Add files using upload-large-folder tool

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_convolution_mode_native.h

@@ -0,0 +1,21 @@
+#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 _convolution_mode_symint(const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::string_view padding, c10::SymIntArrayRef dilation, c10::SymInt groups);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_cummin_helper_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 void _cummin_helper(const at::Tensor & self, at::Tensor & values, at::Tensor & indices, int64_t dim);
+
+} // namespace cpu
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_efficientzerotensor_cpu_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 cpu {
+
+TORCH_API at::Tensor _efficientzerotensor(at::IntArrayRef size, at::TensorOptions options={});
+TORCH_API at::Tensor _efficientzerotensor(at::IntArrayRef size, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory);
+TORCH_API at::Tensor _efficientzerotensor_symint(c10::SymIntArrayRef size, at::TensorOptions options={});
+TORCH_API at::Tensor _efficientzerotensor_symint(c10::SymIntArrayRef size, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory);
+
+} // namespace cpu
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_fake_quantize_learnable_per_tensor_affine_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 & _fake_quantize_learnable_per_tensor_affine_out(const at::Tensor & self, const at::Tensor & scale, const at::Tensor & zero_point, int64_t quant_min, int64_t quant_max, double grad_factor, at::Tensor & out);
+TORCH_API at::Tensor _fake_quantize_learnable_per_tensor_affine(const at::Tensor & self, const at::Tensor & scale, const at::Tensor & zero_point, int64_t quant_min, int64_t quant_max, double grad_factor=1.0);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_foreach_abs_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_abs {
+  using schema = ::std::vector<at::Tensor> (at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_foreach_abs")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_foreach_abs(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_abs_ {
+  using schema = void (at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_foreach_abs_")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_foreach_abs_(Tensor(a!)[] self) -> ()")
+  static void call(at::TensorList self);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self);
+};
+
+struct TORCH_API _foreach_abs_out {
+  using schema = void (at::TensorList, at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_foreach_abs")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_foreach_abs.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

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_foreach_cosh.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_cosh_ops.h>
+
+namespace at {
+
+
+// aten::_foreach_cosh(Tensor[] self) -> Tensor[]
+inline ::std::vector<at::Tensor> _foreach_cosh(at::TensorList self) {
+    return at::_ops::_foreach_cosh::call(self);
+}
+
+// aten::_foreach_cosh_(Tensor(a!)[] self) -> ()
+inline void _foreach_cosh_(at::TensorList self) {
+    return at::_ops::_foreach_cosh_::call(self);
+}
+
+// aten::_foreach_cosh.out(Tensor[] self, *, Tensor(a!)[] out) -> ()
+inline void _foreach_cosh_out(at::TensorList out, at::TensorList self) {
+    return at::_ops::_foreach_cosh_out::call(self, out);
+}
+// aten::_foreach_cosh.out(Tensor[] self, *, Tensor(a!)[] out) -> ()
+inline void _foreach_cosh_outf(at::TensorList self, at::TensorList out) {
+    return at::_ops::_foreach_cosh_out::call(self, out);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_foreach_floor.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_floor_ops.h>
+
+namespace at {
+
+
+// aten::_foreach_floor(Tensor[] self) -> Tensor[]
+inline ::std::vector<at::Tensor> _foreach_floor(at::TensorList self) {
+    return at::_ops::_foreach_floor::call(self);
+}
+
+// aten::_foreach_floor_(Tensor(a!)[] self) -> ()
+inline void _foreach_floor_(at::TensorList self) {
+    return at::_ops::_foreach_floor_::call(self);
+}
+
+// aten::_foreach_floor.out(Tensor[] self, *, Tensor(a!)[] out) -> ()
+inline void _foreach_floor_out(at::TensorList out, at::TensorList self) {
+    return at::_ops::_foreach_floor_out::call(self, out);
+}
+// aten::_foreach_floor.out(Tensor[] self, *, Tensor(a!)[] out) -> ()
+inline void _foreach_floor_outf(at::TensorList self, at::TensorList out) {
+    return at::_ops::_foreach_floor_out::call(self, out);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_foreach_log_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_log(at::TensorList self);
+TORCH_API void _foreach_log_out(at::TensorList out, at::TensorList self);
+TORCH_API void _foreach_log_outf(at::TensorList self, at::TensorList out);
+TORCH_API void _foreach_log_(at::TensorList self);
+
+} // namespace compositeexplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_foreach_minimum_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_minimum(at::TensorList self, const at::Scalar & scalar);
+TORCH_API void _foreach_minimum_out(at::TensorList out, at::TensorList self, const at::Scalar & scalar);
+TORCH_API void _foreach_minimum_outf(at::TensorList self, const at::Scalar & scalar, at::TensorList out);
+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_out(at::TensorList out, at::TensorList self, at::TensorList other);
+TORCH_API void _foreach_minimum_outf(at::TensorList self, at::TensorList other, at::TensorList out);
+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_out(at::TensorList out, at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API void _foreach_minimum_outf(at::TensorList self, at::ArrayRef<at::Scalar> scalars, at::TensorList out);
+TORCH_API void _foreach_minimum_(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+
+} // namespace compositeexplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_linalg_check_errors_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 void _linalg_check_errors(const at::Tensor & info, c10::string_view api_name, bool is_matrix);
+
+} // namespace compositeexplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_linalg_slogdet_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 _linalg_slogdet {
+  using schema = ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor> (const at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_linalg_slogdet")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_linalg_slogdet(Tensor A) -> (Tensor sign, Tensor logabsdet, Tensor LU, Tensor pivots)")
+  static ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor> call(const at::Tensor & A);
+  static ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & A);
+};
+
+struct TORCH_API _linalg_slogdet_sign {
+  using schema = ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> (const at::Tensor &, at::Tensor &, at::Tensor &, at::Tensor &, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_linalg_slogdet")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "sign")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_linalg_slogdet.sign(Tensor A, *, Tensor(a!) sign, Tensor(b!) logabsdet, Tensor(c!) LU, Tensor(d!) pivots) -> (Tensor(a!) sign, Tensor(b!) logabsdet, Tensor(c!) LU, Tensor(d!) pivots)")
+  static ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> call(const at::Tensor & A, at::Tensor & sign, at::Tensor & logabsdet, at::Tensor & LU, at::Tensor & pivots);
+  static ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & A, at::Tensor & sign, at::Tensor & logabsdet, at::Tensor & LU, at::Tensor & pivots);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_lstm_mps_compositeexplicitautograd_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 compositeexplicitautograd {
+
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> _lstm_mps_out(at::Tensor & out0, at::Tensor & out1, at::Tensor & out2, at::Tensor & out3, at::Tensor & out4, at::Tensor & out5, const at::Tensor & input, at::TensorList hx, at::TensorList params, bool has_biases, int64_t num_layers, double dropout, bool train, bool bidirectional, bool batch_first);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> _lstm_mps_outf(const at::Tensor & input, at::TensorList hx, at::TensorList params, bool has_biases, int64_t num_layers, double dropout, bool train, bool bidirectional, bool batch_first, at::Tensor & out0, at::Tensor & out1, at::Tensor & out2, at::Tensor & out3, at::Tensor & out4, at::Tensor & out5);
+
+} // namespace compositeexplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_make_dep_token_cpu_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 cpu {
+
+TORCH_API at::Tensor _make_dep_token(at::TensorOptions options={}, ::std::optional<at::MemoryFormat> memory_format=::std::nullopt);
+TORCH_API at::Tensor _make_dep_token(::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);
+
+} // namespace cpu
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_native_multi_head_attention_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 _native_multi_head_attention {
+  using schema = ::std::tuple<at::Tensor,at::Tensor> (const at::Tensor &, const at::Tensor &, const at::Tensor &, int64_t, int64_t, const at::Tensor &, const at::Tensor &, const at::Tensor &, const at::Tensor &, const ::std::optional<at::Tensor> &, bool, bool, ::std::optional<int64_t>);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_native_multi_head_attention")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_native_multi_head_attention(Tensor query, Tensor key, Tensor value, int embed_dim, int num_head, Tensor qkv_weight, Tensor qkv_bias, Tensor proj_weight, Tensor proj_bias, Tensor? mask=None, bool need_weights=True, bool average_attn_weights=True, int? mask_type=None) -> (Tensor, Tensor)")
+  static ::std::tuple<at::Tensor,at::Tensor> call(const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, int64_t embed_dim, int64_t num_head, const at::Tensor & qkv_weight, const at::Tensor & qkv_bias, const at::Tensor & proj_weight, const at::Tensor & proj_bias, const ::std::optional<at::Tensor> & mask, bool need_weights, bool average_attn_weights, ::std::optional<int64_t> mask_type);
+  static ::std::tuple<at::Tensor,at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, int64_t embed_dim, int64_t num_head, const at::Tensor & qkv_weight, const at::Tensor & qkv_bias, const at::Tensor & proj_weight, const at::Tensor & proj_bias, const ::std::optional<at::Tensor> & mask, bool need_weights, bool average_attn_weights, ::std::optional<int64_t> mask_type);
+};
+
+struct TORCH_API _native_multi_head_attention_out {
+  using schema = ::std::tuple<at::Tensor &,at::Tensor &> (const at::Tensor &, const at::Tensor &, const at::Tensor &, int64_t, int64_t, const at::Tensor &, const at::Tensor &, const at::Tensor &, const at::Tensor &, const ::std::optional<at::Tensor> &, bool, bool, ::std::optional<int64_t>, at::Tensor &, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_native_multi_head_attention")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_native_multi_head_attention.out(Tensor query, Tensor key, Tensor value, int embed_dim, int num_head, Tensor qkv_weight, Tensor qkv_bias, Tensor proj_weight, Tensor proj_bias, Tensor? mask=None, bool need_weights=True, bool average_attn_weights=True, int? mask_type=None, *, Tensor(a!) out0, Tensor(b!) out1) -> (Tensor(a!), Tensor(b!))")
+  static ::std::tuple<at::Tensor &,at::Tensor &> call(const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, int64_t embed_dim, int64_t num_head, const at::Tensor & qkv_weight, const at::Tensor & qkv_bias, const at::Tensor & proj_weight, const at::Tensor & proj_bias, const ::std::optional<at::Tensor> & mask, bool need_weights, bool average_attn_weights, ::std::optional<int64_t> mask_type, at::Tensor & out0, at::Tensor & out1);
+  static ::std::tuple<at::Tensor &,at::Tensor &> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, int64_t embed_dim, int64_t num_head, const at::Tensor & qkv_weight, const at::Tensor & qkv_bias, const at::Tensor & proj_weight, const at::Tensor & proj_bias, const ::std::optional<at::Tensor> & mask, bool need_weights, bool average_attn_weights, ::std::optional<int64_t> mask_type, at::Tensor & out0, at::Tensor & out1);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_nested_compute_contiguous_strides_offsets_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 _nested_compute_contiguous_strides_offsets {
+  using schema = ::std::tuple<at::Tensor,at::Tensor> (const at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_nested_compute_contiguous_strides_offsets")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_nested_compute_contiguous_strides_offsets(Tensor nested_size) -> (Tensor, Tensor)")
+  static ::std::tuple<at::Tensor,at::Tensor> call(const at::Tensor & nested_size);
+  static ::std::tuple<at::Tensor,at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & nested_size);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_nested_tensor_from_tensor_list_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 _nested_tensor_from_tensor_list(at::TensorList list, ::std::optional<at::ScalarType> dtype=::std::nullopt, ::std::optional<at::Layout> layout=::std::nullopt, ::std::optional<at::Device> device=::std::nullopt, ::std::optional<bool> pin_memory=::std::nullopt);
+TORCH_API at::Tensor & _nested_tensor_from_tensor_list_out(at::TensorList list, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory, at::Tensor & out);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_scaled_dot_product_flash_attention_for_cpu_backward_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 ::std::tuple<at::Tensor,at::Tensor,at::Tensor> _scaled_dot_product_flash_attention_for_cpu_backward(const at::Tensor & grad_out, const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const at::Tensor & out, const at::Tensor & logsumexp, double dropout_p, bool is_causal, const ::std::optional<at::Tensor> & attn_mask={}, ::std::optional<double> scale=::std::nullopt);
+
+} // namespace cpu
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_sparse_compressed_tensor_with_dims_compositeexplicitautograd_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 compositeexplicitautograd {
+
+TORCH_API at::Tensor _sparse_compressed_tensor_with_dims(int64_t nnz, int64_t dense_dim, at::IntArrayRef size, at::IntArrayRef blocksize, at::ScalarType index_dtype, at::TensorOptions options);
+TORCH_API at::Tensor _sparse_compressed_tensor_with_dims(int64_t nnz, int64_t dense_dim, at::IntArrayRef size, at::IntArrayRef blocksize, at::ScalarType index_dtype, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory);
+
+} // namespace compositeexplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_test_autograd_multiple_dispatch_view_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 _test_autograd_multiple_dispatch_view(const at::Tensor & self);
+
+} // namespace compositeexplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_test_warn_in_autograd_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 _test_warn_in_autograd {
+  using schema = at::Tensor (const at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_test_warn_in_autograd")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_test_warn_in_autograd(Tensor self) -> Tensor")
+  static at::Tensor call(const at::Tensor & self);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self);
+};
+
+struct TORCH_API _test_warn_in_autograd_out {
+  using schema = at::Tensor & (const at::Tensor &, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_test_warn_in_autograd")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_test_warn_in_autograd.out(Tensor self, *, Tensor(a!) out) -> Tensor(a!)")
+  static at::Tensor & call(const at::Tensor & self, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, at::Tensor & out);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_trilinear_compositeexplicitautograd_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 compositeexplicitautograd {
+
+TORCH_API at::Tensor & _trilinear_out(at::Tensor & out, const at::Tensor & i1, const at::Tensor & i2, const at::Tensor & i3, at::IntArrayRef expand1, at::IntArrayRef expand2, at::IntArrayRef expand3, at::IntArrayRef sumdim, int64_t unroll_dim=1);
+TORCH_API at::Tensor & _trilinear_outf(const at::Tensor & i1, const at::Tensor & i2, const at::Tensor & i3, at::IntArrayRef expand1, at::IntArrayRef expand2, at::IntArrayRef expand3, at::IntArrayRef sumdim, int64_t unroll_dim, at::Tensor & out);
+
+} // namespace compositeexplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_triton_scaled_dot_attention_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 & _triton_scaled_dot_attention_out(const at::Tensor & q, const at::Tensor & k, const at::Tensor & v, double dropout_p, at::Tensor & out);
+TORCH_API at::Tensor triton_scaled_dot_attention(const at::Tensor & q, const at::Tensor & k, const at::Tensor & v, double dropout_p=0.0);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_validate_sparse_bsc_tensor_args.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/_validate_sparse_bsc_tensor_args_ops.h>
+
+namespace at {
+
+
+// aten::_validate_sparse_bsc_tensor_args(Tensor ccol_indices, Tensor row_indices, Tensor values, int[] size) -> ()
+inline void _validate_sparse_bsc_tensor_args(const at::Tensor & ccol_indices, const at::Tensor & row_indices, const at::Tensor & values, at::IntArrayRef size) {
+    return at::_ops::_validate_sparse_bsc_tensor_args::call(ccol_indices, row_indices, values, size);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_validate_sparse_coo_tensor_args_native.h

@@ -0,0 +1,21 @@
+#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 _validate_sparse_coo_tensor_args(const at::Tensor & indices, const at::Tensor & values, at::IntArrayRef size, ::std::optional<bool> is_coalesced=::std::nullopt);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/acosh_cuda_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 cuda {
+
+TORCH_API at::Tensor acosh(const at::Tensor & self);
+TORCH_API at::Tensor & acosh_out(at::Tensor & out, const at::Tensor & self);
+TORCH_API at::Tensor & acosh_outf(const at::Tensor & self, at::Tensor & out);
+TORCH_API at::Tensor & acosh_(at::Tensor & self);
+
+} // namespace cuda
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/adaptive_max_pool2d_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/adaptive_max_pool2d_meta.h>
+
+namespace at {
+namespace native {
+struct TORCH_API structured_adaptive_max_pool2d_out_cpu : public at::meta::structured_adaptive_max_pool2d {
+void impl(const at::Tensor & self, at::IntArrayRef output_size, const at::Tensor & out, const at::Tensor & indices);
+};
+struct TORCH_API structured_adaptive_max_pool2d_out_cuda : public at::meta::structured_adaptive_max_pool2d {
+void impl(const at::Tensor & self, at::IntArrayRef output_size, const at::Tensor & out, const at::Tensor & indices);
+};
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/addmv_native.h

@@ -0,0 +1,28 @@
+#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/addmv_meta.h>
+
+namespace at {
+namespace native {
+struct TORCH_API structured_addmv_out_cpu : public at::meta::structured_addmv {
+void impl(const at::Tensor & self, const at::Tensor & mat, const at::Tensor & vec, const at::Scalar & beta, const at::Scalar & alpha, const at::Tensor & out);
+};
+struct TORCH_API structured_addmv_out_cuda : public at::meta::structured_addmv {
+void impl(const at::Tensor & self, const at::Tensor & mat, const at::Tensor & vec, const at::Scalar & beta, const at::Scalar & alpha, const at::Tensor & out);
+};
+TORCH_API at::Tensor & addmv_out_sparse_compressed(const at::Tensor & self, const at::Tensor & mat, const at::Tensor & vec, const at::Scalar & beta, const at::Scalar & alpha, at::Tensor & out);
+TORCH_API at::Tensor & addmv_out_sparse_compressed_cuda(const at::Tensor & self, const at::Tensor & mat, const at::Tensor & vec, const at::Scalar & beta, const at::Scalar & alpha, at::Tensor & out);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/argwhere_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 argwhere(const at::Tensor & self);
+
+} // namespace compositeimplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/batch_norm_native.h

@@ -0,0 +1,21 @@
+#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 batch_norm(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);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/broadcast_tensors_native.h

@@ -0,0 +1,21 @@
+#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> broadcast_tensors(at::TensorList tensors);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/chunk_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 ::std::vector<at::Tensor> chunk(const at::Tensor & self, int64_t chunks, int64_t dim=0);
+TORCH_API ::std::vector<at::Tensor> chunk_nested_tensor(const at::Tensor & self, int64_t chunks, int64_t dim=0);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/conj_physical_cpu_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 cpu {
+
+TORCH_API at::Tensor & conj_physical_out(at::Tensor & out, const at::Tensor & self);
+TORCH_API at::Tensor & conj_physical_outf(const at::Tensor & self, at::Tensor & out);
+
+} // namespace cpu
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/constant_pad_nd_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 constant_pad_nd(const at::Tensor & self, at::IntArrayRef pad, const at::Scalar & value=0);
+TORCH_API at::Tensor & constant_pad_nd_out_symint(const at::Tensor & self, c10::SymIntArrayRef pad, const at::Scalar & value, at::Tensor & out);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/cosh_compositeexplicitautogradnonfunctional_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 compositeexplicitautogradnonfunctional {
+
+TORCH_API at::Tensor cosh(const at::Tensor & self);
+TORCH_API at::Tensor & cosh_(at::Tensor & self);
+
+} // namespace compositeexplicitautogradnonfunctional
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/cumsum_meta.h

@@ -0,0 +1,27 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeMetaFunction.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/TensorIterator.h>
+#include <ATen/TensorMeta.h>
+#include <tuple>
+#include <vector>
+
+namespace at {
+namespace meta {
+
+struct TORCH_API structured_cumsum : public at::impl::MetaBase {
+    
+    
+    void meta(const at::Tensor & self, int64_t dim, ::std::optional<at::ScalarType> dtype);
+};
+
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/empty_strided_cpu_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 cpu {
+
+TORCH_API at::Tensor empty_strided(at::IntArrayRef size, at::IntArrayRef stride, at::TensorOptions options={});
+TORCH_API at::Tensor empty_strided(at::IntArrayRef size, at::IntArrayRef stride, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory);
+TORCH_API at::Tensor empty_strided_symint(c10::SymIntArrayRef size, c10::SymIntArrayRef stride, at::TensorOptions options={});
+TORCH_API at::Tensor empty_strided_symint(c10::SymIntArrayRef size, c10::SymIntArrayRef stride, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory);
+
+} // namespace cpu
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/equal_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 bool cpu_equal(const at::Tensor & self, const at::Tensor & other);
+TORCH_API bool cuda_equal(const at::Tensor & self, const at::Tensor & other);
+TORCH_API bool equal_quantized_cpu(const at::Tensor & self, const at::Tensor & other);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/erf_cpu_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 cpu {
+
+TORCH_API at::Tensor erf(const at::Tensor & self);
+TORCH_API at::Tensor & erf_out(at::Tensor & out, const at::Tensor & self);
+TORCH_API at::Tensor & erf_outf(const at::Tensor & self, at::Tensor & out);
+TORCH_API at::Tensor & erf_(at::Tensor & self);
+
+} // namespace cpu
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/expand_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 & expand_copy_out_symint(const at::Tensor & self, c10::SymIntArrayRef size, bool implicit, at::Tensor & out);
+TORCH_API at::Tensor expand_copy_symint(const at::Tensor & self, c10::SymIntArrayRef size, bool implicit=false);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/exponential_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 exponential(const at::Tensor & self, double lambd=1, ::std::optional<at::Generator> generator=::std::nullopt);
+TORCH_API at::Tensor & exponential_out(const at::Tensor & self, double lambd, ::std::optional<at::Generator> generator, at::Tensor & out);
+TORCH_API at::Tensor & exponential_(at::Tensor & self, double lambd=1, ::std::optional<at::Generator> generator=::std::nullopt);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/fake_quantize_per_tensor_affine_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 {
+  using schema = at::Tensor (const at::Tensor &, double, int64_t, int64_t, int64_t);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::fake_quantize_per_tensor_affine")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "fake_quantize_per_tensor_affine(Tensor self, float scale, int zero_point, int quant_min, int quant_max) -> Tensor")
+  static at::Tensor call(const at::Tensor & self, double scale, int64_t zero_point, int64_t quant_min, int64_t quant_max);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, double scale, int64_t zero_point, int64_t quant_min, int64_t quant_max);
+};
+
+struct TORCH_API fake_quantize_per_tensor_affine_tensor_qparams {
+  using schema = 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_STR_INL_EXCEPT_WIN_CUDA(name, "aten::fake_quantize_per_tensor_affine")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "tensor_qparams")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "fake_quantize_per_tensor_affine.tensor_qparams(Tensor self, Tensor scale, Tensor zero_point, int quant_min, int quant_max) -> Tensor")
+  static at::Tensor call(const at::Tensor & self, const at::Tensor & scale, const at::Tensor & zero_point, int64_t quant_min, int64_t quant_max);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Tensor & scale, const at::Tensor & zero_point, int64_t quant_min, int64_t quant_max);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/feature_dropout.h

@@ -0,0 +1,35 @@
+#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/feature_dropout_ops.h>
+
+namespace at {
+
+
+// aten::feature_dropout(Tensor input, float p, bool train) -> Tensor
+inline at::Tensor feature_dropout(const at::Tensor & input, double p, bool train) {
+    return at::_ops::feature_dropout::call(input, p, train);
+}
+
+// aten::feature_dropout_(Tensor(a!) self, float p, bool train) -> Tensor(a!)
+inline at::Tensor & feature_dropout_(at::Tensor & self, double p, bool train) {
+    return at::_ops::feature_dropout_::call(self, p, train);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/fft_hfft2_compositeimplicitautograd_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 compositeimplicitautograd {
+
+TORCH_API at::Tensor fft_hfft2(const at::Tensor & self, at::OptionalIntArrayRef s=::std::nullopt, at::IntArrayRef dim={-2,-1}, ::std::optional<c10::string_view> norm=::std::nullopt);
+TORCH_API at::Tensor fft_hfft2_symint(const at::Tensor & self, at::OptionalSymIntArrayRef s=::std::nullopt, at::IntArrayRef dim={-2,-1}, ::std::optional<c10::string_view> norm=::std::nullopt);
+TORCH_API const at::Tensor & fft_hfft2_out(const at::Tensor & out, const at::Tensor & self, at::OptionalIntArrayRef s=::std::nullopt, at::IntArrayRef dim={-2,-1}, ::std::optional<c10::string_view> norm=::std::nullopt);
+TORCH_API const at::Tensor & fft_hfft2_outf(const at::Tensor & self, at::OptionalIntArrayRef s, at::IntArrayRef dim, ::std::optional<c10::string_view> norm, const at::Tensor & out);
+TORCH_API const at::Tensor & fft_hfft2_symint_out(const at::Tensor & out, const at::Tensor & self, at::OptionalSymIntArrayRef s=::std::nullopt, at::IntArrayRef dim={-2,-1}, ::std::optional<c10::string_view> norm=::std::nullopt);
+TORCH_API const at::Tensor & fft_hfft2_symint_outf(const at::Tensor & self, at::OptionalSymIntArrayRef s, at::IntArrayRef dim, ::std::optional<c10::string_view> norm, const at::Tensor & out);
+
+} // namespace compositeimplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/fft_ihfft_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 fft_ihfft {
+  using schema = at::Tensor (const at::Tensor &, ::std::optional<c10::SymInt>, int64_t, ::std::optional<c10::string_view>);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::fft_ihfft")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "fft_ihfft(Tensor self, SymInt? n=None, int dim=-1, str? norm=None) -> Tensor")
+  static at::Tensor call(const at::Tensor & self, ::std::optional<c10::SymInt> n, int64_t dim, ::std::optional<c10::string_view> norm);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, ::std::optional<c10::SymInt> n, int64_t dim, ::std::optional<c10::string_view> norm);
+};
+
+struct TORCH_API fft_ihfft_out {
+  using schema = at::Tensor & (const at::Tensor &, ::std::optional<c10::SymInt>, int64_t, ::std::optional<c10::string_view>, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::fft_ihfft")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "fft_ihfft.out(Tensor self, SymInt? n=None, int dim=-1, str? norm=None, *, Tensor(a!) out) -> Tensor(a!)")
+  static at::Tensor & call(const at::Tensor & self, ::std::optional<c10::SymInt> n, int64_t dim, ::std::optional<c10::string_view> norm, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, ::std::optional<c10::SymInt> n, int64_t dim, ::std::optional<c10::string_view> norm, at::Tensor & out);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/fill_cpu_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 cpu {
+
+TORCH_API at::Tensor & fill_(at::Tensor & self, const at::Scalar & value);
+TORCH_API at::Tensor & fill_(at::Tensor & self, const at::Tensor & value);
+
+} // namespace cpu
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/floor.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/floor_ops.h>
+
+namespace at {
+
+
+// aten::floor(Tensor self) -> Tensor
+inline at::Tensor floor(const at::Tensor & self) {
+    return at::_ops::floor::call(self);
+}
+
+// aten::floor_(Tensor(a!) self) -> Tensor(a!)
+inline at::Tensor & floor_(at::Tensor & self) {
+    return at::_ops::floor_::call(self);
+}
+
+// aten::floor.out(Tensor self, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & floor_out(at::Tensor & out, const at::Tensor & self) {
+    return at::_ops::floor_out::call(self, out);
+}
+// aten::floor.out(Tensor self, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & floor_outf(const at::Tensor & self, at::Tensor & out) {
+    return at::_ops::floor_out::call(self, out);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/fused_moving_avg_obs_fake_quant.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/fused_moving_avg_obs_fake_quant_ops.h>
+
+namespace at {
+
+
+// aten::fused_moving_avg_obs_fake_quant(Tensor self, Tensor observer_on, Tensor fake_quant_on, Tensor(a!) running_min, Tensor(b!) running_max, Tensor(c!) scale, Tensor(d!) zero_point, float averaging_const, int quant_min, int quant_max, int ch_axis, bool per_row_fake_quant=False, bool symmetric_quant=False) -> Tensor
+inline at::Tensor fused_moving_avg_obs_fake_quant(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) {
+    return at::_ops::fused_moving_avg_obs_fake_quant::call(self, observer_on, fake_quant_on, running_min, running_max, scale, zero_point, averaging_const, quant_min, quant_max, ch_axis, per_row_fake_quant, symmetric_quant);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/is_distributed.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/is_distributed_ops.h>
+
+namespace at {
+
+
+// aten::is_distributed(Tensor self) -> bool
+inline bool is_distributed(const at::Tensor & self) {
+    return at::_ops::is_distributed::call(self);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/is_nonzero_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 bool is_nonzero(const at::Tensor & self);
+
+} // namespace compositeimplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_matrix_rank.h

@@ -0,0 +1,81 @@
+#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/linalg_matrix_rank_ops.h>
+
+namespace at {
+
+
+// aten::linalg_matrix_rank.atol_rtol_tensor(Tensor input, *, Tensor? atol=None, Tensor? rtol=None, bool hermitian=False) -> Tensor
+inline at::Tensor linalg_matrix_rank(const at::Tensor & input, const ::std::optional<at::Tensor> & atol={}, const ::std::optional<at::Tensor> & rtol={}, bool hermitian=false) {
+    return at::_ops::linalg_matrix_rank_atol_rtol_tensor::call(input, atol, rtol, hermitian);
+}
+
+// aten::linalg_matrix_rank.atol_rtol_tensor_out(Tensor input, *, Tensor? atol=None, Tensor? rtol=None, bool hermitian=False, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & linalg_matrix_rank_out(at::Tensor & out, const at::Tensor & input, const ::std::optional<at::Tensor> & atol={}, const ::std::optional<at::Tensor> & rtol={}, bool hermitian=false) {
+    return at::_ops::linalg_matrix_rank_atol_rtol_tensor_out::call(input, atol, rtol, hermitian, out);
+}
+// aten::linalg_matrix_rank.atol_rtol_tensor_out(Tensor input, *, Tensor? atol=None, Tensor? rtol=None, bool hermitian=False, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & linalg_matrix_rank_outf(const at::Tensor & input, const ::std::optional<at::Tensor> & atol, const ::std::optional<at::Tensor> & rtol, bool hermitian, at::Tensor & out) {
+    return at::_ops::linalg_matrix_rank_atol_rtol_tensor_out::call(input, atol, rtol, hermitian, out);
+}
+
+// aten::linalg_matrix_rank.atol_rtol_float(Tensor self, *, float? atol=None, float? rtol=None, bool hermitian=False) -> Tensor
+inline at::Tensor linalg_matrix_rank(const at::Tensor & self, ::std::optional<double> atol, ::std::optional<double> rtol, bool hermitian=false) {
+    return at::_ops::linalg_matrix_rank_atol_rtol_float::call(self, atol, rtol, hermitian);
+}
+
+// aten::linalg_matrix_rank.atol_rtol_float_out(Tensor self, *, float? atol=None, float? rtol=None, bool hermitian=False, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & linalg_matrix_rank_out(at::Tensor & out, const at::Tensor & self, ::std::optional<double> atol, ::std::optional<double> rtol, bool hermitian=false) {
+    return at::_ops::linalg_matrix_rank_atol_rtol_float_out::call(self, atol, rtol, hermitian, out);
+}
+// aten::linalg_matrix_rank.atol_rtol_float_out(Tensor self, *, float? atol=None, float? rtol=None, bool hermitian=False, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & linalg_matrix_rank_outf(const at::Tensor & self, ::std::optional<double> atol, ::std::optional<double> rtol, bool hermitian, at::Tensor & out) {
+    return at::_ops::linalg_matrix_rank_atol_rtol_float_out::call(self, atol, rtol, hermitian, out);
+}
+
+// aten::linalg_matrix_rank(Tensor self, float tol, bool hermitian=False) -> Tensor
+inline at::Tensor linalg_matrix_rank(const at::Tensor & self, double tol, bool hermitian=false) {
+    return at::_ops::linalg_matrix_rank::call(self, tol, hermitian);
+}
+
+// aten::linalg_matrix_rank.out(Tensor self, float tol, bool hermitian=False, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & linalg_matrix_rank_out(at::Tensor & out, const at::Tensor & self, double tol, bool hermitian=false) {
+    return at::_ops::linalg_matrix_rank_out::call(self, tol, hermitian, out);
+}
+// aten::linalg_matrix_rank.out(Tensor self, float tol, bool hermitian=False, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & linalg_matrix_rank_outf(const at::Tensor & self, double tol, bool hermitian, at::Tensor & out) {
+    return at::_ops::linalg_matrix_rank_out::call(self, tol, hermitian, out);
+}
+
+// aten::linalg_matrix_rank.tol_tensor(Tensor input, Tensor tol, bool hermitian=False) -> Tensor
+inline at::Tensor linalg_matrix_rank(const at::Tensor & input, const at::Tensor & tol, bool hermitian=false) {
+    return at::_ops::linalg_matrix_rank_tol_tensor::call(input, tol, hermitian);
+}
+
+// aten::linalg_matrix_rank.out_tol_tensor(Tensor input, Tensor tol, bool hermitian=False, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & linalg_matrix_rank_out(at::Tensor & out, const at::Tensor & input, const at::Tensor & tol, bool hermitian=false) {
+    return at::_ops::linalg_matrix_rank_out_tol_tensor::call(input, tol, hermitian, out);
+}
+// aten::linalg_matrix_rank.out_tol_tensor(Tensor input, Tensor tol, bool hermitian=False, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & linalg_matrix_rank_outf(const at::Tensor & input, const at::Tensor & tol, bool hermitian, at::Tensor & out) {
+    return at::_ops::linalg_matrix_rank_out_tol_tensor::call(input, tol, hermitian, out);
+}
+
+}