp b/torch_xla/csrc/init_python_bindings.cpp
index ce6012c1073..253c569e3a2 100644
--- a/torch_xla/csrc/init_python_bindings.cpp
+++ b/torch_xla/csrc/init_python_bindings.cpp
@@ -1,7 +1,4 @@
-#include <ATen/dlpack.h>
 #include <Python.h>
-#include <c10/core/Device.h>
-#include <c10/util/Metaprogramming.h>
 #include <google/protobuf/text_format.h>
 
 #include <algorithm>
@@ -17,6 +14,9 @@
 #include <unordered_map>
 #include <vector>
 
+#include <ATen/dlpack.h>
+#include <c10/core/Device.h>
+#include <c10/util/Metaprogramming.h>
 #include <torch/csrc/Exceptions.h>
 #include <torch/csrc/autograd/utils/wrap_outputs.h>
 #include <torch/csrc/autograd/variable.h>
diff --git a/torch_xla/csrc/ir.h b/torch_xla/csrc/ir.h
index 17944b3c73d..eea82a8fe2b 100644
--- a/torch_xla/csrc/ir.h
+++ b/torch_xla/csrc/ir.h
@@ -1,8 +1,6 @@
 #ifndef XLA_TORCH_XLA_CSRC_IR_H_
 #define XLA_TORCH_XLA_CSRC_IR_H_
 
-#include <ATen/core/interned_strings.h>
-
 #include <functional>
 #include <iostream>
 #include <memory>
@@ -12,6 +10,7 @@
 #include <utility>
 #include <vector>
 
+#include <ATen/core/interned_strings.h>
 #include <torch/csrc/lazy/core/hash.h>
 #include <torch/csrc/lazy/core/ir.h>
 #include <torch/csrc/lazy/core/ir_builder.h>
diff --git a/torch_xla/csrc/ops/cast.h b/torch_xla/csrc/ops/cast.h
index 73490cdf1da..4785cfd03f7 100644
--- a/torch_xla/csrc/ops/cast.h
+++ b/torch_xla/csrc/ops/cast.h
@@ -1,10 +1,10 @@
 #ifndef XLA_TORCH_XLA_CSRC_OPS_CAST_H_
 #define XLA_TORCH_XLA_CSRC_OPS_CAST_H_
 
-#include <c10/core/ScalarType.h>
-
 #include <optional>
 
+#include <c10/core/ScalarType.h>
+
 #include "torch_xla/csrc/ir.h"
 
 namespace torch_xla {
diff --git a/torch_xla/csrc/ops/cumprod.h b/torch_xla/csrc/ops/cumprod.h
index c650505e2f5..b948781c465 100644
--- a/torch_xla/csrc/ops/cumprod.h
+++ b/torch_xla/csrc/ops/cumprod.h
@@ -1,10 +1,10 @@
 #ifndef XLA_TORCH_XLA_CSRC_OPS_CUMPROD_H_
 #define XLA_TORCH_XLA_CSRC_OPS_CUMPROD_H_
 
-#include <c10/core/ScalarType.h>
-
 #include <optional>
 
+#include <c10/core/ScalarType.h>
+
 #include "torch_xla/csrc/ir.h"
 
 namespace torch_xla {
diff --git a/torch_xla/csrc/ops/cumsum.h b/torch_xla/csrc/ops/cumsum.h
index 76832378f61..682d4e6db12 100644
--- a/torch_xla/csrc/ops/cumsum.h
+++ b/torch_xla/csrc/ops/cumsum.h
@@ -1,10 +1,10 @@
 #ifndef XLA_TORCH_XLA_CSRC_OPS_CUMSUM_H_
 #define XLA_TORCH_XLA_CSRC_OPS_CUMSUM_H_
 
-#include <c10/core/ScalarType.h>
-
 #include <optional>
 
+#include <c10/core/ScalarType.h>
+
 #include "torch_xla/csrc/ir.h"
 
 namespace torch_xla {
diff --git a/torch_xla/csrc/ops/dot_general.cpp b/torch_xla/csrc/ops/dot_general.cpp
index 7a54418fda1..345bfc3a08f 100644
--- a/torch_xla/csrc/ops/dot_general.cpp
+++ b/torch_xla/csrc/ops/dot_general.cpp
@@ -1,7 +1,6 @@
 #include "torch_xla/csrc/ops/dot_general.h"
 
 #include <c10/core/ScalarType.h>
-
 #include <torch/csrc/lazy/core/tensor_util.h>
 
 #include "torch_xla/csrc/dtype.h"
diff --git a/torch_xla/csrc/ops/index_ops.cpp b/torch_xla/csrc/ops/index_ops.cpp
index ba306c373d7..ccca4ba9715 100644
--- a/torch_xla/csrc/ops/index_ops.cpp
+++ b/torch_xla/csrc/ops/index_ops.cpp
@@ -3,7 +3,6 @@
 #include <ATen/ExpandUtils.h>
 #include <ATen/Functions.h>
 #include <ATen/ops/select_copy.h>
-
 #include <torch/csrc/lazy/core/util.h>
 
 #include "xla/permutation_util.h"
diff --git a/torch_xla/csrc/ops/log_softmax.h b/torch_xla/csrc/ops/log_softmax.h
index 068080721ba..c8abd0363ce 100644
--- a/torch_xla/csrc/ops/log_softmax.h
+++ b/torch_xla/csrc/ops/log_softmax.h
@@ -1,10 +1,10 @@
 #ifndef XLA_TORCH_XLA_CSRC_OPS_LOG_SOFTMAX_H_
 #define XLA_TORCH_XLA_CSRC_OPS_LOG_SOFTMAX_H_
 
-#include <c10/core/ScalarType.h>
-
 #include <optional>
 
+#include <c10/core/ScalarType.h>
+
 #include "torch_xla/csrc/ir.h"
 
 namespace torch_xla {
diff --git a/torch_xla/csrc/ops/mean.h b/torch_xla/csrc/ops/mean.h
index 380ba9eef70..244acdbd951 100644
--- a/torch_xla/csrc/ops/mean.h
+++ b/torch_xla/csrc/ops/mean.h
@@ -1,11 +1,11 @@
 #ifndef XLA_TORCH_XLA_CSRC_OPS_MEAN_H_
 #define XLA_TORCH_XLA_CSRC_OPS_MEAN_H_
 
-#include <c10/core/ScalarType.h>
-
 #include <optional>
 #include <vector>
 
+#include <c10/core/ScalarType.h>
+
 #include "xla/types.h"
 
 #include "torch_xla/csrc/ir.h"
diff --git a/torch_xla/csrc/ops/mse_loss.cpp b/torch_xla/csrc/ops/mse_loss.cpp
index 196865423a9..c816c32e4f5 100644
--- a/torch_xla/csrc/ops/mse_loss.cpp
+++ b/torch_xla/csrc/ops/mse_loss.cpp
@@ -1,7 +1,6 @@
 #include "torch_xla/csrc/ops/mse_loss.h"
 
 #include <ATen/core/Reduction.h>
-
 #include <torch/csrc/lazy/core/util.h>
 
 #include "torch_xla/csrc/lowering_context.h"
diff --git a/torch_xla/csrc/ops/prod.h b/torch_xla/csrc/ops/prod.h
index a2153aba2b2..0da69b58790 100644
--- a/torch_xla/csrc/ops/prod.h
+++ b/torch_xla/csrc/ops/prod.h
@@ -1,10 +1,10 @@
 #ifndef XLA_TORCH_XLA_CSRC_OPS_PROD_H_
 #define XLA_TORCH_XLA_CSRC_OPS_PROD_H_
 
-#include <c10/core/ScalarType.h>
-
 #include <optional>
 
+#include <c10/core/ScalarType.h>
+
 #include "absl/types/span.h"
 
 #include "torch_xla/csrc/ir.h"
diff --git a/torch_xla/csrc/ops/scalar.h b/torch_xla/csrc/ops/scalar.h
index deae74a4c6f..183edd27981 100644
--- a/torch_xla/csrc/ops/scalar.h
+++ b/torch_xla/csrc/ops/scalar.h
@@ -1,11 +1,11 @@
 #ifndef XLA_TORCH_XLA_CSRC_OPS_SCALAR_H_
 #define XLA_TORCH_XLA_CSRC_OPS_SCALAR_H_
 
+#include <iostream>
+
 #include <ATen/core/Formatting.h>
 #include <c10/core/Scalar.h>
 
-#include <iostream>
-
 #include "torch_xla/csrc/ir.h"
 #include "torch_xla/csrc/runtime/types.h"
 
diff --git a/torch_xla/csrc/ops/softmax.h b/torch_xla/csrc/ops/softmax.h
index 10fea7be9dc..36c9d7b2174 100644
--- a/torch_xla/csrc/ops/softmax.h
+++ b/torch_xla/csrc/ops/softmax.h
@@ -1,10 +1,10 @@
 #ifndef XLA_TORCH_XLA_CSRC_OPS_SOFTMAX_H_
 #define XLA_TORCH_XLA_CSRC_OPS_SOFTMAX_H_
 
-#include <c10/core/ScalarType.h>
-
 #include <optional>
 
+#include <c10/core/ScalarType.h>
+
 #include "torch_xla/csrc/ir.h"
 
 namespace torch_xla {
diff --git a/torch_xla/csrc/ops/sum.h b/torch_xla/csrc/ops/sum.h
index 67ecfdc0542..2eb890592fc 100644
--- a/torch_xla/csrc/ops/sum.h
+++ b/torch_xla/csrc/ops/sum.h
@@ -1,10 +1,10 @@
 #ifndef XLA_TORCH_XLA_CSRC_OPS_SUM_H_
 #define XLA_TORCH_XLA_CSRC_OPS_SUM_H_
 
-#include <c10/core/ScalarType.h>
-
 #include <optional>
 
+#include <c10/core/ScalarType.h>
+
 #include "absl/types/span.h"
 
 #include "torch_xla/csrc/ir.h"
diff --git a/torch_xla/csrc/reduction.cpp b/torch_xla/csrc/reduction.cpp
index 9e759aef700..71cdb606c6a 100644
--- a/torch_xla/csrc/reduction.cpp
+++ b/torch_xla/csrc/reduction.cpp
@@ -1,10 +1,9 @@
 #include "torch_xla/csrc/reduction.h"
 
-#include <ATen/core/Reduction.h>
-
 #include <cmath>
 #include <unordered_set>
 
+#include <ATen/core/Reduction.h>
 #include <torch/csrc/lazy/core/helpers.h>
 #include <torch/csrc/lazy/core/util.h>
 
diff --git a/torch_xla/csrc/runtime/computation_client.h b/torch_xla/csrc/runtime/computation_client.h
index a045a01e548..66466997958 100644
--- a/torch_xla/csrc/runtime/computation_client.h
+++ b/torch_xla/csrc/runtime/computation_client.h
@@ -1,8 +1,6 @@
 #ifndef XLA_CLIENT_COMPUTATION_CLIENT_H_
 #define XLA_CLIENT_COMPUTATION_CLIENT_H_
 
-#include <ATen/Tensor.h>
-
 #include <algorithm>
 #include <cmath>
 #include <map>
@@ -10,6 +8,7 @@
 #include <string>
 #include <vector>
 
+#include <ATen/Tensor.h>
 #include <torch/csrc/lazy/backend/backend_data.h>
 #include <torch/csrc/lazy/backend/lowering_context.h>
 #include <torch/csrc/lazy/core/hash.h>
diff --git a/torch_xla/csrc/runtime/tensor_source.h b/torch_xla/csrc/runtime/tensor_source.h
index 20c9f674c9e..206215ee798 100644
--- a/torch_xla/csrc/runtime/tensor_source.h
+++ b/torch_xla/csrc/runtime/tensor_source.h
@@ -1,12 +1,11 @@
 #ifndef XLA_CLIENT_TENSOR_SOURCE_H_
 #define XLA_CLIENT_TENSOR_SOURCE_H_
 
-#include <ATen/Tensor.h>
-
 #include <string>
 #include <utility>
 #include <vector>
 
+#include <ATen/Tensor.h>
 #include <torch/csrc/lazy/core/metrics.h>
 
 #include "xla/literal.h"
diff --git a/torch_xla/csrc/runtime/tf_logging.cpp b/torch_xla/csrc/runtime/tf_logging.cpp
index a8c2d7644aa..2d2d84d5a68 100644
--- a/torch_xla/csrc/runtime/tf_logging.cpp
+++ b/torch_xla/csrc/runtime/tf_logging.cpp
@@ -1,9 +1,8 @@
 #include "torch_xla/csrc/runtime/tf_logging.h"
 
-#include <c10/util/Exception.h>
-
 #include <stdexcept>
 
+#include <c10/util/Exception.h>
 #include <torch/csrc/utils/cpp_stacktraces.h>
 
 #include "tsl/platform/stacktrace.h"
diff --git a/torch_xla/csrc/status.cpp b/torch_xla/csrc/status.cpp
index e9145a79aea..2666d2c8bbd 100644
--- a/torch_xla/csrc/status.cpp
+++ b/torch_xla/csrc/status.cpp
@@ -1,11 +1,10 @@
 #include "torch_xla/csrc/status.h"
 
-#include <c10/util/Exception.h>
-
 #include <iostream>
 #include <sstream>
 #include <stdexcept>
 
+#include <c10/util/Exception.h>
 #include <torch/csrc/utils/cpp_stacktraces.h>
 
 #include "absl/log/absl_check.h"
diff --git a/torch_xla/csrc/tensor.h b/torch_xla/csrc/tensor.h
index 66a8e14e913..e26bb266cf8 100644
--- a/torch_xla/csrc/tensor.h
+++ b/torch_xla/csrc/tensor.h
@@ -1,11 +1,10 @@
 #ifndef XLA_TORCH_XLA_CSRC_TENSOR_H_
 #define XLA_TORCH_XLA_CSRC_TENSOR_H_
 
-#include <c10/core/SymNodeImpl.h>
-
 #include <memory>
 #include <string>
 
+#include <c10/core/SymNodeImpl.h>
 #include <torch/csrc/autograd/variable.h>
 #include <torch/csrc/lazy/core/ir_metadata.h>
 #include <torch/csrc/lazy/core/ir_util.h>
diff --git a/torch_xla/csrc/tensor_impl.cpp b/torch_xla/csrc/tensor_impl.cpp
index 2c79eff110f..5293d5de6d0 100644
--- a/torch_xla/csrc/tensor_impl.cpp
+++ b/torch_xla/csrc/tensor_impl.cpp
@@ -4,7 +4,6 @@
 #include <c10/core/impl/DeviceGuardImplInterface.h>
 #include <c10/core/impl/LocalDispatchKeySet.h>
 #include <c10/macros/Macros.h>
-
 #include <torch/csrc/lazy/backend/backend_interface.h>
 #include <torch/csrc/lazy/core/tensor.h>
 #include <torch/csrc/lazy/core/tensor_util.h>
diff --git a/torch_xla/csrc/tensor_methods.cpp b/torch_xla/csrc/tensor_methods.cpp
index 65d4e1b4ac8..7b448591955 100644
--- a/torch_xla/csrc/tensor_methods.cpp
+++ b/torch_xla/csrc/tensor_methods.cpp
@@ -1,12 +1,11 @@
 #include "torch_xla/csrc/tensor_methods.h"
 
-#include <ATen/OpMathType.h>
-#include <ATen/core/Reduction.h>
-
 #include <algorithm>
 #include <functional>
 #include <iterator>
 
+#include <ATen/OpMathType.h>
+#include <ATen/core/Reduction.h>
 #include <torch/csrc/autograd/variable.h>
 #include <torch/csrc/lazy/core/helpers.h>
 #include <torch/csrc/lazy/core/util.h>
diff --git a/torch_xla/csrc/tensor_util.cpp b/torch_xla/csrc/tensor_util.cpp
index cef072f4c53..078e8a593c5 100644
--- a/torch_xla/csrc/tensor_util.cpp
+++ b/torch_xla/csrc/tensor_util.cpp
@@ -1,8 +1,5 @@
 #include "torch_xla/csrc/tensor_util.h"
 
-#include <ATen/Formatting.h>
-#include <ATen/Functions.h>
-
 #include <algorithm>
 #include <cstring>
 #include <functional>
@@ -10,6 +7,8 @@
 #include <numeric>
 #include <thread>
 
+#include <ATen/Formatting.h>
+#include <ATen/Functions.h>
 #include <torch/csrc/lazy/core/hash.h>
 #include <torch/csrc/lazy/core/util.h>
 
diff --git a/torch_xla/csrc/tensor_util.h b/torch_xla/csrc/tensor_util.h
index 4a6f73bd3bd..fc840054bcc 100644
--- a/torch_xla/csrc/tensor_util.h
+++ b/torch_xla/csrc/tensor_util.h
@@ -1,11 +1,10 @@
 #ifndef XLA_TORCH_XLA_CSRC_TENSOR_UTIL_H_
 #define XLA_TORCH_XLA_CSRC_TENSOR_UTIL_H_
 
-#include <ATen/ExpandUtils.h>
-
 #include <string>
 #include <vector>
 
+#include <ATen/ExpandUtils.h>
 #include <torch/csrc/autograd/variable.h>
 #include <torch/csrc/lazy/core/hash.h>
 
diff --git a/torch_xla/csrc/torch_util.h b/torch_xla/csrc/torch_util.h
index b45695ad6d6..1422c38fd4a 100644
--- a/torch_xla/csrc/torch_util.h
+++ b/torch_xla/csrc/torch_util.h
@@ -1,12 +1,11 @@
 #ifndef XLA_TORCH_XLA_CSRC_TORCH_UTIL_H_
 #define XLA_TORCH_XLA_CSRC_TORCH_UTIL_H_
 
+#include <optional>
+
 #include <ATen/ATen.h>
 #include <ATen/FunctionalTensorWrapper.h>
 #include <c10/core/ScalarType.h>
-
-#include <optional>
-
 #include <torch/csrc/lazy/core/dynamic_ir.h>
 #include <torch/csrc/lazy/core/hash.h>
 #include <torch/csrc/lazy/core/tensor.h>
diff --git a/torch_xla/csrc/xla_generator.cpp b/torch_xla/csrc/xla_generator.cpp
index cb334b1ba95..a04625cce6d 100644
--- a/torch_xla/csrc/xla_generator.cpp
+++ b/torch_xla/csrc/xla_generator.cpp
@@ -1,5 +1,9 @@
 #include "xla_generator.h"
 
+#include <cstring>
+#include <deque>
+#include <vector>
+
 #include <ATen/Functions.h>
 #include <ATen/core/ScalarType.h>
 #include <ATen/core/Tensor.h>
@@ -10,10 +14,6 @@
 #include <c10/util/CallOnce.h>
 #include <c10/util/intrusive_ptr.h>
 
-#include <cstring>
-#include <deque>
-#include <vector>
-
 #include "absl/status/status.h"
 
 #include "torch_xla/csrc/aten_xla_bridge.h"
diff --git a/torch_xla/csrc/xla_generator.h b/torch_xla/csrc/xla_generator.h
index 8001737e795..868ed7bee98 100644
--- a/torch_xla/csrc/xla_generator.h
+++ b/torch_xla/csrc/xla_generator.h
@@ -1,5 +1,7 @@
 #pragma once
 
+#include <cstdint>
+
 #include <ATen/core/Generator.h>
 #include <ATen/core/Tensor.h>
 #include <c10/core/Device.h>
@@ -8,8 +10,6 @@
 #include <c10/core/TensorImpl.h>
 #include <c10/util/intrusive_ptr.h>
 
-#include <cstdint>
-
 #include "absl/status/status.h"
 #include "absl/status/statusor.h"
 
diff --git a/torch_xla/csrc/xla_graph_executor.h b/torch_xla/csrc/xla_graph_executor.h
index e4dcedce826..96219551394 100644
--- a/torch_xla/csrc/xla_graph_executor.h
+++ b/torch_xla/csrc/xla_graph_executor.h
@@ -1,13 +1,12 @@
 #ifndef XLA_TORCH_XLA_CSRC_XLA_GRAPH_EXECUTOR_H_
 #define XLA_TORCH_XLA_CSRC_XLA_GRAPH_EXECUTOR_H_
 
-#include <c10/core/SymNodeImpl.h>
-
 #include <iostream>
 #include <memory>
 #include <string>
 #include <unordered_map>
 
+#include <c10/core/SymNodeImpl.h>
 #include <torch/csrc/autograd/variable.h>
 #include <torch/csrc/lazy/core/ir_util.h>
 
diff --git a/torch_xla/csrc/xla_manual_registration.cpp b/torch_xla/csrc/xla_manual_registration.cpp
index 6aad7994040..f5e5f74f380 100644
--- a/torch_xla/csrc/xla_manual_registration.cpp
+++ b/torch_xla/csrc/xla_manual_registration.cpp
@@ -1,6 +1,5 @@
 #include <ATen/ATen.h>
 #include <ATen/native/CPUFallback.h>
-
 #include <torch/library.h>
 
 #include "torch_xla/csrc/XLANativeFunctions.h"
diff --git a/torch_xla/csrc/xla_sharding_util.cpp b/torch_xla/csrc/xla_sharding_util.cpp
index 2f90a6a0bca..6ecf6c65bf1 100644
--- a/torch_xla/csrc/xla_sharding_util.cpp
+++ b/torch_xla/csrc/xla_sharding_util.cpp
@@ -1,10 +1,9 @@
 #include "torch_xla/csrc/xla_sharding_util.h"
 
-#include <ATen/TensorIndexing.h>
-
 #include <cmath>
 #include <unordered_map>
 
+#include <ATen/TensorIndexing.h>
 #include <torch/csrc/lazy/core/ir_util.h>
 
 #include "absl/synchronization/blocking_counter.h"