Add files using upload-large-folder tool

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_conv_depthwise2d.h

@@ -0,0 +1,91 @@
+#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/_conv_depthwise2d_ops.h>
+
+namespace at {
+
+
+// aten::_conv_depthwise2d.out(Tensor self, Tensor weight, SymInt[2] kernel_size, Tensor? bias, SymInt[2] stride, SymInt[2] padding, SymInt[2] dilation, *, Tensor(a!) out) -> Tensor(a!)
+inline const at::Tensor & _conv_depthwise2d_out(const at::Tensor & out, const at::Tensor & self, const at::Tensor & weight, at::IntArrayRef kernel_size, const ::std::optional<at::Tensor> & bias, at::IntArrayRef stride, at::IntArrayRef padding, at::IntArrayRef dilation) {
+    return at::_ops::_conv_depthwise2d_out::call(self, weight, c10::fromIntArrayRefSlow(kernel_size), bias, c10::fromIntArrayRefSlow(stride), c10::fromIntArrayRefSlow(padding), c10::fromIntArrayRefSlow(dilation), out);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
+  const at::Tensor & _conv_depthwise2d_out(const at::Tensor & out, const at::Tensor & self, const at::Tensor & weight, at::IntArrayRef kernel_size, const ::std::optional<at::Tensor> & bias, at::IntArrayRef stride, at::IntArrayRef padding, at::IntArrayRef dilation) {
+    return at::_ops::_conv_depthwise2d_out::call(self, weight, c10::fromIntArrayRefSlow(kernel_size), bias, c10::fromIntArrayRefSlow(stride), c10::fromIntArrayRefSlow(padding), c10::fromIntArrayRefSlow(dilation), out);
+  }
+}
+
+// aten::_conv_depthwise2d.out(Tensor self, Tensor weight, SymInt[2] kernel_size, Tensor? bias, SymInt[2] stride, SymInt[2] padding, SymInt[2] dilation, *, Tensor(a!) out) -> Tensor(a!)
+inline const at::Tensor & _conv_depthwise2d_outf(const at::Tensor & self, const at::Tensor & weight, at::IntArrayRef kernel_size, const ::std::optional<at::Tensor> & bias, at::IntArrayRef stride, at::IntArrayRef padding, at::IntArrayRef dilation, const at::Tensor & out) {
+    return at::_ops::_conv_depthwise2d_out::call(self, weight, c10::fromIntArrayRefSlow(kernel_size), bias, c10::fromIntArrayRefSlow(stride), c10::fromIntArrayRefSlow(padding), c10::fromIntArrayRefSlow(dilation), out);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
+  const at::Tensor & _conv_depthwise2d_outf(const at::Tensor & self, const at::Tensor & weight, at::IntArrayRef kernel_size, const ::std::optional<at::Tensor> & bias, at::IntArrayRef stride, at::IntArrayRef padding, at::IntArrayRef dilation, const at::Tensor & out) {
+    return at::_ops::_conv_depthwise2d_out::call(self, weight, c10::fromIntArrayRefSlow(kernel_size), bias, c10::fromIntArrayRefSlow(stride), c10::fromIntArrayRefSlow(padding), c10::fromIntArrayRefSlow(dilation), out);
+  }
+}
+
+// aten::_conv_depthwise2d.out(Tensor self, Tensor weight, SymInt[2] kernel_size, Tensor? bias, SymInt[2] stride, SymInt[2] padding, SymInt[2] dilation, *, Tensor(a!) out) -> Tensor(a!)
+inline const at::Tensor & _conv_depthwise2d_symint_out(const at::Tensor & out, const at::Tensor & self, const at::Tensor & weight, c10::SymIntArrayRef kernel_size, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::SymIntArrayRef padding, c10::SymIntArrayRef dilation) {
+    return at::_ops::_conv_depthwise2d_out::call(self, weight, kernel_size, bias, stride, padding, dilation, out);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
+  const at::Tensor & _conv_depthwise2d_out(const at::Tensor & out, const at::Tensor & self, const at::Tensor & weight, c10::SymIntArrayRef kernel_size, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::SymIntArrayRef padding, c10::SymIntArrayRef dilation) {
+    return at::_ops::_conv_depthwise2d_out::call(self, weight, kernel_size, bias, stride, padding, dilation, out);
+  }
+}
+
+// aten::_conv_depthwise2d.out(Tensor self, Tensor weight, SymInt[2] kernel_size, Tensor? bias, SymInt[2] stride, SymInt[2] padding, SymInt[2] dilation, *, Tensor(a!) out) -> Tensor(a!)
+inline const at::Tensor & _conv_depthwise2d_symint_outf(const at::Tensor & self, const at::Tensor & weight, c10::SymIntArrayRef kernel_size, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::SymIntArrayRef padding, c10::SymIntArrayRef dilation, const at::Tensor & out) {
+    return at::_ops::_conv_depthwise2d_out::call(self, weight, kernel_size, bias, stride, padding, dilation, out);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
+  const at::Tensor & _conv_depthwise2d_outf(const at::Tensor & self, const at::Tensor & weight, c10::SymIntArrayRef kernel_size, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::SymIntArrayRef padding, c10::SymIntArrayRef dilation, const at::Tensor & out) {
+    return at::_ops::_conv_depthwise2d_out::call(self, weight, kernel_size, bias, stride, padding, dilation, out);
+  }
+}
+
+// aten::_conv_depthwise2d(Tensor self, Tensor weight, SymInt[2] kernel_size, Tensor? bias, SymInt[2] stride, SymInt[2] padding, SymInt[2] dilation) -> Tensor
+inline at::Tensor _conv_depthwise2d(const at::Tensor & self, const at::Tensor & weight, at::IntArrayRef kernel_size, const ::std::optional<at::Tensor> & bias, at::IntArrayRef stride, at::IntArrayRef padding, at::IntArrayRef dilation) {
+    return at::_ops::_conv_depthwise2d::call(self, weight, c10::fromIntArrayRefSlow(kernel_size), bias, c10::fromIntArrayRefSlow(stride), c10::fromIntArrayRefSlow(padding), c10::fromIntArrayRefSlow(dilation));
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
+  at::Tensor _conv_depthwise2d(const at::Tensor & self, const at::Tensor & weight, at::IntArrayRef kernel_size, const ::std::optional<at::Tensor> & bias, at::IntArrayRef stride, at::IntArrayRef padding, at::IntArrayRef dilation) {
+    return at::_ops::_conv_depthwise2d::call(self, weight, c10::fromIntArrayRefSlow(kernel_size), bias, c10::fromIntArrayRefSlow(stride), c10::fromIntArrayRefSlow(padding), c10::fromIntArrayRefSlow(dilation));
+  }
+}
+
+// aten::_conv_depthwise2d(Tensor self, Tensor weight, SymInt[2] kernel_size, Tensor? bias, SymInt[2] stride, SymInt[2] padding, SymInt[2] dilation) -> Tensor
+inline at::Tensor _conv_depthwise2d_symint(const at::Tensor & self, const at::Tensor & weight, c10::SymIntArrayRef kernel_size, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::SymIntArrayRef padding, c10::SymIntArrayRef dilation) {
+    return at::_ops::_conv_depthwise2d::call(self, weight, kernel_size, bias, stride, padding, dilation);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
+  at::Tensor _conv_depthwise2d(const at::Tensor & self, const at::Tensor & weight, c10::SymIntArrayRef kernel_size, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::SymIntArrayRef padding, c10::SymIntArrayRef dilation) {
+    return at::_ops::_conv_depthwise2d::call(self, weight, kernel_size, bias, stride, padding, dilation);
+  }
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_convert_indices_from_coo_to_csr_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 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 meta
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_cslt_sparse_mm_search_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 _cslt_sparse_mm_search {
+  using schema = int64_t (const at::Tensor &, const at::Tensor &, const ::std::optional<at::Tensor> &, const ::std::optional<at::Tensor> &, ::std::optional<at::ScalarType>, bool);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_cslt_sparse_mm_search")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_cslt_sparse_mm_search(Tensor compressed_A, Tensor dense_B, Tensor? bias=None, Tensor? alpha=None, ScalarType? out_dtype=None, bool transpose_result=False) -> int")
+  static int64_t call(const at::Tensor & compressed_A, const at::Tensor & dense_B, const ::std::optional<at::Tensor> & bias, const ::std::optional<at::Tensor> & alpha, ::std::optional<at::ScalarType> out_dtype, bool transpose_result);
+  static int64_t redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & compressed_A, const at::Tensor & dense_B, const ::std::optional<at::Tensor> & bias, const ::std::optional<at::Tensor> & alpha, ::std::optional<at::ScalarType> out_dtype, bool transpose_result);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_embedding_bag_backward_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 _embedding_bag_backward_symint(const at::Tensor & grad, const at::Tensor & indices, const at::Tensor & offsets, const at::Tensor & offset2bag, const at::Tensor & bag_size, const at::Tensor & maximum_indices, c10::SymInt num_weights, bool scale_grad_by_freq, int64_t mode, bool sparse, const ::std::optional<at::Tensor> & per_sample_weights, int64_t padding_idx=-1);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_fake_quantize_learnable_per_channel_affine_backward_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::tuple<at::Tensor,at::Tensor,at::Tensor> _fake_quantize_learnable_per_channel_affine_backward(const at::Tensor & grad, const at::Tensor & self, const at::Tensor & scale, const at::Tensor & zero_point, int64_t axis, 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/_flash_attention_forward_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 _flash_attention_forward {
+  using schema = ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,at::Tensor> (const at::Tensor &, const at::Tensor &, const at::Tensor &, const ::std::optional<at::Tensor> &, const ::std::optional<at::Tensor> &, c10::SymInt, c10::SymInt, double, bool, bool, ::std::optional<double>, ::std::optional<c10::SymInt>, ::std::optional<c10::SymInt>, const ::std::optional<at::Tensor> &, const ::std::optional<at::Tensor> &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_flash_attention_forward")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_flash_attention_forward(Tensor query, Tensor key, Tensor value, Tensor? cum_seq_q, Tensor? cum_seq_k, SymInt max_q, SymInt max_k, float dropout_p, bool is_causal, bool return_debug_mask, *, float? scale=None, SymInt? window_size_left=None, SymInt? window_size_right=None, Tensor? seqused_k=None, Tensor? alibi_slopes=None) -> (Tensor output, Tensor softmax_logsumexp, Tensor philox_seed, Tensor philox_offset, Tensor debug_attn_mask)")
+  static ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,at::Tensor> call(const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const ::std::optional<at::Tensor> & cum_seq_q, const ::std::optional<at::Tensor> & cum_seq_k, c10::SymInt max_q, c10::SymInt max_k, double dropout_p, bool is_causal, bool return_debug_mask, ::std::optional<double> scale, ::std::optional<c10::SymInt> window_size_left, ::std::optional<c10::SymInt> window_size_right, const ::std::optional<at::Tensor> & seqused_k, const ::std::optional<at::Tensor> & alibi_slopes);
+  static ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const ::std::optional<at::Tensor> & cum_seq_q, const ::std::optional<at::Tensor> & cum_seq_k, c10::SymInt max_q, c10::SymInt max_k, double dropout_p, bool is_causal, bool return_debug_mask, ::std::optional<double> scale, ::std::optional<c10::SymInt> window_size_left, ::std::optional<c10::SymInt> window_size_right, const ::std::optional<at::Tensor> & seqused_k, const ::std::optional<at::Tensor> & alibi_slopes);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_foreach_clamp_min_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_clamp_min(at::TensorList self, const at::Scalar & scalar);
+TORCH_API void _foreach_clamp_min_out(at::TensorList out, at::TensorList self, const at::Scalar & scalar);
+TORCH_API void _foreach_clamp_min_outf(at::TensorList self, const at::Scalar & scalar, at::TensorList out);
+TORCH_API void _foreach_clamp_min_(at::TensorList self, const at::Scalar & scalar);
+TORCH_API ::std::vector<at::Tensor> _foreach_clamp_min(at::TensorList self, at::TensorList other);
+TORCH_API void _foreach_clamp_min_out(at::TensorList out, at::TensorList self, at::TensorList other);
+TORCH_API void _foreach_clamp_min_outf(at::TensorList self, at::TensorList other, at::TensorList out);
+TORCH_API void _foreach_clamp_min_(at::TensorList self, at::TensorList other);
+TORCH_API ::std::vector<at::Tensor> _foreach_clamp_min(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API void _foreach_clamp_min_out(at::TensorList out, at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API void _foreach_clamp_min_outf(at::TensorList self, at::ArrayRef<at::Scalar> scalars, at::TensorList out);
+TORCH_API void _foreach_clamp_min_(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+
+} // namespace compositeexplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_foreach_div_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_div_scalar_kernel_slow(at::TensorList self, const at::Scalar & scalar);
+TORCH_API void _foreach_div_Scalar_out(at::TensorList self, const at::Scalar & scalar, at::TensorList out);
+TORCH_API void foreach_tensor_div_scalar_kernel_slow_(at::TensorList self, const at::Scalar & scalar);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_div_scalar_kernel_cuda(at::TensorList self, const at::Scalar & scalar);
+TORCH_API void foreach_tensor_div_scalar_kernel_cuda_(at::TensorList self, const at::Scalar & scalar);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_div_list_kernel_slow(at::TensorList self, at::TensorList other);
+TORCH_API void _foreach_div_List_out(at::TensorList self, at::TensorList other, at::TensorList out);
+TORCH_API void foreach_tensor_div_list_kernel_slow_(at::TensorList self, at::TensorList other);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_div_list_kernel_cuda(at::TensorList self, at::TensorList other);
+TORCH_API void foreach_tensor_div_list_kernel_cuda_(at::TensorList self, at::TensorList other);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_div_scalarlist_kernel_slow(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API void _foreach_div_ScalarList_out(at::TensorList self, at::ArrayRef<at::Scalar> scalars, at::TensorList out);
+TORCH_API void foreach_tensor_div_scalarlist_kernel_slow_(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_div_scalarlist_kernel_cuda(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API void foreach_tensor_div_scalarlist_kernel_cuda_(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_div_tensor_kernel_slow(at::TensorList self, const at::Tensor & other);
+TORCH_API void _foreach_div_Tensor_out(at::TensorList self, const at::Tensor & other, at::TensorList out);
+TORCH_API void foreach_tensor_div_tensor_kernel_slow_(at::TensorList self, const at::Tensor & other);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_div_tensor_kernel_cuda(at::TensorList self, const at::Tensor & other);
+TORCH_API void foreach_tensor_div_tensor_kernel_cuda_(at::TensorList self, const at::Tensor & other);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_foreach_expm1_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_expm1_slow(at::TensorList self);
+TORCH_API void _foreach_expm1_out(at::TensorList self, at::TensorList out);
+TORCH_API void foreach_tensor_expm1_slow_(at::TensorList self);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_expm1_cuda(at::TensorList self);
+TORCH_API void foreach_tensor_expm1_cuda_(at::TensorList self);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_foreach_log10_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_log10_slow(at::TensorList self);
+TORCH_API void _foreach_log10_out(at::TensorList self, at::TensorList out);
+TORCH_API void foreach_tensor_log10_slow_(at::TensorList self);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_log10_cuda(at::TensorList self);
+TORCH_API void foreach_tensor_log10_cuda_(at::TensorList self);
+} // namespace native
+} // namespace at

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

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_linalg_solve_ex_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 ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor> _linalg_solve_ex(const at::Tensor & A, const at::Tensor & B, bool left=true, bool check_errors=false);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> _linalg_solve_ex_out(at::Tensor & result, at::Tensor & LU, at::Tensor & pivots, at::Tensor & info, const at::Tensor & A, const at::Tensor & B, bool left=true, bool check_errors=false);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> _linalg_solve_ex_outf(const at::Tensor & A, const at::Tensor & B, bool left, bool check_errors, at::Tensor & result, at::Tensor & LU, at::Tensor & pivots, at::Tensor & info);
+
+} // namespace cuda
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_native_batch_norm_legit_cpu_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 cpu {
+
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> _native_batch_norm_legit_out(at::Tensor & out, at::Tensor & save_mean, at::Tensor & save_invstd, const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & bias, at::Tensor & running_mean, at::Tensor & running_var, bool training, double momentum, double eps);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> _native_batch_norm_legit_outf(const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & bias, at::Tensor & running_mean, at::Tensor & running_var, bool training, double momentum, double eps, at::Tensor & out, at::Tensor & save_mean, at::Tensor & save_invstd);
+TORCH_API ::std::tuple<at::Tensor,at::Tensor,at::Tensor> _native_batch_norm_legit(const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & bias, at::Tensor & running_mean, at::Tensor & running_var, bool training, double momentum, double eps);
+TORCH_API ::std::tuple<at::Tensor,at::Tensor,at::Tensor> _native_batch_norm_legit(const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & bias, bool training, double momentum, double eps);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> _native_batch_norm_legit_out(at::Tensor & out, at::Tensor & save_mean, at::Tensor & save_invstd, const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & bias, bool training, double momentum, double eps);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> _native_batch_norm_legit_outf(const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & bias, bool training, double momentum, double eps, at::Tensor & out, at::Tensor & save_mean, at::Tensor & save_invstd);
+
+} // namespace cpu
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_nnpack_available_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 _nnpack_available();
+
+} // namespace compositeimplicitautograd
+} // namespace at

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

@@ -0,0 +1,26 @@
+#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/_nnz_ops.h>
+
+namespace at {
+
+
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_reshape_copy_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 _reshape_copy(const at::Tensor & self, at::IntArrayRef size);
+TORCH_API at::Tensor _reshape_copy_symint(const at::Tensor & self, c10::SymIntArrayRef size);
+
+} // namespace compositeexplicitautograd
+} // namespace at

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

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_sparse_addmm.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/_sparse_addmm_ops.h>
+
+namespace at {
+
+
+// aten::_sparse_addmm(Tensor self, Tensor mat1, Tensor mat2, *, Scalar beta=1, Scalar alpha=1) -> Tensor
+inline at::Tensor _sparse_addmm(const at::Tensor & self, const at::Tensor & mat1, const at::Tensor & mat2, const at::Scalar & beta=1, const at::Scalar & alpha=1) {
+    return at::_ops::_sparse_addmm::call(self, mat1, mat2, beta, alpha);
+}
+
+// aten::_sparse_addmm.out(Tensor self, Tensor mat1, Tensor mat2, *, Scalar beta=1, Scalar alpha=1, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & _sparse_addmm_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & mat1, const at::Tensor & mat2, const at::Scalar & beta=1, const at::Scalar & alpha=1) {
+    return at::_ops::_sparse_addmm_out::call(self, mat1, mat2, beta, alpha, out);
+}
+// aten::_sparse_addmm.out(Tensor self, Tensor mat1, Tensor mat2, *, Scalar beta=1, Scalar alpha=1, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & _sparse_addmm_outf(const at::Tensor & self, const at::Tensor & mat1, const at::Tensor & mat2, const at::Scalar & beta, const at::Scalar & alpha, at::Tensor & out) {
+    return at::_ops::_sparse_addmm_out::call(self, mat1, mat2, beta, alpha, out);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_sparse_coo_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_coo_tensor_with_dims_out(at::Tensor & out, int64_t sparse_dim, int64_t dense_dim, at::IntArrayRef size);
+TORCH_API at::Tensor & _sparse_coo_tensor_with_dims_outf(int64_t sparse_dim, int64_t dense_dim, at::IntArrayRef size, at::Tensor & out);
+
+} // namespace compositeexplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_sparse_log_softmax_backward_data.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/_sparse_log_softmax_backward_data_ops.h>
+
+namespace at {
+
+
+// aten::_sparse_log_softmax_backward_data(Tensor grad_output, Tensor output, int dim, Tensor self) -> Tensor
+inline at::Tensor _sparse_log_softmax_backward_data(const at::Tensor & grad_output, const at::Tensor & output, int64_t dim, const at::Tensor & self) {
+    return at::_ops::_sparse_log_softmax_backward_data::call(grad_output, output, dim, self);
+}
+
+// aten::_sparse_log_softmax_backward_data.out(Tensor grad_output, Tensor output, int dim, Tensor self, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & _sparse_log_softmax_backward_data_out(at::Tensor & out, const at::Tensor & grad_output, const at::Tensor & output, int64_t dim, const at::Tensor & self) {
+    return at::_ops::_sparse_log_softmax_backward_data_out::call(grad_output, output, dim, self, out);
+}
+// aten::_sparse_log_softmax_backward_data.out(Tensor grad_output, Tensor output, int dim, Tensor self, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & _sparse_log_softmax_backward_data_outf(const at::Tensor & grad_output, const at::Tensor & output, int64_t dim, const at::Tensor & self, at::Tensor & out) {
+    return at::_ops::_sparse_log_softmax_backward_data_out::call(grad_output, output, dim, self, out);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_to_dense_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 _to_dense {
+  using schema = at::Tensor (const at::Tensor &, ::std::optional<at::ScalarType>, ::std::optional<bool>);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_to_dense")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_to_dense(Tensor self, ScalarType? dtype=None, bool? masked_grad=None) -> Tensor")
+  static at::Tensor call(const at::Tensor & self, ::std::optional<at::ScalarType> dtype, ::std::optional<bool> masked_grad);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, ::std::optional<at::ScalarType> dtype, ::std::optional<bool> masked_grad);
+};
+
+struct TORCH_API _to_dense_out {
+  using schema = at::Tensor & (const at::Tensor &, ::std::optional<at::ScalarType>, ::std::optional<bool>, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_to_dense")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_to_dense.out(Tensor self, ScalarType? dtype=None, bool? masked_grad=None, *, Tensor(a!) out) -> Tensor(a!)")
+  static at::Tensor & call(const at::Tensor & self, ::std::optional<at::ScalarType> dtype, ::std::optional<bool> masked_grad, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, ::std::optional<at::ScalarType> dtype, ::std::optional<bool> masked_grad, at::Tensor & out);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_transformer_encoder_layer_fwd_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 _transformer_encoder_layer_fwd(const at::Tensor & src, int64_t embed_dim, int64_t num_heads, const at::Tensor & qkv_weight, const at::Tensor & qkv_bias, const at::Tensor & proj_weight, const at::Tensor & proj_bias, bool use_gelu, bool norm_first, double eps, const at::Tensor & norm_weight_1, const at::Tensor & norm_bias_1, const at::Tensor & norm_weight_2, const at::Tensor & norm_bias_2, const at::Tensor & ffn_weight_1, const at::Tensor & ffn_bias_1, const at::Tensor & ffn_weight_2, const at::Tensor & ffn_bias_2, const ::std::optional<at::Tensor> & mask={}, ::std::optional<int64_t> mask_type=::std::nullopt);
+
+} // namespace cuda
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_upsample_nearest_exact3d_backward.h

@@ -0,0 +1,91 @@
+#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/_upsample_nearest_exact3d_backward_ops.h>
+
+namespace at {
+
+
+// aten::_upsample_nearest_exact3d_backward.grad_input(Tensor grad_output, SymInt[3] output_size, SymInt[5] input_size, float? scales_d=None, float? scales_h=None, float? scales_w=None, *, Tensor(a!) grad_input) -> Tensor(a!)
+inline at::Tensor & _upsample_nearest_exact3d_backward_out(at::Tensor & grad_input, const at::Tensor & grad_output, at::IntArrayRef output_size, at::IntArrayRef input_size, ::std::optional<double> scales_d=::std::nullopt, ::std::optional<double> scales_h=::std::nullopt, ::std::optional<double> scales_w=::std::nullopt) {
+    return at::_ops::_upsample_nearest_exact3d_backward_grad_input::call(grad_output, c10::fromIntArrayRefSlow(output_size), c10::fromIntArrayRefSlow(input_size), scales_d, scales_h, scales_w, grad_input);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
+  at::Tensor & _upsample_nearest_exact3d_backward_out(at::Tensor & grad_input, const at::Tensor & grad_output, at::IntArrayRef output_size, at::IntArrayRef input_size, ::std::optional<double> scales_d=::std::nullopt, ::std::optional<double> scales_h=::std::nullopt, ::std::optional<double> scales_w=::std::nullopt) {
+    return at::_ops::_upsample_nearest_exact3d_backward_grad_input::call(grad_output, c10::fromIntArrayRefSlow(output_size), c10::fromIntArrayRefSlow(input_size), scales_d, scales_h, scales_w, grad_input);
+  }
+}
+
+// aten::_upsample_nearest_exact3d_backward.grad_input(Tensor grad_output, SymInt[3] output_size, SymInt[5] input_size, float? scales_d=None, float? scales_h=None, float? scales_w=None, *, Tensor(a!) grad_input) -> Tensor(a!)
+inline at::Tensor & _upsample_nearest_exact3d_backward_outf(const at::Tensor & grad_output, at::IntArrayRef output_size, at::IntArrayRef input_size, ::std::optional<double> scales_d, ::std::optional<double> scales_h, ::std::optional<double> scales_w, at::Tensor & grad_input) {
+    return at::_ops::_upsample_nearest_exact3d_backward_grad_input::call(grad_output, c10::fromIntArrayRefSlow(output_size), c10::fromIntArrayRefSlow(input_size), scales_d, scales_h, scales_w, grad_input);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
+  at::Tensor & _upsample_nearest_exact3d_backward_outf(const at::Tensor & grad_output, at::IntArrayRef output_size, at::IntArrayRef input_size, ::std::optional<double> scales_d, ::std::optional<double> scales_h, ::std::optional<double> scales_w, at::Tensor & grad_input) {
+    return at::_ops::_upsample_nearest_exact3d_backward_grad_input::call(grad_output, c10::fromIntArrayRefSlow(output_size), c10::fromIntArrayRefSlow(input_size), scales_d, scales_h, scales_w, grad_input);
+  }
+}
+
+// aten::_upsample_nearest_exact3d_backward.grad_input(Tensor grad_output, SymInt[3] output_size, SymInt[5] input_size, float? scales_d=None, float? scales_h=None, float? scales_w=None, *, Tensor(a!) grad_input) -> Tensor(a!)
+inline at::Tensor & _upsample_nearest_exact3d_backward_symint_out(at::Tensor & grad_input, const at::Tensor & grad_output, c10::SymIntArrayRef output_size, c10::SymIntArrayRef input_size, ::std::optional<double> scales_d=::std::nullopt, ::std::optional<double> scales_h=::std::nullopt, ::std::optional<double> scales_w=::std::nullopt) {
+    return at::_ops::_upsample_nearest_exact3d_backward_grad_input::call(grad_output, output_size, input_size, scales_d, scales_h, scales_w, grad_input);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
+  at::Tensor & _upsample_nearest_exact3d_backward_out(at::Tensor & grad_input, const at::Tensor & grad_output, c10::SymIntArrayRef output_size, c10::SymIntArrayRef input_size, ::std::optional<double> scales_d=::std::nullopt, ::std::optional<double> scales_h=::std::nullopt, ::std::optional<double> scales_w=::std::nullopt) {
+    return at::_ops::_upsample_nearest_exact3d_backward_grad_input::call(grad_output, output_size, input_size, scales_d, scales_h, scales_w, grad_input);
+  }
+}
+
+// aten::_upsample_nearest_exact3d_backward.grad_input(Tensor grad_output, SymInt[3] output_size, SymInt[5] input_size, float? scales_d=None, float? scales_h=None, float? scales_w=None, *, Tensor(a!) grad_input) -> Tensor(a!)
+inline at::Tensor & _upsample_nearest_exact3d_backward_symint_outf(const at::Tensor & grad_output, c10::SymIntArrayRef output_size, c10::SymIntArrayRef input_size, ::std::optional<double> scales_d, ::std::optional<double> scales_h, ::std::optional<double> scales_w, at::Tensor & grad_input) {
+    return at::_ops::_upsample_nearest_exact3d_backward_grad_input::call(grad_output, output_size, input_size, scales_d, scales_h, scales_w, grad_input);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
+  at::Tensor & _upsample_nearest_exact3d_backward_outf(const at::Tensor & grad_output, c10::SymIntArrayRef output_size, c10::SymIntArrayRef input_size, ::std::optional<double> scales_d, ::std::optional<double> scales_h, ::std::optional<double> scales_w, at::Tensor & grad_input) {
+    return at::_ops::_upsample_nearest_exact3d_backward_grad_input::call(grad_output, output_size, input_size, scales_d, scales_h, scales_w, grad_input);
+  }
+}
+
+// aten::_upsample_nearest_exact3d_backward(Tensor grad_output, SymInt[3] output_size, SymInt[5] input_size, float? scales_d=None, float? scales_h=None, float? scales_w=None) -> Tensor
+inline at::Tensor _upsample_nearest_exact3d_backward(const at::Tensor & grad_output, at::IntArrayRef output_size, at::IntArrayRef input_size, ::std::optional<double> scales_d=::std::nullopt, ::std::optional<double> scales_h=::std::nullopt, ::std::optional<double> scales_w=::std::nullopt) {
+    return at::_ops::_upsample_nearest_exact3d_backward::call(grad_output, c10::fromIntArrayRefSlow(output_size), c10::fromIntArrayRefSlow(input_size), scales_d, scales_h, scales_w);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
+  at::Tensor _upsample_nearest_exact3d_backward(const at::Tensor & grad_output, at::IntArrayRef output_size, at::IntArrayRef input_size, ::std::optional<double> scales_d=::std::nullopt, ::std::optional<double> scales_h=::std::nullopt, ::std::optional<double> scales_w=::std::nullopt) {
+    return at::_ops::_upsample_nearest_exact3d_backward::call(grad_output, c10::fromIntArrayRefSlow(output_size), c10::fromIntArrayRefSlow(input_size), scales_d, scales_h, scales_w);
+  }
+}
+
+// aten::_upsample_nearest_exact3d_backward(Tensor grad_output, SymInt[3] output_size, SymInt[5] input_size, float? scales_d=None, float? scales_h=None, float? scales_w=None) -> Tensor
+inline at::Tensor _upsample_nearest_exact3d_backward_symint(const at::Tensor & grad_output, c10::SymIntArrayRef output_size, c10::SymIntArrayRef input_size, ::std::optional<double> scales_d=::std::nullopt, ::std::optional<double> scales_h=::std::nullopt, ::std::optional<double> scales_w=::std::nullopt) {
+    return at::_ops::_upsample_nearest_exact3d_backward::call(grad_output, output_size, input_size, scales_d, scales_h, scales_w);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
+  at::Tensor _upsample_nearest_exact3d_backward(const at::Tensor & grad_output, c10::SymIntArrayRef output_size, c10::SymIntArrayRef input_size, ::std::optional<double> scales_d=::std::nullopt, ::std::optional<double> scales_h=::std::nullopt, ::std::optional<double> scales_w=::std::nullopt) {
+    return at::_ops::_upsample_nearest_exact3d_backward::call(grad_output, output_size, input_size, scales_d, scales_h, scales_w);
+  }
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_upsample_nearest_exact3d_backward_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/_upsample_nearest_exact3d_backward_meta.h>
+
+namespace at {
+namespace native {
+struct TORCH_API structured__upsample_nearest_exact3d_backward_out_cpu : public at::meta::structured__upsample_nearest_exact3d_backward {
+void impl(const at::Tensor & grad_output, at::ArrayRef<int64_t> output_size, at::ArrayRef<int64_t> input_size, ::std::optional<double> scales_d, ::std::optional<double> scales_h, ::std::optional<double> scales_w, const at::Tensor & grad_input);
+};
+struct TORCH_API structured__upsample_nearest_exact3d_backward_out_cuda : public at::meta::structured__upsample_nearest_exact3d_backward {
+void impl(const at::Tensor & grad_output, at::ArrayRef<int64_t> output_size, at::ArrayRef<int64_t> input_size, ::std::optional<double> scales_d, ::std::optional<double> scales_h, ::std::optional<double> scales_w, const at::Tensor & grad_input);
+};
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_weight_norm_interface_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 &> _weight_norm_interface_out(at::Tensor & out0, at::Tensor & out1, const at::Tensor & v, const at::Tensor & g, int64_t dim=0);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &> _weight_norm_interface_outf(const at::Tensor & v, const at::Tensor & g, int64_t dim, at::Tensor & out0, at::Tensor & out1);
+
+} // namespace compositeexplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/addbmm_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 addbmm(const at::Tensor & self, const at::Tensor & batch1, const at::Tensor & batch2, const at::Scalar & beta=1, const at::Scalar & alpha=1);
+TORCH_API at::Tensor & addbmm_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & batch1, const at::Tensor & batch2, const at::Scalar & beta=1, const at::Scalar & alpha=1);
+TORCH_API at::Tensor & addbmm_outf(const at::Tensor & self, const at::Tensor & batch1, const at::Tensor & batch2, const at::Scalar & beta, const at::Scalar & alpha, at::Tensor & out);
+TORCH_API at::Tensor & addbmm_(at::Tensor & self, const at::Tensor & batch1, const at::Tensor & batch2, const at::Scalar & beta=1, const at::Scalar & alpha=1);
+
+} // namespace cuda
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/addmv.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/addmv_ops.h>
+
+namespace at {
+
+
+// aten::addmv(Tensor self, Tensor mat, Tensor vec, *, Scalar beta=1, Scalar alpha=1) -> Tensor
+inline at::Tensor addmv(const at::Tensor & self, const at::Tensor & mat, const at::Tensor & vec, const at::Scalar & beta=1, const at::Scalar & alpha=1) {
+    return at::_ops::addmv::call(self, mat, vec, beta, alpha);
+}
+
+// aten::addmv_(Tensor(a!) self, Tensor mat, Tensor vec, *, Scalar beta=1, Scalar alpha=1) -> Tensor(a!)
+inline at::Tensor & addmv_(at::Tensor & self, const at::Tensor & mat, const at::Tensor & vec, const at::Scalar & beta=1, const at::Scalar & alpha=1) {
+    return at::_ops::addmv_::call(self, mat, vec, beta, alpha);
+}
+
+// aten::addmv.out(Tensor self, Tensor mat, Tensor vec, *, Scalar beta=1, Scalar alpha=1, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & addmv_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & mat, const at::Tensor & vec, const at::Scalar & beta=1, const at::Scalar & alpha=1) {
+    return at::_ops::addmv_out::call(self, mat, vec, beta, alpha, out);
+}
+// aten::addmv.out(Tensor self, Tensor mat, Tensor vec, *, Scalar beta=1, Scalar alpha=1, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & addmv_outf(const at::Tensor & self, const at::Tensor & mat, const at::Tensor & vec, const at::Scalar & beta, const at::Scalar & alpha, at::Tensor & out) {
+    return at::_ops::addmv_out::call(self, mat, vec, beta, alpha, out);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/alias_copy_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 & alias_copy_out(at::Tensor & out, const at::Tensor & self);
+TORCH_API at::Tensor & alias_copy_outf(const at::Tensor & self, at::Tensor & out);
+
+} // namespace compositeexplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/amin_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_amin : public at::impl::MetaBase {
+    
+    
+    void meta(const at::Tensor & self, at::IntArrayRef dim, bool keepdim);
+};
+
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/as_strided_copy_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 as_strided_copy {
+  using schema = at::Tensor (const at::Tensor &, c10::SymIntArrayRef, c10::SymIntArrayRef, ::std::optional<c10::SymInt>);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::as_strided_copy")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "as_strided_copy(Tensor self, SymInt[] size, SymInt[] stride, SymInt? storage_offset=None) -> Tensor")
+  static at::Tensor call(const at::Tensor & self, c10::SymIntArrayRef size, c10::SymIntArrayRef stride, ::std::optional<c10::SymInt> storage_offset);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, c10::SymIntArrayRef size, c10::SymIntArrayRef stride, ::std::optional<c10::SymInt> storage_offset);
+};
+
+struct TORCH_API as_strided_copy_out {
+  using schema = at::Tensor & (const at::Tensor &, c10::SymIntArrayRef, c10::SymIntArrayRef, ::std::optional<c10::SymInt>, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::as_strided_copy")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "as_strided_copy.out(Tensor self, SymInt[] size, SymInt[] stride, SymInt? storage_offset=None, *, Tensor(a!) out) -> Tensor(a!)")
+  static at::Tensor & call(const at::Tensor & self, c10::SymIntArrayRef size, c10::SymIntArrayRef stride, ::std::optional<c10::SymInt> storage_offset, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, c10::SymIntArrayRef size, c10::SymIntArrayRef stride, ::std::optional<c10::SymInt> storage_offset, at::Tensor & out);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/atan.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/atan_ops.h>
+
+namespace at {
+
+
+// aten::atan(Tensor self) -> Tensor
+inline at::Tensor atan(const at::Tensor & self) {
+    return at::_ops::atan::call(self);
+}
+
+// aten::atan_(Tensor(a!) self) -> Tensor(a!)
+inline at::Tensor & atan_(at::Tensor & self) {
+    return at::_ops::atan_::call(self);
+}
+
+// aten::atan.out(Tensor self, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & atan_out(at::Tensor & out, const at::Tensor & self) {
+    return at::_ops::atan_out::call(self, out);
+}
+// aten::atan.out(Tensor self, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & atan_outf(const at::Tensor & self, at::Tensor & out) {
+    return at::_ops::atan_out::call(self, out);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/atan2_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_atan2 : public TensorIteratorBase {
+    
+    
+    void meta(const at::Tensor & self, const at::Tensor & other);
+};
+
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/binomial.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/binomial_ops.h>
+
+namespace at {
+
+
+// aten::binomial(Tensor count, Tensor prob, Generator? generator=None) -> Tensor
+inline at::Tensor binomial(const at::Tensor & count, const at::Tensor & prob, ::std::optional<at::Generator> generator=::std::nullopt) {
+    return at::_ops::binomial::call(count, prob, generator);
+}
+
+// aten::binomial.out(Tensor count, Tensor prob, Generator? generator=None, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & binomial_out(at::Tensor & out, const at::Tensor & count, const at::Tensor & prob, ::std::optional<at::Generator> generator=::std::nullopt) {
+    return at::_ops::binomial_out::call(count, prob, generator, out);
+}
+// aten::binomial.out(Tensor count, Tensor prob, Generator? generator=None, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & binomial_outf(const at::Tensor & count, const at::Tensor & prob, ::std::optional<at::Generator> generator, at::Tensor & out) {
+    return at::_ops::binomial_out::call(count, prob, generator, out);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/binomial_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 binomial {
+  using schema = at::Tensor (const at::Tensor &, const at::Tensor &, ::std::optional<at::Generator>);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::binomial")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "binomial(Tensor count, Tensor prob, Generator? generator=None) -> Tensor")
+  static at::Tensor call(const at::Tensor & count, const at::Tensor & prob, ::std::optional<at::Generator> generator);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & count, const at::Tensor & prob, ::std::optional<at::Generator> generator);
+};
+
+struct TORCH_API binomial_out {
+  using schema = at::Tensor & (const at::Tensor &, const at::Tensor &, ::std::optional<at::Generator>, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::binomial")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "binomial.out(Tensor count, Tensor prob, Generator? generator=None, *, Tensor(a!) out) -> Tensor(a!)")
+  static at::Tensor & call(const at::Tensor & count, const at::Tensor & prob, ::std::optional<at::Generator> generator, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & count, const at::Tensor & prob, ::std::optional<at::Generator> generator, at::Tensor & out);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/bitwise_or_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 bitwise_or(const at::Tensor & self, const at::Tensor & other);
+TORCH_API at::Tensor & bitwise_or_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & other);
+TORCH_API at::Tensor & bitwise_or_outf(const at::Tensor & self, const at::Tensor & other, at::Tensor & out);
+TORCH_API at::Tensor & bitwise_or_(at::Tensor & self, const at::Tensor & other);
+
+} // namespace cpu
+} // namespace at

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

@@ -0,0 +1,61 @@
+#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 blackman_window {
+  using schema = at::Tensor (int64_t, ::std::optional<at::ScalarType>, ::std::optional<at::Layout>, ::std::optional<at::Device>, ::std::optional<bool>);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::blackman_window")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "blackman_window(int window_length, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor")
+  static at::Tensor call(int64_t window_length, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, int64_t window_length, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory);
+};
+
+struct TORCH_API blackman_window_periodic {
+  using schema = at::Tensor (int64_t, bool, ::std::optional<at::ScalarType>, ::std::optional<at::Layout>, ::std::optional<at::Device>, ::std::optional<bool>);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::blackman_window")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "periodic")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "blackman_window.periodic(int window_length, bool periodic, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor")
+  static at::Tensor call(int64_t window_length, bool periodic, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, int64_t window_length, bool periodic, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory);
+};
+
+struct TORCH_API blackman_window_out {
+  using schema = at::Tensor & (int64_t, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::blackman_window")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "blackman_window.out(int window_length, *, Tensor(a!) out) -> Tensor(a!)")
+  static at::Tensor & call(int64_t window_length, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, int64_t window_length, at::Tensor & out);
+};
+
+struct TORCH_API blackman_window_periodic_out {
+  using schema = at::Tensor & (int64_t, bool, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::blackman_window")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "periodic_out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "blackman_window.periodic_out(int window_length, bool periodic, *, Tensor(a!) out) -> Tensor(a!)")
+  static at::Tensor & call(int64_t window_length, bool periodic, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, int64_t window_length, bool periodic, at::Tensor & out);
+};
+
+}} // namespace at::_ops

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

@@ -0,0 +1,53 @@
+#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/bucketize_ops.h>
+
+namespace at {
+
+
+// aten::bucketize.Tensor(Tensor self, Tensor boundaries, *, bool out_int32=False, bool right=False) -> Tensor
+inline at::Tensor bucketize(const at::Tensor & self, const at::Tensor & boundaries, bool out_int32=false, bool right=false) {
+    return at::_ops::bucketize_Tensor::call(self, boundaries, out_int32, right);
+}
+
+// aten::bucketize.Tensor_out(Tensor self, Tensor boundaries, *, bool out_int32=False, bool right=False, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & bucketize_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & boundaries, bool out_int32=false, bool right=false) {
+    return at::_ops::bucketize_Tensor_out::call(self, boundaries, out_int32, right, out);
+}
+// aten::bucketize.Tensor_out(Tensor self, Tensor boundaries, *, bool out_int32=False, bool right=False, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & bucketize_outf(const at::Tensor & self, const at::Tensor & boundaries, bool out_int32, bool right, at::Tensor & out) {
+    return at::_ops::bucketize_Tensor_out::call(self, boundaries, out_int32, right, out);
+}
+
+// aten::bucketize.Scalar(Scalar self, Tensor boundaries, *, bool out_int32=False, bool right=False) -> Tensor
+inline at::Tensor bucketize(const at::Scalar & self, const at::Tensor & boundaries, bool out_int32=false, bool right=false) {
+    return at::_ops::bucketize_Scalar::call(self, boundaries, out_int32, right);
+}
+
+// aten::bucketize.Scalar_out(Scalar self, Tensor boundaries, *, bool out_int32=False, bool right=False, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & bucketize_out(at::Tensor & out, const at::Scalar & self, const at::Tensor & boundaries, bool out_int32=false, bool right=false) {
+    return at::_ops::bucketize_Scalar_out::call(self, boundaries, out_int32, right, out);
+}
+// aten::bucketize.Scalar_out(Scalar self, Tensor boundaries, *, bool out_int32=False, bool right=False, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & bucketize_outf(const at::Scalar & self, const at::Tensor & boundaries, bool out_int32, bool right, at::Tensor & out) {
+    return at::_ops::bucketize_Scalar_out::call(self, boundaries, out_int32, right, out);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/chunk_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 chunk {
+  using schema = ::std::vector<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::chunk")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "chunk(Tensor(a -> *) self, int chunks, int dim=0) -> Tensor(a)[]")
+  static ::std::vector<at::Tensor> call(const at::Tensor & self, int64_t chunks, int64_t dim);
+  static ::std::vector<at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, int64_t chunks, int64_t dim);
+};
+
+}} // namespace at::_ops

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

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

@@ -0,0 +1,30 @@
+#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 div(const at::Tensor & self, const at::Tensor & other);
+TORCH_API at::Tensor & div_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & other);
+TORCH_API at::Tensor & div_outf(const at::Tensor & self, const at::Tensor & other, at::Tensor & out);
+TORCH_API at::Tensor & div_(at::Tensor & self, const at::Tensor & other);
+TORCH_API at::Tensor div(const at::Tensor & self, const at::Tensor & other, ::std::optional<c10::string_view> rounding_mode);
+TORCH_API at::Tensor & div_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & other, ::std::optional<c10::string_view> rounding_mode);
+TORCH_API at::Tensor & div_outf(const at::Tensor & self, const at::Tensor & other, ::std::optional<c10::string_view> rounding_mode, at::Tensor & out);
+TORCH_API at::Tensor & div_(at::Tensor & self, const at::Tensor & other, ::std::optional<c10::string_view> rounding_mode);
+
+} // namespace cpu
+} // namespace at

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

@@ -0,0 +1,127 @@
+#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 divide_Tensor {
+  using schema = at::Tensor (const 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::divide")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "Tensor")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "divide.Tensor(Tensor self, Tensor other) -> Tensor")
+  static at::Tensor call(const at::Tensor & self, const at::Tensor & other);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Tensor & other);
+};
+
+struct TORCH_API divide__Tensor {
+  using schema = 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::divide_")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "Tensor")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "divide_.Tensor(Tensor(a!) self, Tensor other) -> Tensor(a!)")
+  static at::Tensor & call(at::Tensor & self, const at::Tensor & other);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, at::Tensor & self, const at::Tensor & other);
+};
+
+struct TORCH_API divide_out {
+  using schema = at::Tensor & (const 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::divide")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "divide.out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)")
+  static at::Tensor & call(const at::Tensor & self, const at::Tensor & other, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Tensor & other, at::Tensor & out);
+};
+
+struct TORCH_API divide_Scalar {
+  using schema = at::Tensor (const at::Tensor &, const at::Scalar &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::divide")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "Scalar")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "divide.Scalar(Tensor self, Scalar other) -> Tensor")
+  static at::Tensor call(const at::Tensor & self, const at::Scalar & other);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Scalar & other);
+};
+
+struct TORCH_API divide__Scalar {
+  using schema = at::Tensor & (at::Tensor &, const at::Scalar &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::divide_")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "Scalar")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "divide_.Scalar(Tensor(a!) self, Scalar other) -> Tensor(a!)")
+  static at::Tensor & call(at::Tensor & self, const at::Scalar & other);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, at::Tensor & self, const at::Scalar & other);
+};
+
+struct TORCH_API divide_Tensor_mode {
+  using schema = at::Tensor (const at::Tensor &, const at::Tensor &, ::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::divide")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "Tensor_mode")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "divide.Tensor_mode(Tensor self, Tensor other, *, str? rounding_mode) -> Tensor")
+  static at::Tensor call(const at::Tensor & self, const at::Tensor & other, ::std::optional<c10::string_view> rounding_mode);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Tensor & other, ::std::optional<c10::string_view> rounding_mode);
+};
+
+struct TORCH_API divide__Tensor_mode {
+  using schema = at::Tensor & (at::Tensor &, const at::Tensor &, ::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::divide_")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "Tensor_mode")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "divide_.Tensor_mode(Tensor(a!) self, Tensor other, *, str? rounding_mode) -> Tensor(a!)")
+  static at::Tensor & call(at::Tensor & self, const at::Tensor & other, ::std::optional<c10::string_view> rounding_mode);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, at::Tensor & self, const at::Tensor & other, ::std::optional<c10::string_view> rounding_mode);
+};
+
+struct TORCH_API divide_out_mode {
+  using schema = at::Tensor & (const at::Tensor &, const at::Tensor &, ::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::divide")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out_mode")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "divide.out_mode(Tensor self, Tensor other, *, str? rounding_mode, Tensor(a!) out) -> Tensor(a!)")
+  static at::Tensor & call(const at::Tensor & self, const at::Tensor & other, ::std::optional<c10::string_view> rounding_mode, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Tensor & other, ::std::optional<c10::string_view> rounding_mode, at::Tensor & out);
+};
+
+struct TORCH_API divide_Scalar_mode {
+  using schema = at::Tensor (const at::Tensor &, const at::Scalar &, ::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::divide")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "Scalar_mode")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "divide.Scalar_mode(Tensor self, Scalar other, *, str? rounding_mode) -> Tensor")
+  static at::Tensor call(const at::Tensor & self, const at::Scalar & other, ::std::optional<c10::string_view> rounding_mode);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Scalar & other, ::std::optional<c10::string_view> rounding_mode);
+};
+
+struct TORCH_API divide__Scalar_mode {
+  using schema = at::Tensor & (at::Tensor &, const at::Scalar &, ::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::divide_")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "Scalar_mode")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "divide_.Scalar_mode(Tensor(a!) self, Scalar other, *, str? rounding_mode) -> Tensor(a!)")
+  static at::Tensor & call(at::Tensor & self, const at::Scalar & other, ::std::optional<c10::string_view> rounding_mode);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, at::Tensor & self, const at::Scalar & other, ::std::optional<c10::string_view> rounding_mode);
+};
+
+}} // namespace at::_ops

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

@@ -0,0 +1,30 @@
+#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 empty_permuted(at::IntArrayRef size, at::IntArrayRef physical_layout, at::TensorOptions options={});
+TORCH_API at::Tensor empty_permuted(at::IntArrayRef size, at::IntArrayRef physical_layout, ::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_permuted_symint(c10::SymIntArrayRef size, at::IntArrayRef physical_layout, at::TensorOptions options={});
+TORCH_API at::Tensor empty_permuted_symint(c10::SymIntArrayRef size, at::IntArrayRef physical_layout, ::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_permuted_out(at::Tensor & out, at::IntArrayRef size, at::IntArrayRef physical_layout);
+TORCH_API at::Tensor & empty_permuted_outf(at::IntArrayRef size, at::IntArrayRef physical_layout, at::Tensor & out);
+TORCH_API at::Tensor & empty_permuted_symint_out(at::Tensor & out, c10::SymIntArrayRef size, at::IntArrayRef physical_layout);
+TORCH_API at::Tensor & empty_permuted_symint_outf(c10::SymIntArrayRef size, at::IntArrayRef physical_layout, at::Tensor & out);
+
+} // namespace compositeexplicitautograd
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/empty_permuted_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 empty_permuted_symint(c10::SymIntArrayRef size, at::IntArrayRef physical_layout, ::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_permuted_out_symint(c10::SymIntArrayRef size, at::IntArrayRef physical_layout, at::Tensor & out);
+} // namespace native
+} // namespace at

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

@@ -0,0 +1,30 @@
+#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 & eye_out(at::Tensor & out, int64_t n);
+TORCH_API at::Tensor & eye_outf(int64_t n, at::Tensor & out);
+TORCH_API at::Tensor & eye_symint_out(at::Tensor & out, c10::SymInt n);
+TORCH_API at::Tensor & eye_symint_outf(c10::SymInt n, at::Tensor & out);
+TORCH_API at::Tensor & eye_out(at::Tensor & out, int64_t n, int64_t m);
+TORCH_API at::Tensor & eye_outf(int64_t n, int64_t m, at::Tensor & out);
+TORCH_API at::Tensor & eye_symint_out(at::Tensor & out, c10::SymInt n, c10::SymInt m);
+TORCH_API at::Tensor & eye_symint_outf(c10::SymInt n, c10::SymInt m, at::Tensor & out);
+
+} // namespace cpu
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/feature_alpha_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_alpha_dropout_ops.h>
+
+namespace at {
+
+
+// aten::feature_alpha_dropout(Tensor input, float p, bool train) -> Tensor
+inline at::Tensor feature_alpha_dropout(const at::Tensor & input, double p, bool train) {
+    return at::_ops::feature_alpha_dropout::call(input, p, train);
+}
+
+// aten::feature_alpha_dropout_(Tensor(a!) self, float p, bool train) -> Tensor(a!)
+inline at::Tensor & feature_alpha_dropout_(at::Tensor & self, double p, bool train) {
+    return at::_ops::feature_alpha_dropout_::call(self, p, train);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/gru_cell_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 gru_cell {
+  using schema = at::Tensor (const at::Tensor &, const at::Tensor &, const at::Tensor &, const at::Tensor &, const ::std::optional<at::Tensor> &, const ::std::optional<at::Tensor> &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::gru_cell")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "gru_cell(Tensor input, Tensor hx, Tensor w_ih, Tensor w_hh, Tensor? b_ih=None, Tensor? b_hh=None) -> Tensor")
+  static at::Tensor call(const at::Tensor & input, const at::Tensor & hx, const at::Tensor & w_ih, const at::Tensor & w_hh, const ::std::optional<at::Tensor> & b_ih, const ::std::optional<at::Tensor> & b_hh);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & input, const at::Tensor & hx, const at::Tensor & w_ih, const at::Tensor & w_hh, const ::std::optional<at::Tensor> & b_ih, const ::std::optional<at::Tensor> & b_hh);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/histogramdd_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 histogramdd {
+  using schema = ::std::tuple<at::Tensor,::std::vector<at::Tensor>> (const at::Tensor &, at::IntArrayRef, ::std::optional<at::ArrayRef<double>>, const ::std::optional<at::Tensor> &, bool);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::histogramdd")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "histogramdd(Tensor self, int[] bins, float[]? range=None, Tensor? weight=None, bool density=False) -> (Tensor hist, Tensor[] bin_edges)")
+  static ::std::tuple<at::Tensor,::std::vector<at::Tensor>> call(const at::Tensor & self, at::IntArrayRef bins, ::std::optional<at::ArrayRef<double>> range, const ::std::optional<at::Tensor> & weight, bool density);
+  static ::std::tuple<at::Tensor,::std::vector<at::Tensor>> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, at::IntArrayRef bins, ::std::optional<at::ArrayRef<double>> range, const ::std::optional<at::Tensor> & weight, bool density);
+};
+
+struct TORCH_API histogramdd_int_bins {
+  using schema = ::std::tuple<at::Tensor,::std::vector<at::Tensor>> (const at::Tensor &, int64_t, ::std::optional<at::ArrayRef<double>>, const ::std::optional<at::Tensor> &, bool);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::histogramdd")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "int_bins")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "histogramdd.int_bins(Tensor self, int bins, float[]? range=None, Tensor? weight=None, bool density=False) -> (Tensor hist, Tensor[] bin_edges)")
+  static ::std::tuple<at::Tensor,::std::vector<at::Tensor>> call(const at::Tensor & self, int64_t bins, ::std::optional<at::ArrayRef<double>> range, const ::std::optional<at::Tensor> & weight, bool density);
+  static ::std::tuple<at::Tensor,::std::vector<at::Tensor>> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, int64_t bins, ::std::optional<at::ArrayRef<double>> range, const ::std::optional<at::Tensor> & weight, bool density);
+};
+
+struct TORCH_API histogramdd_TensorList_bins {
+  using schema = ::std::tuple<at::Tensor,::std::vector<at::Tensor>> (const at::Tensor &, at::TensorList, ::std::optional<at::ArrayRef<double>>, const ::std::optional<at::Tensor> &, bool);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::histogramdd")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "TensorList_bins")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "histogramdd.TensorList_bins(Tensor self, Tensor[] bins, float[]? range=None, Tensor? weight=None, bool density=False) -> (Tensor hist, Tensor[] bin_edges)")
+  static ::std::tuple<at::Tensor,::std::vector<at::Tensor>> call(const at::Tensor & self, at::TensorList bins, ::std::optional<at::ArrayRef<double>> range, const ::std::optional<at::Tensor> & weight, bool density);
+  static ::std::tuple<at::Tensor,::std::vector<at::Tensor>> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, at::TensorList bins, ::std::optional<at::ArrayRef<double>> range, const ::std::optional<at::Tensor> & weight, bool density);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/index_copy_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 index_copy(const at::Tensor & self, int64_t dim, const at::Tensor & index, const at::Tensor & source);
+TORCH_API at::Tensor & index_copy_(at::Tensor & self, int64_t dim, const at::Tensor & index, const at::Tensor & source);
+
+} // namespace compositeexplicitautogradnonfunctional
+} // namespace at

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

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/is_same_size_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 bool is_same_size(const at::Tensor & self, const at::Tensor & other);
+
+} // namespace compositeexplicitautograd
+} // namespace at