Add files using upload-large-folder tool

lib/python3.12/site-packages/torch/include/ATen/ops/_adaptive_avg_pool3d_backward_native.h

@@ -0,0 +1,23 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API at::Tensor & _adaptive_avg_pool3d_backward_out(const at::Tensor & grad_output, const at::Tensor & self, at::Tensor & out);
+TORCH_API at::Tensor adaptive_avg_pool3d_backward_cpu(const at::Tensor & grad_output, const at::Tensor & self);
+TORCH_API at::Tensor adaptive_avg_pool3d_backward_cuda(const at::Tensor & grad_output, const at::Tensor & self);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_add_relu_ops.h

@@ -0,0 +1,83 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _add_relu_Tensor {
+  using schema = at::Tensor (const at::Tensor &, const at::Tensor &, const at::Scalar &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_add_relu";
+  static constexpr const char* overload_name = "Tensor";
+  static constexpr const char* schema_str = "_add_relu.Tensor(Tensor self, Tensor other, *, Scalar alpha=1) -> Tensor";
+  static at::Tensor call(const at::Tensor & self, const at::Tensor & other, const at::Scalar & alpha);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Tensor & other, const at::Scalar & alpha);
+};
+
+struct TORCH_API _add_relu__Tensor {
+  using schema = at::Tensor & (at::Tensor &, const at::Tensor &, const at::Scalar &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_add_relu_";
+  static constexpr const char* overload_name = "Tensor";
+  static constexpr const char* schema_str = "_add_relu_.Tensor(Tensor(a!) self, Tensor other, *, Scalar alpha=1) -> Tensor(a!)";
+  static at::Tensor & call(at::Tensor & self, const at::Tensor & other, const at::Scalar & alpha);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, at::Tensor & self, const at::Tensor & other, const at::Scalar & alpha);
+};
+
+struct TORCH_API _add_relu_out {
+  using schema = at::Tensor & (const at::Tensor &, const at::Tensor &, const at::Scalar &, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_add_relu";
+  static constexpr const char* overload_name = "out";
+  static constexpr const char* schema_str = "_add_relu.out(Tensor self, Tensor other, *, Scalar alpha=1, Tensor(a!) out) -> Tensor(a!)";
+  static at::Tensor & call(const at::Tensor & self, const at::Tensor & other, const at::Scalar & alpha, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Tensor & other, const at::Scalar & alpha, at::Tensor & out);
+};
+
+struct TORCH_API _add_relu_Scalar {
+  using schema = at::Tensor (const at::Tensor &, const at::Scalar &, const at::Scalar &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_add_relu";
+  static constexpr const char* overload_name = "Scalar";
+  static constexpr const char* schema_str = "_add_relu.Scalar(Tensor self, Scalar other, Scalar alpha=1) -> Tensor";
+  static at::Tensor call(const at::Tensor & self, const at::Scalar & other, const at::Scalar & alpha);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Scalar & other, const at::Scalar & alpha);
+};
+
+struct TORCH_API _add_relu__Scalar {
+  using schema = at::Tensor & (at::Tensor &, const at::Scalar &, const at::Scalar &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_add_relu_";
+  static constexpr const char* overload_name = "Scalar";
+  static constexpr const char* schema_str = "_add_relu_.Scalar(Tensor(a!) self, Scalar other, Scalar alpha=1) -> Tensor(a!)";
+  static at::Tensor & call(at::Tensor & self, const at::Scalar & other, const at::Scalar & alpha);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, at::Tensor & self, const at::Scalar & other, const at::Scalar & alpha);
+};
+
+struct TORCH_API _add_relu_Scalar_out {
+  using schema = at::Tensor & (const at::Tensor &, const at::Scalar &, const at::Scalar &, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_add_relu";
+  static constexpr const char* overload_name = "Scalar_out";
+  static constexpr const char* schema_str = "_add_relu.Scalar_out(Tensor self, Scalar other, Scalar alpha=1, *, Tensor(a!) out) -> Tensor(a!)";
+  static at::Tensor & call(const at::Tensor & self, const at::Scalar & other, const at::Scalar & alpha, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Scalar & other, const at::Scalar & alpha, at::Tensor & out);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_amp_foreach_non_finite_check_and_unscale_compositeexplicitautograd_dispatch.h

@@ -0,0 +1,25 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautograd {
+
+TORCH_API ::std::tuple<::std::vector<at::Tensor>,at::Tensor> _amp_foreach_non_finite_check_and_unscale(at::TensorList self, const at::Tensor & found_inf, const at::Tensor & inv_scale);
+TORCH_API void _amp_foreach_non_finite_check_and_unscale_out(at::TensorList out, at::TensorList self, at::Tensor & found_inf, const at::Tensor & inv_scale);
+TORCH_API void _amp_foreach_non_finite_check_and_unscale_outf(at::TensorList self, at::Tensor & found_inf, const at::Tensor & inv_scale, at::TensorList out);
+
+} // namespace compositeexplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_assert_tensor_metadata.h

@@ -0,0 +1,47 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_assert_tensor_metadata_ops.h>
+
+namespace at {
+
+
+// aten::_assert_tensor_metadata(Tensor a, SymInt[]? size=None, SymInt[]? stride=None, ScalarType? dtype=None, *, Device? device=None, Layout? layout=None) -> ()
+inline void _assert_tensor_metadata(const at::Tensor & a, at::OptionalIntArrayRef size=::std::nullopt, at::OptionalIntArrayRef stride=::std::nullopt, ::std::optional<at::ScalarType> dtype=::std::nullopt, ::std::optional<at::Device> device=::std::nullopt, ::std::optional<at::Layout> layout=::std::nullopt) {
+    return at::_ops::_assert_tensor_metadata::call(a, size.has_value() ? ::std::make_optional(c10::fromIntArrayRefSlow(*size)) : ::std::nullopt, stride.has_value() ? ::std::make_optional(c10::fromIntArrayRefSlow(*stride)) : ::std::nullopt, dtype, device, layout);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same_v<T, int64_t>>>
+  void _assert_tensor_metadata(const at::Tensor & a, at::OptionalIntArrayRef size=::std::nullopt, at::OptionalIntArrayRef stride=::std::nullopt, ::std::optional<at::ScalarType> dtype=::std::nullopt, ::std::optional<at::Device> device=::std::nullopt, ::std::optional<at::Layout> layout=::std::nullopt) {
+    return at::_ops::_assert_tensor_metadata::call(a, size.has_value() ? ::std::make_optional(c10::fromIntArrayRefSlow(*size)) : ::std::nullopt, stride.has_value() ? ::std::make_optional(c10::fromIntArrayRefSlow(*stride)) : ::std::nullopt, dtype, device, layout);
+  }
+}
+
+// aten::_assert_tensor_metadata(Tensor a, SymInt[]? size=None, SymInt[]? stride=None, ScalarType? dtype=None, *, Device? device=None, Layout? layout=None) -> ()
+inline void _assert_tensor_metadata_symint(const at::Tensor & a, at::OptionalSymIntArrayRef size=::std::nullopt, at::OptionalSymIntArrayRef stride=::std::nullopt, ::std::optional<at::ScalarType> dtype=::std::nullopt, ::std::optional<at::Device> device=::std::nullopt, ::std::optional<at::Layout> layout=::std::nullopt) {
+    return at::_ops::_assert_tensor_metadata::call(a, size, stride, dtype, device, layout);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same_v<T, c10::SymInt>>>
+  void _assert_tensor_metadata(const at::Tensor & a, at::OptionalSymIntArrayRef size=::std::nullopt, at::OptionalSymIntArrayRef stride=::std::nullopt, ::std::optional<at::ScalarType> dtype=::std::nullopt, ::std::optional<at::Device> device=::std::nullopt, ::std::optional<at::Layout> layout=::std::nullopt) {
+    return at::_ops::_assert_tensor_metadata::call(a, size, stride, dtype, device, layout);
+  }
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_assert_tensor_metadata_meta_dispatch.h

@@ -0,0 +1,24 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace meta {
+
+TORCH_API void _assert_tensor_metadata(const at::Tensor & a, at::OptionalIntArrayRef size=::std::nullopt, at::OptionalIntArrayRef stride=::std::nullopt, ::std::optional<at::ScalarType> dtype=::std::nullopt, ::std::optional<at::Device> device=::std::nullopt, ::std::optional<at::Layout> layout=::std::nullopt);
+TORCH_API void _assert_tensor_metadata_symint(const at::Tensor & a, at::OptionalSymIntArrayRef size=::std::nullopt, at::OptionalSymIntArrayRef stride=::std::nullopt, ::std::optional<at::ScalarType> dtype=::std::nullopt, ::std::optional<at::Device> device=::std::nullopt, ::std::optional<at::Layout> layout=::std::nullopt);
+
+} // namespace meta
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_batch_norm_impl_index.h

@@ -0,0 +1,30 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_batch_norm_impl_index_ops.h>
+
+namespace at {
+
+
+// aten::_batch_norm_impl_index(Tensor input, Tensor? weight, Tensor? bias, Tensor? running_mean, Tensor? running_var, bool training, float momentum, float eps, bool cudnn_enabled) -> (Tensor, Tensor, Tensor, Tensor, int)
+inline ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,int64_t> _batch_norm_impl_index(const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & bias, const ::std::optional<at::Tensor> & running_mean, const ::std::optional<at::Tensor> & running_var, bool training, double momentum, double eps, bool cudnn_enabled) {
+    return at::_ops::_batch_norm_impl_index::call(input, weight, bias, running_mean, running_var, training, momentum, eps, cudnn_enabled);
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_batch_norm_impl_index_backward.h

@@ -0,0 +1,30 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_batch_norm_impl_index_backward_ops.h>
+
+namespace at {
+
+
+// aten::_batch_norm_impl_index_backward(int impl_index, Tensor input, Tensor grad_output, Tensor? weight, Tensor? running_mean, Tensor? running_var, Tensor? save_mean, Tensor? save_var_transform, bool train, float eps, bool[3] output_mask, Tensor reservedSpace) -> (Tensor, Tensor, Tensor)
+inline ::std::tuple<at::Tensor,at::Tensor,at::Tensor> _batch_norm_impl_index_backward(int64_t impl_index, const at::Tensor & input, const at::Tensor & grad_output, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & running_mean, const ::std::optional<at::Tensor> & running_var, const ::std::optional<at::Tensor> & save_mean, const ::std::optional<at::Tensor> & save_var_transform, bool train, double eps, ::std::array<bool,3> output_mask, const at::Tensor & reservedSpace) {
+    return at::_ops::_batch_norm_impl_index_backward::call(impl_index, input, grad_output, weight, running_mean, running_var, save_mean, save_var_transform, train, eps, output_mask, reservedSpace);
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_cast_Char_compositeimplicitautograd_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeimplicitautograd {
+
+TORCH_API at::Tensor _cast_Char(const at::Tensor & self, bool non_blocking=false);
+
+} // namespace compositeimplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_cast_Short.h

@@ -0,0 +1,30 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_cast_Short_ops.h>
+
+namespace at {
+
+
+// aten::_cast_Short(Tensor self, bool non_blocking=False) -> Tensor
+inline at::Tensor _cast_Short(const at::Tensor & self, bool non_blocking=false) {
+    return at::_ops::_cast_Short::call(self, non_blocking);
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_cdist_backward_cuda_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API at::Tensor _cdist_backward(const at::Tensor & grad, const at::Tensor & x1, const at::Tensor & x2, double p, const at::Tensor & cdist);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_convert_indices_from_coo_to_csr_cuda_dispatch.h

@@ -0,0 +1,25 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API at::Tensor _convert_indices_from_coo_to_csr(const at::Tensor & self, int64_t size, bool out_int32=false);
+TORCH_API at::Tensor & _convert_indices_from_coo_to_csr_out(at::Tensor & out, const at::Tensor & self, int64_t size, bool out_int32=false);
+TORCH_API at::Tensor & _convert_indices_from_coo_to_csr_outf(const at::Tensor & self, int64_t size, bool out_int32, at::Tensor & out);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_convert_indices_from_coo_to_csr_native.h

@@ -0,0 +1,26 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+#include <ATen/ops/_convert_indices_from_coo_to_csr_meta.h>
+
+namespace at {
+namespace native {
+struct TORCH_API structured__convert_indices_from_coo_to_csr_structured_cpu : public at::meta::structured__convert_indices_from_coo_to_csr {
+void impl(const at::Tensor & self, int64_t size, bool out_int32, const at::Tensor & out);
+};
+struct TORCH_API structured__convert_indices_from_coo_to_csr_structured_cuda : public at::meta::structured__convert_indices_from_coo_to_csr {
+void impl(const at::Tensor & self, int64_t size, bool out_int32, const at::Tensor & out);
+};
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_convert_indices_from_coo_to_csr_ops.h

@@ -0,0 +1,39 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _convert_indices_from_coo_to_csr {
+  using schema = at::Tensor (const at::Tensor &, int64_t, bool);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_convert_indices_from_coo_to_csr";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_convert_indices_from_coo_to_csr(Tensor self, int size, *, bool out_int32=False) -> Tensor";
+  static at::Tensor call(const at::Tensor & self, int64_t size, bool out_int32);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, int64_t size, bool out_int32);
+};
+
+struct TORCH_API _convert_indices_from_coo_to_csr_out {
+  using schema = at::Tensor & (const at::Tensor &, int64_t, bool, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_convert_indices_from_coo_to_csr";
+  static constexpr const char* overload_name = "out";
+  static constexpr const char* schema_str = "_convert_indices_from_coo_to_csr.out(Tensor self, int size, *, bool out_int32=False, Tensor(a!) out) -> Tensor(a!)";
+  static at::Tensor & call(const at::Tensor & self, int64_t size, bool out_int32, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, int64_t size, bool out_int32, at::Tensor & out);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_convert_indices_from_csr_to_coo_cuda_dispatch.h

@@ -0,0 +1,25 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API at::Tensor _convert_indices_from_csr_to_coo(const at::Tensor & crow_indices, const at::Tensor & col_indices, bool out_int32=false, bool transpose=false);
+TORCH_API at::Tensor & _convert_indices_from_csr_to_coo_out(at::Tensor & out, const at::Tensor & crow_indices, const at::Tensor & col_indices, bool out_int32=false, bool transpose=false);
+TORCH_API at::Tensor & _convert_indices_from_csr_to_coo_outf(const at::Tensor & crow_indices, const at::Tensor & col_indices, bool out_int32, bool transpose, at::Tensor & out);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_cudnn_rnn_backward_native.h

@@ -0,0 +1,22 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API void _cudnn_rnn_backward_out_symint(const at::Tensor & input, at::TensorList weight, int64_t weight_stride0, const at::Tensor & weight_buf, const at::Tensor & hx, const ::std::optional<at::Tensor> & cx, const at::Tensor & output, const ::std::optional<at::Tensor> & grad_output, const ::std::optional<at::Tensor> & grad_hy, const ::std::optional<at::Tensor> & grad_cy, int64_t mode, c10::SymInt hidden_size, c10::SymInt proj_size, int64_t num_layers, bool batch_first, double dropout, bool train, bool bidirectional, c10::SymIntArrayRef batch_sizes, const ::std::optional<at::Tensor> & dropout_state, const at::Tensor & reserve, ::std::array<bool,4> output_mask, at::Tensor & out0, at::Tensor & out1, at::Tensor & out2, at::TensorList out3);
+TORCH_API ::std::tuple<at::Tensor,at::Tensor,at::Tensor,::std::vector<at::Tensor>> _cudnn_rnn_backward(const at::Tensor & input, at::TensorList weight, int64_t weight_stride0, const at::Tensor & weight_buf, const at::Tensor & hx, const ::std::optional<at::Tensor> & cx, const at::Tensor & output, const ::std::optional<at::Tensor> & grad_output, const ::std::optional<at::Tensor> & grad_hy, const ::std::optional<at::Tensor> & grad_cy, int64_t mode, int64_t hidden_size, int64_t proj_size, int64_t num_layers, bool batch_first, double dropout, bool train, bool bidirectional, at::IntArrayRef batch_sizes, const ::std::optional<at::Tensor> & dropout_state, const at::Tensor & reserve, ::std::array<bool,4> output_mask);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_cudnn_rnn_compositeexplicitautograd_dispatch.h

@@ -0,0 +1,26 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautograd {
+
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> _cudnn_rnn_out(at::Tensor & out0, at::Tensor & out1, at::Tensor & out2, at::Tensor & out3, at::Tensor & out4, const at::Tensor & input, at::TensorList weight, int64_t weight_stride0, const ::std::optional<at::Tensor> & weight_buf, const at::Tensor & hx, const ::std::optional<at::Tensor> & cx, int64_t mode, int64_t hidden_size, int64_t proj_size, int64_t num_layers, bool batch_first, double dropout, bool train, bool bidirectional, at::IntArrayRef batch_sizes, const ::std::optional<at::Tensor> & dropout_state);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> _cudnn_rnn_outf(const at::Tensor & input, at::TensorList weight, int64_t weight_stride0, const ::std::optional<at::Tensor> & weight_buf, const at::Tensor & hx, const ::std::optional<at::Tensor> & cx, int64_t mode, int64_t hidden_size, int64_t proj_size, int64_t num_layers, bool batch_first, double dropout, bool train, bool bidirectional, at::IntArrayRef batch_sizes, const ::std::optional<at::Tensor> & dropout_state, at::Tensor & out0, at::Tensor & out1, at::Tensor & out2, at::Tensor & out3, at::Tensor & out4);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> _cudnn_rnn_symint_out(at::Tensor & out0, at::Tensor & out1, at::Tensor & out2, at::Tensor & out3, at::Tensor & out4, const at::Tensor & input, at::TensorList weight, int64_t weight_stride0, const ::std::optional<at::Tensor> & weight_buf, const at::Tensor & hx, const ::std::optional<at::Tensor> & cx, int64_t mode, c10::SymInt hidden_size, c10::SymInt proj_size, int64_t num_layers, bool batch_first, double dropout, bool train, bool bidirectional, c10::SymIntArrayRef batch_sizes, const ::std::optional<at::Tensor> & dropout_state);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> _cudnn_rnn_symint_outf(const at::Tensor & input, at::TensorList weight, int64_t weight_stride0, const ::std::optional<at::Tensor> & weight_buf, const at::Tensor & hx, const ::std::optional<at::Tensor> & cx, int64_t mode, c10::SymInt hidden_size, c10::SymInt proj_size, int64_t num_layers, bool batch_first, double dropout, bool train, bool bidirectional, c10::SymIntArrayRef batch_sizes, const ::std::optional<at::Tensor> & dropout_state, at::Tensor & out0, at::Tensor & out1, at::Tensor & out2, at::Tensor & out3, at::Tensor & out4);
+
+} // namespace compositeexplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_cufft_get_plan_cache_max_size_ops.h

@@ -0,0 +1,28 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _cufft_get_plan_cache_max_size {
+  using schema = int64_t (at::DeviceIndex);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_cufft_get_plan_cache_max_size";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_cufft_get_plan_cache_max_size(DeviceIndex device_index) -> int";
+  static int64_t call(at::DeviceIndex device_index);
+  static int64_t redispatch(c10::DispatchKeySet dispatchKeySet, at::DeviceIndex device_index);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_cufft_set_plan_cache_max_size.h

@@ -0,0 +1,30 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_cufft_set_plan_cache_max_size_ops.h>
+
+namespace at {
+
+
+// aten::_cufft_set_plan_cache_max_size(DeviceIndex device_index, int max_size) -> ()
+inline void _cufft_set_plan_cache_max_size(at::DeviceIndex device_index, int64_t max_size) {
+    return at::_ops::_cufft_set_plan_cache_max_size::call(device_index, max_size);
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_cufft_set_plan_cache_max_size_ops.h

@@ -0,0 +1,28 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _cufft_set_plan_cache_max_size {
+  using schema = void (at::DeviceIndex, int64_t);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_cufft_set_plan_cache_max_size";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_cufft_set_plan_cache_max_size(DeviceIndex device_index, int max_size) -> ()";
+  static void call(at::DeviceIndex device_index, int64_t max_size);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::DeviceIndex device_index, int64_t max_size);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_cummax_helper.h

@@ -0,0 +1,30 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_cummax_helper_ops.h>
+
+namespace at {
+
+
+// aten::_cummax_helper(Tensor self, Tensor(a!) values, Tensor(b!) indices, int dim) -> ()
+inline void _cummax_helper(const at::Tensor & self, at::Tensor & values, at::Tensor & indices, int64_t dim) {
+    return at::_ops::_cummax_helper::call(self, values, indices, dim);
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_efficient_attention_forward.h

@@ -0,0 +1,47 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_efficient_attention_forward_ops.h>
+
+namespace at {
+
+
+// aten::_efficient_attention_forward(Tensor query, Tensor key, Tensor value, Tensor? bias, Tensor? cu_seqlens_q, Tensor? cu_seqlens_k, SymInt? max_seqlen_q, SymInt? max_seqlen_k, float dropout_p, int custom_mask_type, bool compute_log_sumexp=False, *, float? scale=None, Tensor? seqlen_k=None, int? window_size=None) -> (Tensor output, Tensor logsumexp, Tensor philox_seed, Tensor philox_offset, SymInt max_seqlen_batch_q, SymInt max_seqlen_batch_k)
+inline ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,c10::SymInt,c10::SymInt> _efficient_attention_forward(const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const ::std::optional<at::Tensor> & bias, const ::std::optional<at::Tensor> & cu_seqlens_q, const ::std::optional<at::Tensor> & cu_seqlens_k, ::std::optional<int64_t> max_seqlen_q, ::std::optional<int64_t> max_seqlen_k, double dropout_p, int64_t custom_mask_type, bool compute_log_sumexp=false, ::std::optional<double> scale=::std::nullopt, const ::std::optional<at::Tensor> & seqlen_k={}, ::std::optional<int64_t> window_size=::std::nullopt) {
+    return at::_ops::_efficient_attention_forward::call(query, key, value, bias, cu_seqlens_q, cu_seqlens_k, max_seqlen_q.has_value() ? ::std::make_optional(c10::SymInt(*max_seqlen_q)) : ::std::nullopt, max_seqlen_k.has_value() ? ::std::make_optional(c10::SymInt(*max_seqlen_k)) : ::std::nullopt, dropout_p, custom_mask_type, compute_log_sumexp, scale, seqlen_k, window_size);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same_v<T, int64_t>>>
+  ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,c10::SymInt,c10::SymInt> _efficient_attention_forward(const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const ::std::optional<at::Tensor> & bias, const ::std::optional<at::Tensor> & cu_seqlens_q, const ::std::optional<at::Tensor> & cu_seqlens_k, ::std::optional<int64_t> max_seqlen_q, ::std::optional<int64_t> max_seqlen_k, double dropout_p, int64_t custom_mask_type, bool compute_log_sumexp=false, ::std::optional<double> scale=::std::nullopt, const ::std::optional<at::Tensor> & seqlen_k={}, ::std::optional<int64_t> window_size=::std::nullopt) {
+    return at::_ops::_efficient_attention_forward::call(query, key, value, bias, cu_seqlens_q, cu_seqlens_k, max_seqlen_q.has_value() ? ::std::make_optional(c10::SymInt(*max_seqlen_q)) : ::std::nullopt, max_seqlen_k.has_value() ? ::std::make_optional(c10::SymInt(*max_seqlen_k)) : ::std::nullopt, dropout_p, custom_mask_type, compute_log_sumexp, scale, seqlen_k, window_size);
+  }
+}
+
+// aten::_efficient_attention_forward(Tensor query, Tensor key, Tensor value, Tensor? bias, Tensor? cu_seqlens_q, Tensor? cu_seqlens_k, SymInt? max_seqlen_q, SymInt? max_seqlen_k, float dropout_p, int custom_mask_type, bool compute_log_sumexp=False, *, float? scale=None, Tensor? seqlen_k=None, int? window_size=None) -> (Tensor output, Tensor logsumexp, Tensor philox_seed, Tensor philox_offset, SymInt max_seqlen_batch_q, SymInt max_seqlen_batch_k)
+inline ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,c10::SymInt,c10::SymInt> _efficient_attention_forward_symint(const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const ::std::optional<at::Tensor> & bias, const ::std::optional<at::Tensor> & cu_seqlens_q, const ::std::optional<at::Tensor> & cu_seqlens_k, ::std::optional<c10::SymInt> max_seqlen_q, ::std::optional<c10::SymInt> max_seqlen_k, double dropout_p, int64_t custom_mask_type, bool compute_log_sumexp=false, ::std::optional<double> scale=::std::nullopt, const ::std::optional<at::Tensor> & seqlen_k={}, ::std::optional<int64_t> window_size=::std::nullopt) {
+    return at::_ops::_efficient_attention_forward::call(query, key, value, bias, cu_seqlens_q, cu_seqlens_k, max_seqlen_q, max_seqlen_k, dropout_p, custom_mask_type, compute_log_sumexp, scale, seqlen_k, window_size);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same_v<T, c10::SymInt>>>
+  ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,c10::SymInt,c10::SymInt> _efficient_attention_forward(const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const ::std::optional<at::Tensor> & bias, const ::std::optional<at::Tensor> & cu_seqlens_q, const ::std::optional<at::Tensor> & cu_seqlens_k, ::std::optional<c10::SymInt> max_seqlen_q, ::std::optional<c10::SymInt> max_seqlen_k, double dropout_p, int64_t custom_mask_type, bool compute_log_sumexp=false, ::std::optional<double> scale=::std::nullopt, const ::std::optional<at::Tensor> & seqlen_k={}, ::std::optional<int64_t> window_size=::std::nullopt) {
+    return at::_ops::_efficient_attention_forward::call(query, key, value, bias, cu_seqlens_q, cu_seqlens_k, max_seqlen_q, max_seqlen_k, dropout_p, custom_mask_type, compute_log_sumexp, scale, seqlen_k, window_size);
+  }
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_efficient_attention_forward_cuda_dispatch.h

@@ -0,0 +1,24 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,c10::SymInt,c10::SymInt> _efficient_attention_forward(const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const ::std::optional<at::Tensor> & bias, const ::std::optional<at::Tensor> & cu_seqlens_q, const ::std::optional<at::Tensor> & cu_seqlens_k, ::std::optional<int64_t> max_seqlen_q, ::std::optional<int64_t> max_seqlen_k, double dropout_p, int64_t custom_mask_type, bool compute_log_sumexp=false, ::std::optional<double> scale=::std::nullopt, const ::std::optional<at::Tensor> & seqlen_k={}, ::std::optional<int64_t> window_size=::std::nullopt);
+TORCH_API ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,c10::SymInt,c10::SymInt> _efficient_attention_forward_symint(const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const ::std::optional<at::Tensor> & bias, const ::std::optional<at::Tensor> & cu_seqlens_q, const ::std::optional<at::Tensor> & cu_seqlens_k, ::std::optional<c10::SymInt> max_seqlen_q, ::std::optional<c10::SymInt> max_seqlen_k, double dropout_p, int64_t custom_mask_type, bool compute_log_sumexp=false, ::std::optional<double> scale=::std::nullopt, const ::std::optional<at::Tensor> & seqlen_k={}, ::std::optional<int64_t> window_size=::std::nullopt);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_empty_per_channel_affine_quantized_native.h

@@ -0,0 +1,23 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API at::Tensor & _empty_per_channel_affine_quantized_out_symint(c10::SymIntArrayRef size, const at::Tensor & scales, const at::Tensor & zero_points, int64_t axis, ::std::optional<at::MemoryFormat> memory_format, at::Tensor & out);
+TORCH_API at::Tensor empty_per_channel_affine_quantized_other_backends_stub(at::IntArrayRef size, const at::Tensor & scales, const at::Tensor & zero_points, int64_t axis, ::std::optional<at::ScalarType> dtype={}, ::std::optional<at::Layout> layout={}, ::std::optional<at::Device> device={}, ::std::optional<bool> pin_memory={}, ::std::optional<at::MemoryFormat> memory_format=c10::MemoryFormat::Contiguous);
+TORCH_API at::Tensor empty_per_channel_affine_quantized(at::IntArrayRef size, const at::Tensor & scales, const at::Tensor & zero_points, int64_t axis, ::std::optional<at::ScalarType> dtype={}, ::std::optional<at::Layout> layout={}, ::std::optional<at::Device> device={}, ::std::optional<bool> pin_memory={}, ::std::optional<at::MemoryFormat> memory_format=c10::MemoryFormat::Contiguous);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_euclidean_dist_native.h

@@ -0,0 +1,22 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API at::Tensor _euclidean_dist(const at::Tensor & x1, const at::Tensor & x2);
+TORCH_API at::Tensor & _euclidean_dist_out(const at::Tensor & x1, const at::Tensor & x2, at::Tensor & out);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_fake_quantize_per_tensor_affine_cachemask_tensor_qparams_ops.h

@@ -0,0 +1,39 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _fake_quantize_per_tensor_affine_cachemask_tensor_qparams {
+  using schema = ::std::tuple<at::Tensor,at::Tensor> (const at::Tensor &, const at::Tensor &, const at::Tensor &, const at::Tensor &, int64_t, int64_t);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_fake_quantize_per_tensor_affine_cachemask_tensor_qparams";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_fake_quantize_per_tensor_affine_cachemask_tensor_qparams(Tensor self, Tensor scale, Tensor zero_point, Tensor fake_quant_enabled, int quant_min, int quant_max) -> (Tensor output, Tensor mask)";
+  static ::std::tuple<at::Tensor,at::Tensor> call(const at::Tensor & self, const at::Tensor & scale, const at::Tensor & zero_point, const at::Tensor & fake_quant_enabled, int64_t quant_min, int64_t quant_max);
+  static ::std::tuple<at::Tensor,at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Tensor & scale, const at::Tensor & zero_point, const at::Tensor & fake_quant_enabled, int64_t quant_min, int64_t quant_max);
+};
+
+struct TORCH_API _fake_quantize_per_tensor_affine_cachemask_tensor_qparams_out {
+  using schema = ::std::tuple<at::Tensor &,at::Tensor &> (const at::Tensor &, const at::Tensor &, const at::Tensor &, const at::Tensor &, int64_t, int64_t, at::Tensor &, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_fake_quantize_per_tensor_affine_cachemask_tensor_qparams";
+  static constexpr const char* overload_name = "out";
+  static constexpr const char* schema_str = "_fake_quantize_per_tensor_affine_cachemask_tensor_qparams.out(Tensor self, Tensor scale, Tensor zero_point, Tensor fake_quant_enabled, int quant_min, int quant_max, *, Tensor(a!) out0, Tensor(b!) out1) -> (Tensor(a!), Tensor(b!))";
+  static ::std::tuple<at::Tensor &,at::Tensor &> call(const at::Tensor & self, const at::Tensor & scale, const at::Tensor & zero_point, const at::Tensor & fake_quant_enabled, int64_t quant_min, int64_t quant_max, at::Tensor & out0, at::Tensor & out1);
+  static ::std::tuple<at::Tensor &,at::Tensor &> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Tensor & scale, const at::Tensor & zero_point, const at::Tensor & fake_quant_enabled, int64_t quant_min, int64_t quant_max, at::Tensor & out0, at::Tensor & out1);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_acos_ops.h

@@ -0,0 +1,50 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _foreach_acos {
+  using schema = ::std::vector<at::Tensor> (at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_acos";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_foreach_acos(Tensor[] self) -> Tensor[]";
+  static ::std::vector<at::Tensor> call(at::TensorList self);
+  static ::std::vector<at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self);
+};
+
+struct TORCH_API _foreach_acos_ {
+  using schema = void (at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_acos_";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_foreach_acos_(Tensor(a!)[] self) -> ()";
+  static void call(at::TensorList self);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self);
+};
+
+struct TORCH_API _foreach_acos_out {
+  using schema = void (at::TensorList, at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_acos";
+  static constexpr const char* overload_name = "out";
+  static constexpr const char* schema_str = "_foreach_acos.out(Tensor[] self, *, Tensor(a!)[] out) -> ()";
+  static void call(at::TensorList self, at::TensorList out);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::TensorList out);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_asin_cuda_dispatch.h

@@ -0,0 +1,24 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_asin(at::TensorList self);
+TORCH_API void _foreach_asin_(at::TensorList self);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_cos_cuda_dispatch.h

@@ -0,0 +1,24 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_cos(at::TensorList self);
+TORCH_API void _foreach_cos_(at::TensorList self);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_erfc_native.h

@@ -0,0 +1,25 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_erfc_slow(at::TensorList self);
+TORCH_API void _foreach_erfc_out(at::TensorList self, at::TensorList out);
+TORCH_API void foreach_tensor_erfc_slow_(at::TensorList self);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_erfc_cuda(at::TensorList self);
+TORCH_API void foreach_tensor_erfc_cuda_(at::TensorList self);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_lerp.h

@@ -0,0 +1,82 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_foreach_lerp_ops.h>
+
+namespace at {
+
+
+// aten::_foreach_lerp.List(Tensor[] self, Tensor[] tensors1, Tensor[] weights) -> Tensor[]
+inline ::std::vector<at::Tensor> _foreach_lerp(at::TensorList self, at::TensorList tensors1, at::TensorList weights) {
+    return at::_ops::_foreach_lerp_List::call(self, tensors1, weights);
+}
+
+// aten::_foreach_lerp_.List(Tensor(a!)[] self, Tensor[] tensors1, Tensor[] weights) -> ()
+inline void _foreach_lerp_(at::TensorList self, at::TensorList tensors1, at::TensorList weights) {
+    return at::_ops::_foreach_lerp__List::call(self, tensors1, weights);
+}
+
+// aten::_foreach_lerp.Scalar(Tensor[] self, Tensor[] tensors1, Scalar weight) -> Tensor[]
+inline ::std::vector<at::Tensor> _foreach_lerp(at::TensorList self, at::TensorList tensors1, const at::Scalar & weight) {
+    return at::_ops::_foreach_lerp_Scalar::call(self, tensors1, weight);
+}
+
+// aten::_foreach_lerp_.Scalar(Tensor(a!)[] self, Tensor[] tensors1, Scalar weight) -> ()
+inline void _foreach_lerp_(at::TensorList self, at::TensorList tensors1, const at::Scalar & weight) {
+    return at::_ops::_foreach_lerp__Scalar::call(self, tensors1, weight);
+}
+
+// aten::_foreach_lerp.ScalarList(Tensor[] self, Tensor[] tensors1, Scalar[] weight) -> Tensor[]
+inline ::std::vector<at::Tensor> _foreach_lerp(at::TensorList self, at::TensorList tensors1, at::ArrayRef<at::Scalar> weight) {
+    return at::_ops::_foreach_lerp_ScalarList::call(self, tensors1, weight);
+}
+
+// aten::_foreach_lerp_.ScalarList(Tensor(a!)[] self, Tensor[] tensors1, Scalar[] weight) -> ()
+inline void _foreach_lerp_(at::TensorList self, at::TensorList tensors1, at::ArrayRef<at::Scalar> weight) {
+    return at::_ops::_foreach_lerp__ScalarList::call(self, tensors1, weight);
+}
+
+// aten::_foreach_lerp.List_out(Tensor[] self, Tensor[] tensors1, Tensor[] weights, *, Tensor(a!)[] out) -> ()
+inline void _foreach_lerp_out(at::TensorList out, at::TensorList self, at::TensorList tensors1, at::TensorList weights) {
+    return at::_ops::_foreach_lerp_List_out::call(self, tensors1, weights, out);
+}
+// aten::_foreach_lerp.List_out(Tensor[] self, Tensor[] tensors1, Tensor[] weights, *, Tensor(a!)[] out) -> ()
+inline void _foreach_lerp_outf(at::TensorList self, at::TensorList tensors1, at::TensorList weights, at::TensorList out) {
+    return at::_ops::_foreach_lerp_List_out::call(self, tensors1, weights, out);
+}
+
+// aten::_foreach_lerp.Scalar_out(Tensor[] self, Tensor[] tensors1, Scalar weight, *, Tensor(a!)[] out) -> ()
+inline void _foreach_lerp_out(at::TensorList out, at::TensorList self, at::TensorList tensors1, const at::Scalar & weight) {
+    return at::_ops::_foreach_lerp_Scalar_out::call(self, tensors1, weight, out);
+}
+// aten::_foreach_lerp.Scalar_out(Tensor[] self, Tensor[] tensors1, Scalar weight, *, Tensor(a!)[] out) -> ()
+inline void _foreach_lerp_outf(at::TensorList self, at::TensorList tensors1, const at::Scalar & weight, at::TensorList out) {
+    return at::_ops::_foreach_lerp_Scalar_out::call(self, tensors1, weight, out);
+}
+
+// aten::_foreach_lerp.ScalarList_out(Tensor[] self, Tensor[] tensors1, Scalar[] weight, *, Tensor(a!)[] out) -> ()
+inline void _foreach_lerp_out(at::TensorList out, at::TensorList self, at::TensorList tensors1, at::ArrayRef<at::Scalar> weight) {
+    return at::_ops::_foreach_lerp_ScalarList_out::call(self, tensors1, weight, out);
+}
+// aten::_foreach_lerp.ScalarList_out(Tensor[] self, Tensor[] tensors1, Scalar[] weight, *, Tensor(a!)[] out) -> ()
+inline void _foreach_lerp_outf(at::TensorList self, at::TensorList tensors1, at::ArrayRef<at::Scalar> weight, at::TensorList out) {
+    return at::_ops::_foreach_lerp_ScalarList_out::call(self, tensors1, weight, out);
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_lerp_compositeexplicitautograd_dispatch.h

@@ -0,0 +1,34 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautograd {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_lerp(at::TensorList self, at::TensorList tensors1, at::TensorList weights);
+TORCH_API void _foreach_lerp_out(at::TensorList out, at::TensorList self, at::TensorList tensors1, at::TensorList weights);
+TORCH_API void _foreach_lerp_outf(at::TensorList self, at::TensorList tensors1, at::TensorList weights, at::TensorList out);
+TORCH_API void _foreach_lerp_(at::TensorList self, at::TensorList tensors1, at::TensorList weights);
+TORCH_API ::std::vector<at::Tensor> _foreach_lerp(at::TensorList self, at::TensorList tensors1, const at::Scalar & weight);
+TORCH_API void _foreach_lerp_out(at::TensorList out, at::TensorList self, at::TensorList tensors1, const at::Scalar & weight);
+TORCH_API void _foreach_lerp_outf(at::TensorList self, at::TensorList tensors1, const at::Scalar & weight, at::TensorList out);
+TORCH_API void _foreach_lerp_(at::TensorList self, at::TensorList tensors1, const at::Scalar & weight);
+TORCH_API ::std::vector<at::Tensor> _foreach_lerp(at::TensorList self, at::TensorList tensors1, at::ArrayRef<at::Scalar> weight);
+TORCH_API void _foreach_lerp_out(at::TensorList out, at::TensorList self, at::TensorList tensors1, at::ArrayRef<at::Scalar> weight);
+TORCH_API void _foreach_lerp_outf(at::TensorList self, at::TensorList tensors1, at::ArrayRef<at::Scalar> weight, at::TensorList out);
+TORCH_API void _foreach_lerp_(at::TensorList self, at::TensorList tensors1, at::ArrayRef<at::Scalar> weight);
+
+} // namespace compositeexplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_lgamma_compositeexplicitautograd_dispatch.h

@@ -0,0 +1,26 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautograd {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_lgamma(at::TensorList self);
+TORCH_API void _foreach_lgamma_out(at::TensorList out, at::TensorList self);
+TORCH_API void _foreach_lgamma_outf(at::TensorList self, at::TensorList out);
+TORCH_API void _foreach_lgamma_(at::TensorList self);
+
+} // namespace compositeexplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_log10_compositeexplicitautograd_dispatch.h

@@ -0,0 +1,26 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautograd {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_log10(at::TensorList self);
+TORCH_API void _foreach_log10_out(at::TensorList out, at::TensorList self);
+TORCH_API void _foreach_log10_outf(at::TensorList self, at::TensorList out);
+TORCH_API void _foreach_log10_(at::TensorList self);
+
+} // namespace compositeexplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_maximum_cuda_dispatch.h

@@ -0,0 +1,28 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_maximum(at::TensorList self, const at::Scalar & scalar);
+TORCH_API void _foreach_maximum_(at::TensorList self, const at::Scalar & scalar);
+TORCH_API ::std::vector<at::Tensor> _foreach_maximum(at::TensorList self, at::TensorList other);
+TORCH_API void _foreach_maximum_(at::TensorList self, at::TensorList other);
+TORCH_API ::std::vector<at::Tensor> _foreach_maximum(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API void _foreach_maximum_(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_minimum_cuda_dispatch.h

@@ -0,0 +1,28 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_minimum(at::TensorList self, const at::Scalar & scalar);
+TORCH_API void _foreach_minimum_(at::TensorList self, const at::Scalar & scalar);
+TORCH_API ::std::vector<at::Tensor> _foreach_minimum(at::TensorList self, at::TensorList other);
+TORCH_API void _foreach_minimum_(at::TensorList self, at::TensorList other);
+TORCH_API ::std::vector<at::Tensor> _foreach_minimum(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API void _foreach_minimum_(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_minimum_native.h

@@ -0,0 +1,35 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_clamp_max_scalar_kernel_slow(at::TensorList self, const at::Scalar & scalar);
+TORCH_API void _foreach_minimum_Scalar_out(at::TensorList self, const at::Scalar & scalar, at::TensorList out);
+TORCH_API void foreach_tensor_clamp_max_scalar_kernel_slow_(at::TensorList self, const at::Scalar & scalar);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_clamp_max_scalar_kernel_cuda(at::TensorList self, const at::Scalar & scalar);
+TORCH_API void foreach_tensor_clamp_max_scalar_kernel_cuda_(at::TensorList self, const at::Scalar & scalar);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_clamp_max_list_kernel_slow(at::TensorList self, at::TensorList other);
+TORCH_API void _foreach_minimum_List_out(at::TensorList self, at::TensorList other, at::TensorList out);
+TORCH_API void foreach_tensor_clamp_max_list_kernel_slow_(at::TensorList self, at::TensorList other);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_clamp_max_list_kernel_cuda(at::TensorList self, at::TensorList other);
+TORCH_API void foreach_tensor_clamp_max_list_kernel_cuda_(at::TensorList self, at::TensorList other);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_clamp_max_scalarlist_kernel_slow(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API void _foreach_minimum_ScalarList_out(at::TensorList self, at::ArrayRef<at::Scalar> scalars, at::TensorList out);
+TORCH_API void foreach_tensor_clamp_max_scalarlist_kernel_slow_(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_clamp_max_scalarlist_kernel_cuda(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API void foreach_tensor_clamp_max_scalarlist_kernel_cuda_(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_mul_native.h

@@ -0,0 +1,40 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_mul_scalar_kernel_slow(at::TensorList self, const at::Scalar & scalar);
+TORCH_API void _foreach_mul_Scalar_out(at::TensorList self, const at::Scalar & scalar, at::TensorList out);
+TORCH_API void foreach_tensor_mul_scalar_kernel_slow_(at::TensorList self, const at::Scalar & scalar);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_mul_scalar_kernel_cuda(at::TensorList self, const at::Scalar & scalar);
+TORCH_API void foreach_tensor_mul_scalar_kernel_cuda_(at::TensorList self, const at::Scalar & scalar);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_mul_list_kernel_slow(at::TensorList self, at::TensorList other);
+TORCH_API void _foreach_mul_List_out(at::TensorList self, at::TensorList other, at::TensorList out);
+TORCH_API void foreach_tensor_mul_list_kernel_slow_(at::TensorList self, at::TensorList other);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_mul_list_kernel_cuda(at::TensorList self, at::TensorList other);
+TORCH_API void foreach_tensor_mul_list_kernel_cuda_(at::TensorList self, at::TensorList other);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_mul_scalarlist_kernel_slow(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API void _foreach_mul_ScalarList_out(at::TensorList self, at::ArrayRef<at::Scalar> scalars, at::TensorList out);
+TORCH_API void foreach_tensor_mul_scalarlist_kernel_slow_(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_mul_scalarlist_kernel_cuda(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API void foreach_tensor_mul_scalarlist_kernel_cuda_(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_mul_tensor_kernel_slow(at::TensorList self, const at::Tensor & other);
+TORCH_API void _foreach_mul_Tensor_out(at::TensorList self, const at::Tensor & other, at::TensorList out);
+TORCH_API void foreach_tensor_mul_tensor_kernel_slow_(at::TensorList self, const at::Tensor & other);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_mul_tensor_kernel_cuda(at::TensorList self, const at::Tensor & other);
+TORCH_API void foreach_tensor_mul_tensor_kernel_cuda_(at::TensorList self, const at::Tensor & other);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_norm_ops.h

@@ -0,0 +1,39 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _foreach_norm_Scalar {
+  using schema = ::std::vector<at::Tensor> (at::TensorList, const at::Scalar &, ::std::optional<at::ScalarType>);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_norm";
+  static constexpr const char* overload_name = "Scalar";
+  static constexpr const char* schema_str = "_foreach_norm.Scalar(Tensor[] self, Scalar ord=2, ScalarType? dtype=None) -> Tensor[]";
+  static ::std::vector<at::Tensor> call(at::TensorList self, const at::Scalar & ord, ::std::optional<at::ScalarType> dtype);
+  static ::std::vector<at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, const at::Scalar & ord, ::std::optional<at::ScalarType> dtype);
+};
+
+struct TORCH_API _foreach_norm_Scalar_out {
+  using schema = void (at::TensorList, const at::Scalar &, ::std::optional<at::ScalarType>, at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_norm";
+  static constexpr const char* overload_name = "Scalar_out";
+  static constexpr const char* schema_str = "_foreach_norm.Scalar_out(Tensor[] self, Scalar ord=2, ScalarType? dtype=None, *, Tensor(a!)[] out) -> ()";
+  static void call(at::TensorList self, const at::Scalar & ord, ::std::optional<at::ScalarType> dtype, at::TensorList out);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, const at::Scalar & ord, ::std::optional<at::ScalarType> dtype, at::TensorList out);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_sinh_cuda_dispatch.h

@@ -0,0 +1,24 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_sinh(at::TensorList self);
+TORCH_API void _foreach_sinh_(at::TensorList self);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_sqrt_ops.h

@@ -0,0 +1,50 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _foreach_sqrt {
+  using schema = ::std::vector<at::Tensor> (at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_sqrt";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_foreach_sqrt(Tensor[] self) -> Tensor[]";
+  static ::std::vector<at::Tensor> call(at::TensorList self);
+  static ::std::vector<at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self);
+};
+
+struct TORCH_API _foreach_sqrt_ {
+  using schema = void (at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_sqrt_";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_foreach_sqrt_(Tensor(a!)[] self) -> ()";
+  static void call(at::TensorList self);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self);
+};
+
+struct TORCH_API _foreach_sqrt_out {
+  using schema = void (at::TensorList, at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_sqrt";
+  static constexpr const char* overload_name = "out";
+  static constexpr const char* schema_str = "_foreach_sqrt.out(Tensor[] self, *, Tensor(a!)[] out) -> ()";
+  static void call(at::TensorList self, at::TensorList out);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::TensorList out);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_zero.h

@@ -0,0 +1,44 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_foreach_zero_ops.h>
+
+namespace at {
+
+
+// aten::_foreach_zero_(Tensor(a!)[] self) -> ()
+inline void _foreach_zero_(at::TensorList self) {
+    return at::_ops::_foreach_zero_::call(self);
+}
+
+// aten::_foreach_zero.out(Tensor[] self, *, Tensor(a!)[] out) -> ()
+inline void _foreach_zero_out(at::TensorList out, at::TensorList self) {
+    return at::_ops::_foreach_zero_out::call(self, out);
+}
+// aten::_foreach_zero.out(Tensor[] self, *, Tensor(a!)[] out) -> ()
+inline void _foreach_zero_outf(at::TensorList self, at::TensorList out) {
+    return at::_ops::_foreach_zero_out::call(self, out);
+}
+
+// aten::_foreach_zero(Tensor[] self) -> Tensor[] self_out
+inline ::std::vector<at::Tensor> _foreach_zero(at::TensorList self) {
+    return at::_ops::_foreach_zero::call(self);
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_functional_assert_scalar.h

@@ -0,0 +1,30 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_functional_assert_scalar_ops.h>
+
+namespace at {
+
+
+// aten::_functional_assert_scalar(Scalar self, str assert_msg, Tensor dep_token) -> Tensor
+inline at::Tensor _functional_assert_scalar(const at::Scalar & self, c10::string_view assert_msg, const at::Tensor & dep_token) {
+    return at::_ops::_functional_assert_scalar::call(self, assert_msg, dep_token);
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_fused_moving_avg_obs_fq_helper_cuda_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API ::std::tuple<at::Tensor,at::Tensor> _fused_moving_avg_obs_fq_helper(const at::Tensor & self, const at::Tensor & observer_on, const at::Tensor & fake_quant_on, at::Tensor & running_min, at::Tensor & running_max, at::Tensor & scale, at::Tensor & zero_point, double averaging_const, int64_t quant_min, int64_t quant_max, int64_t ch_axis, bool per_row_fake_quant=false, bool symmetric_quant=false);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_index_put_impl_cpu_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cpu {
+
+TORCH_API at::Tensor & _index_put_impl_(at::Tensor & self, const c10::List<::std::optional<at::Tensor>> & indices, const at::Tensor & values, bool accumulate=false, bool unsafe=false);
+
+} // namespace cpu
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_indices_copy_native.h

@@ -0,0 +1,22 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API at::Tensor & _indices_copy_out(const at::Tensor & self, at::Tensor & out);
+TORCH_API at::Tensor _indices_copy(const at::Tensor & self);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_indices_ops.h

@@ -0,0 +1,28 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _indices {
+  using schema = at::Tensor (const at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_indices";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_indices(Tensor(a) self) -> Tensor(a)";
+  static at::Tensor call(const at::Tensor & self);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_int_mm.h

@@ -0,0 +1,39 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_int_mm_ops.h>
+
+namespace at {
+
+
+// aten::_int_mm(Tensor self, Tensor mat2) -> Tensor
+inline at::Tensor _int_mm(const at::Tensor & self, const at::Tensor & mat2) {
+    return at::_ops::_int_mm::call(self, mat2);
+}
+
+// aten::_int_mm.out(Tensor self, Tensor mat2, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & _int_mm_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & mat2) {
+    return at::_ops::_int_mm_out::call(self, mat2, out);
+}
+// aten::_int_mm.out(Tensor self, Tensor mat2, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & _int_mm_outf(const at::Tensor & self, const at::Tensor & mat2, at::Tensor & out) {
+    return at::_ops::_int_mm_out::call(self, mat2, out);
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_lazy_clone_compositeexplicitautograd_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautograd {
+
+TORCH_API at::Tensor _lazy_clone(const at::Tensor & self);
+
+} // namespace compositeexplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_linalg_svd_meta_dispatch.h

@@ -0,0 +1,25 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace meta {
+
+TORCH_API ::std::tuple<at::Tensor,at::Tensor,at::Tensor> _linalg_svd(const at::Tensor & A, bool full_matrices=false, bool compute_uv=true, ::std::optional<c10::string_view> driver=::std::nullopt);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> _linalg_svd_out(at::Tensor & U, at::Tensor & S, at::Tensor & Vh, const at::Tensor & A, bool full_matrices=false, bool compute_uv=true, ::std::optional<c10::string_view> driver=::std::nullopt);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> _linalg_svd_outf(const at::Tensor & A, bool full_matrices, bool compute_uv, ::std::optional<c10::string_view> driver, at::Tensor & U, at::Tensor & S, at::Tensor & Vh);
+
+} // namespace meta
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_log_softmax_compositeexplicitautogradnonfunctional_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautogradnonfunctional {
+
+TORCH_API at::Tensor _log_softmax(const at::Tensor & self, int64_t dim, bool half_to_float);
+
+} // namespace compositeexplicitautogradnonfunctional
+} // namespace at