Add files using upload-large-folder tool

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

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

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_amp_foreach_non_finite_check_and_unscale_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 _amp_foreach_non_finite_check_and_unscale_ {
+  using schema = void (at::TensorList, 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::_amp_foreach_non_finite_check_and_unscale_")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_amp_foreach_non_finite_check_and_unscale_(Tensor(a!)[] self, Tensor(b!) found_inf, Tensor inv_scale) -> ()")
+  static void call(at::TensorList self, at::Tensor & found_inf, const at::Tensor & inv_scale);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::Tensor & found_inf, const at::Tensor & inv_scale);
+};
+
+struct TORCH_API _amp_foreach_non_finite_check_and_unscale_out {
+  using schema = void (at::TensorList, at::Tensor &, const at::Tensor &, at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_amp_foreach_non_finite_check_and_unscale")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_amp_foreach_non_finite_check_and_unscale.out(Tensor[] self, Tensor(b!) found_inf, Tensor inv_scale, *, Tensor(a!)[] out) -> ()")
+  static void call(at::TensorList self, at::Tensor & found_inf, const at::Tensor & inv_scale, at::TensorList out);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::Tensor & found_inf, const at::Tensor & inv_scale, at::TensorList out);
+};
+
+struct TORCH_API _amp_foreach_non_finite_check_and_unscale {
+  using schema = ::std::tuple<::std::vector<at::Tensor>,at::Tensor> (at::TensorList, 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::_amp_foreach_non_finite_check_and_unscale")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_amp_foreach_non_finite_check_and_unscale(Tensor[] self, Tensor found_inf, Tensor inv_scale) -> (Tensor[] self_out, Tensor found_inf_out)")
+  static ::std::tuple<::std::vector<at::Tensor>,at::Tensor> call(at::TensorList self, const at::Tensor & found_inf, const at::Tensor & inv_scale);
+  static ::std::tuple<::std::vector<at::Tensor>,at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, const at::Tensor & found_inf, const at::Tensor & inv_scale);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_ctc_loss_meta_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 meta {
+
+TORCH_API ::std::tuple<at::Tensor,at::Tensor> _ctc_loss(const at::Tensor & log_probs, const at::Tensor & targets, at::IntArrayRef input_lengths, at::IntArrayRef target_lengths, int64_t blank=0, bool zero_infinity=false);
+
+} // namespace meta
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_fft_c2r_cuda_dispatch.h

@@ -0,0 +1,28 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API at::Tensor _fft_c2r(const at::Tensor & self, at::IntArrayRef dim, int64_t normalization, int64_t last_dim_size);
+TORCH_API at::Tensor _fft_c2r_symint(const at::Tensor & self, at::IntArrayRef dim, int64_t normalization, c10::SymInt last_dim_size);
+TORCH_API at::Tensor & _fft_c2r_out(at::Tensor & out, const at::Tensor & self, at::IntArrayRef dim, int64_t normalization, int64_t last_dim_size);
+TORCH_API at::Tensor & _fft_c2r_outf(const at::Tensor & self, at::IntArrayRef dim, int64_t normalization, int64_t last_dim_size, at::Tensor & out);
+TORCH_API at::Tensor & _fft_c2r_symint_out(at::Tensor & out, const at::Tensor & self, at::IntArrayRef dim, int64_t normalization, c10::SymInt last_dim_size);
+TORCH_API at::Tensor & _fft_c2r_symint_outf(const at::Tensor & self, at::IntArrayRef dim, int64_t normalization, c10::SymInt last_dim_size, at::Tensor & out);
+
+} // namespace cuda
+} // namespace at

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

@@ -0,0 +1,47 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_flash_attention_forward_ops.h>
+
+namespace at {
+
+
+// aten::_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)
+inline ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,at::Tensor> _flash_attention_forward(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, int64_t max_q, int64_t max_k, double dropout_p, bool is_causal, bool return_debug_mask, ::std::optional<double> scale=::std::nullopt, ::std::optional<int64_t> window_size_left=::std::nullopt, ::std::optional<int64_t> window_size_right=::std::nullopt, const ::std::optional<at::Tensor> & seqused_k={}, const ::std::optional<at::Tensor> & alibi_slopes={}) {
+    return at::_ops::_flash_attention_forward::call(query, key, value, cum_seq_q, cum_seq_k, max_q, max_k, dropout_p, is_causal, return_debug_mask, scale, window_size_left.has_value() ? ::std::make_optional(c10::SymInt(*window_size_left)) : ::std::nullopt, window_size_right.has_value() ? ::std::make_optional(c10::SymInt(*window_size_right)) : ::std::nullopt, seqused_k, alibi_slopes);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
+  ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,at::Tensor> _flash_attention_forward(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, int64_t max_q, int64_t max_k, double dropout_p, bool is_causal, bool return_debug_mask, ::std::optional<double> scale=::std::nullopt, ::std::optional<int64_t> window_size_left=::std::nullopt, ::std::optional<int64_t> window_size_right=::std::nullopt, const ::std::optional<at::Tensor> & seqused_k={}, const ::std::optional<at::Tensor> & alibi_slopes={}) {
+    return at::_ops::_flash_attention_forward::call(query, key, value, cum_seq_q, cum_seq_k, max_q, max_k, dropout_p, is_causal, return_debug_mask, scale, window_size_left.has_value() ? ::std::make_optional(c10::SymInt(*window_size_left)) : ::std::nullopt, window_size_right.has_value() ? ::std::make_optional(c10::SymInt(*window_size_right)) : ::std::nullopt, seqused_k, alibi_slopes);
+  }
+}
+
+// aten::_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)
+inline ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,at::Tensor> _flash_attention_forward_symint(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::nullopt, ::std::optional<c10::SymInt> window_size_left=::std::nullopt, ::std::optional<c10::SymInt> window_size_right=::std::nullopt, const ::std::optional<at::Tensor> & seqused_k={}, const ::std::optional<at::Tensor> & alibi_slopes={}) {
+    return at::_ops::_flash_attention_forward::call(query, key, value, cum_seq_q, cum_seq_k, max_q, max_k, dropout_p, is_causal, return_debug_mask, scale, window_size_left, window_size_right, seqused_k, alibi_slopes);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
+  ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,at::Tensor> _flash_attention_forward(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::nullopt, ::std::optional<c10::SymInt> window_size_left=::std::nullopt, ::std::optional<c10::SymInt> window_size_right=::std::nullopt, const ::std::optional<at::Tensor> & seqused_k={}, const ::std::optional<at::Tensor> & alibi_slopes={}) {
+    return at::_ops::_flash_attention_forward::call(query, key, value, cum_seq_q, cum_seq_k, max_q, max_k, dropout_p, is_causal, return_debug_mask, scale, window_size_left, window_size_right, seqused_k, alibi_slopes);
+  }
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_foreach_log1p_cuda_dispatch.h

@@ -0,0 +1,24 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_log1p(at::TensorList self);
+TORCH_API void _foreach_log1p_(at::TensorList self);
+
+} // namespace cuda
+} // namespace at

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

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_functional_assert_scalar_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 _functional_assert_scalar(const at::Scalar & self, c10::string_view assert_msg, const at::Tensor & dep_token);
+} // namespace native
+} // namespace at

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

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

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_int_mm_cpu_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 cpu {
+
+TORCH_API at::Tensor _int_mm(const at::Tensor & self, const at::Tensor & mat2);
+TORCH_API at::Tensor & _int_mm_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & mat2);
+TORCH_API at::Tensor & _int_mm_outf(const at::Tensor & self, const at::Tensor & mat2, at::Tensor & out);
+
+} // namespace cpu
+} // namespace at

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

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

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

@@ -0,0 +1,39 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _linalg_svd {
+  using schema = ::std::tuple<at::Tensor,at::Tensor,at::Tensor> (const at::Tensor &, bool, bool, ::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::_linalg_svd")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_linalg_svd(Tensor A, bool full_matrices=False, bool compute_uv=True, *, str? driver=None) -> (Tensor U, Tensor S, Tensor Vh)")
+  static ::std::tuple<at::Tensor,at::Tensor,at::Tensor> call(const at::Tensor & A, bool full_matrices, bool compute_uv, ::std::optional<c10::string_view> driver);
+  static ::std::tuple<at::Tensor,at::Tensor,at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & A, bool full_matrices, bool compute_uv, ::std::optional<c10::string_view> driver);
+};
+
+struct TORCH_API _linalg_svd_U {
+  using schema = ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> (const at::Tensor &, bool, bool, ::std::optional<c10::string_view>, at::Tensor &, at::Tensor &, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_linalg_svd")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "U")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_linalg_svd.U(Tensor A, bool full_matrices=False, bool compute_uv=True, *, str? driver=None, Tensor(a!) U, Tensor(b!) S, Tensor(c!) Vh) -> (Tensor(a!) U, Tensor(b!) S, Tensor(c!) Vh)")
+  static ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> call(const at::Tensor & A, bool full_matrices, bool compute_uv, ::std::optional<c10::string_view> driver, at::Tensor & U, at::Tensor & S, at::Tensor & Vh);
+  static ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & A, bool full_matrices, bool compute_uv, ::std::optional<c10::string_view> driver, at::Tensor & U, at::Tensor & S, at::Tensor & Vh);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_nested_from_padded.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/_nested_from_padded_ops.h>
+
+namespace at {
+
+
+// aten::_nested_from_padded(Tensor padded, Tensor cpu_nested_shape_example, bool fuse_transform_0213=False) -> Tensor
+inline at::Tensor _nested_from_padded(const at::Tensor & padded, const at::Tensor & cpu_nested_shape_example, bool fuse_transform_0213=false) {
+    return at::_ops::_nested_from_padded::call(padded, cpu_nested_shape_example, fuse_transform_0213);
+}
+
+// aten::_nested_from_padded.out(Tensor padded, Tensor cpu_nested_shape_example, bool fuse_transform_0213=False, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & _nested_from_padded_out(at::Tensor & out, const at::Tensor & padded, const at::Tensor & cpu_nested_shape_example, bool fuse_transform_0213=false) {
+    return at::_ops::_nested_from_padded_out::call(padded, cpu_nested_shape_example, fuse_transform_0213, out);
+}
+// aten::_nested_from_padded.out(Tensor padded, Tensor cpu_nested_shape_example, bool fuse_transform_0213=False, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & _nested_from_padded_outf(const at::Tensor & padded, const at::Tensor & cpu_nested_shape_example, bool fuse_transform_0213, at::Tensor & out) {
+    return at::_ops::_nested_from_padded_out::call(padded, cpu_nested_shape_example, fuse_transform_0213, out);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_nnpack_spatial_convolution_compositeexplicitautograd_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 compositeexplicitautograd {
+
+TORCH_API at::Tensor _nnpack_spatial_convolution(const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, at::IntArrayRef padding, at::IntArrayRef stride=1);
+TORCH_API at::Tensor _nnpack_spatial_convolution_symint(const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef padding, c10::SymIntArrayRef stride=c10::SymInt(1));
+TORCH_API at::Tensor & _nnpack_spatial_convolution_out(at::Tensor & out, const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, at::IntArrayRef padding, at::IntArrayRef stride=1);
+TORCH_API at::Tensor & _nnpack_spatial_convolution_outf(const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, at::IntArrayRef padding, at::IntArrayRef stride, at::Tensor & out);
+TORCH_API at::Tensor & _nnpack_spatial_convolution_symint_out(at::Tensor & out, const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef padding, c10::SymIntArrayRef stride=c10::SymInt(1));
+TORCH_API at::Tensor & _nnpack_spatial_convolution_symint_outf(const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef padding, c10::SymIntArrayRef stride, at::Tensor & out);
+
+} // namespace compositeexplicitautograd
+} // namespace at

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

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

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_spdiags.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/_spdiags_ops.h>
+
+namespace at {
+
+
+// aten::_spdiags(Tensor diagonals, Tensor offsets, int[] shape, Layout? layout=None) -> Tensor
+inline at::Tensor _spdiags(const at::Tensor & diagonals, const at::Tensor & offsets, at::IntArrayRef shape, ::std::optional<at::Layout> layout=::std::nullopt) {
+    return at::_ops::_spdiags::call(diagonals, offsets, shape, layout);
+}
+
+// aten::_spdiags.out(Tensor diagonals, Tensor offsets, int[] shape, Layout? layout=None, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & _spdiags_out(at::Tensor & out, const at::Tensor & diagonals, const at::Tensor & offsets, at::IntArrayRef shape, ::std::optional<at::Layout> layout=::std::nullopt) {
+    return at::_ops::_spdiags_out::call(diagonals, offsets, shape, layout, out);
+}
+// aten::_spdiags.out(Tensor diagonals, Tensor offsets, int[] shape, Layout? layout=None, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & _spdiags_outf(const at::Tensor & diagonals, const at::Tensor & offsets, at::IntArrayRef shape, ::std::optional<at::Layout> layout, at::Tensor & out) {
+    return at::_ops::_spdiags_out::call(diagonals, offsets, shape, layout, out);
+}
+
+}

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

@@ -0,0 +1,39 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _to_sparse_csr {
+  using schema = at::Tensor (const at::Tensor &, ::std::optional<int64_t>);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_to_sparse_csr")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_to_sparse_csr(Tensor self, int? dense_dim=None) -> Tensor")
+  static at::Tensor call(const at::Tensor & self, ::std::optional<int64_t> dense_dim);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, ::std::optional<int64_t> dense_dim);
+};
+
+struct TORCH_API _to_sparse_csr_out {
+  using schema = at::Tensor & (const at::Tensor &, ::std::optional<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::_to_sparse_csr")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_to_sparse_csr.out(Tensor self, int? dense_dim=None, *, Tensor(a!) out) -> Tensor(a!)")
+  static at::Tensor & call(const at::Tensor & self, ::std::optional<int64_t> dense_dim, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, ::std::optional<int64_t> dense_dim, at::Tensor & out);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_transform_bias_rescale_qkv.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/_transform_bias_rescale_qkv_ops.h>
+
+namespace at {
+
+
+// aten::_transform_bias_rescale_qkv(Tensor qkv, Tensor qkv_bias, int num_heads) -> (Tensor, Tensor, Tensor)
+inline ::std::tuple<at::Tensor,at::Tensor,at::Tensor> _transform_bias_rescale_qkv(const at::Tensor & qkv, const at::Tensor & qkv_bias, int64_t num_heads) {
+    return at::_ops::_transform_bias_rescale_qkv::call(qkv, qkv_bias, num_heads);
+}
+
+// aten::_transform_bias_rescale_qkv.out(Tensor qkv, Tensor qkv_bias, int num_heads, *, Tensor(a!) out0, Tensor(b!) out1, Tensor(c!) out2) -> (Tensor(a!), Tensor(b!), Tensor(c!))
+inline ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> _transform_bias_rescale_qkv_out(at::Tensor & out0, at::Tensor & out1, at::Tensor & out2, const at::Tensor & qkv, const at::Tensor & qkv_bias, int64_t num_heads) {
+    return at::_ops::_transform_bias_rescale_qkv_out::call(qkv, qkv_bias, num_heads, out0, out1, out2);
+}
+// aten::_transform_bias_rescale_qkv.out(Tensor qkv, Tensor qkv_bias, int num_heads, *, Tensor(a!) out0, Tensor(b!) out1, Tensor(c!) out2) -> (Tensor(a!), Tensor(b!), Tensor(c!))
+inline ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> _transform_bias_rescale_qkv_outf(const at::Tensor & qkv, const at::Tensor & qkv_bias, int64_t num_heads, at::Tensor & out0, at::Tensor & out1, at::Tensor & out2) {
+    return at::_ops::_transform_bias_rescale_qkv_out::call(qkv, qkv_bias, num_heads, out0, out1, out2);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_upsample_bilinear2d_aa_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 at::Tensor _upsample_bilinear2d_aa(const at::Tensor & self, at::IntArrayRef output_size, bool align_corners, ::std::optional<double> scales_h=::std::nullopt, ::std::optional<double> scales_w=::std::nullopt);
+TORCH_API at::Tensor _upsample_bilinear2d_aa_symint(const at::Tensor & self, c10::SymIntArrayRef output_size, bool align_corners, ::std::optional<double> scales_h=::std::nullopt, ::std::optional<double> scales_w=::std::nullopt);
+TORCH_API at::Tensor & _upsample_bilinear2d_aa_out(at::Tensor & out, const at::Tensor & self, at::IntArrayRef output_size, bool align_corners, ::std::optional<double> scales_h=::std::nullopt, ::std::optional<double> scales_w=::std::nullopt);
+TORCH_API at::Tensor & _upsample_bilinear2d_aa_outf(const at::Tensor & self, at::IntArrayRef output_size, bool align_corners, ::std::optional<double> scales_h, ::std::optional<double> scales_w, at::Tensor & out);
+TORCH_API at::Tensor & _upsample_bilinear2d_aa_symint_out(at::Tensor & out, const at::Tensor & self, c10::SymIntArrayRef output_size, bool align_corners, ::std::optional<double> scales_h=::std::nullopt, ::std::optional<double> scales_w=::std::nullopt);
+TORCH_API at::Tensor & _upsample_bilinear2d_aa_symint_outf(const at::Tensor & self, c10::SymIntArrayRef output_size, bool align_corners, ::std::optional<double> scales_h, ::std::optional<double> scales_w, at::Tensor & out);
+
+} // namespace cpu
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/acos_meta_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 meta {
+
+TORCH_API at::Tensor acos(const at::Tensor & self);
+TORCH_API at::Tensor & acos_out(at::Tensor & out, const at::Tensor & self);
+TORCH_API at::Tensor & acos_outf(const at::Tensor & self, at::Tensor & out);
+TORCH_API at::Tensor & acos_(at::Tensor & self);
+
+} // namespace meta
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/aminmax.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/aminmax_ops.h>
+
+namespace at {
+
+
+// aten::aminmax(Tensor self, *, int? dim=None, bool keepdim=False) -> (Tensor min, Tensor max)
+inline ::std::tuple<at::Tensor,at::Tensor> aminmax(const at::Tensor & self, ::std::optional<int64_t> dim=::std::nullopt, bool keepdim=false) {
+    return at::_ops::aminmax::call(self, dim, keepdim);
+}
+
+// aten::aminmax.out(Tensor self, *, int? dim=None, bool keepdim=False, Tensor(a!) min, Tensor(b!) max) -> (Tensor(a!) min, Tensor(b!) max)
+inline ::std::tuple<at::Tensor &,at::Tensor &> aminmax_out(at::Tensor & min, at::Tensor & max, const at::Tensor & self, ::std::optional<int64_t> dim=::std::nullopt, bool keepdim=false) {
+    return at::_ops::aminmax_out::call(self, dim, keepdim, min, max);
+}
+// aten::aminmax.out(Tensor self, *, int? dim=None, bool keepdim=False, Tensor(a!) min, Tensor(b!) max) -> (Tensor(a!) min, Tensor(b!) max)
+inline ::std::tuple<at::Tensor &,at::Tensor &> aminmax_outf(const at::Tensor & self, ::std::optional<int64_t> dim, bool keepdim, at::Tensor & min, at::Tensor & max) {
+    return at::_ops::aminmax_out::call(self, dim, keepdim, min, max);
+}
+
+}

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

@@ -0,0 +1,31 @@
+#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 any(const at::Tensor & self, int64_t dim, bool keepdim=false);
+TORCH_API at::Tensor & any_out(at::Tensor & out, const at::Tensor & self, int64_t dim, bool keepdim=false);
+TORCH_API at::Tensor & any_outf(const at::Tensor & self, int64_t dim, bool keepdim, at::Tensor & out);
+TORCH_API at::Tensor any(const at::Tensor & self, at::OptionalIntArrayRef dim, bool keepdim=false);
+TORCH_API at::Tensor & any_out(at::Tensor & out, const at::Tensor & self, at::OptionalIntArrayRef dim, bool keepdim=false);
+TORCH_API at::Tensor & any_outf(const at::Tensor & self, at::OptionalIntArrayRef dim, bool keepdim, at::Tensor & out);
+TORCH_API at::Tensor any(const at::Tensor & self);
+TORCH_API at::Tensor & any_out(at::Tensor & out, const at::Tensor & self);
+TORCH_API at::Tensor & any_outf(const at::Tensor & self, at::Tensor & out);
+
+} // namespace meta
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/asin_meta_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 meta {
+
+TORCH_API at::Tensor asin(const at::Tensor & self);
+TORCH_API at::Tensor & asin_out(at::Tensor & out, const at::Tensor & self);
+TORCH_API at::Tensor & asin_outf(const at::Tensor & self, at::Tensor & out);
+TORCH_API at::Tensor & asin_(at::Tensor & self);
+
+} // namespace meta
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/atan_meta_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 meta {
+
+TORCH_API at::Tensor atan(const at::Tensor & self);
+TORCH_API at::Tensor & atan_out(at::Tensor & out, const at::Tensor & self);
+TORCH_API at::Tensor & atan_outf(const at::Tensor & self, at::Tensor & out);
+TORCH_API at::Tensor & atan_(at::Tensor & self);
+
+} // namespace meta
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/batch_norm_backward_reduce_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 batch_norm_backward_reduce {
+  using schema = ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor> (const at::Tensor &, const at::Tensor &, const at::Tensor &, const at::Tensor &, const ::std::optional<at::Tensor> &, bool, bool, bool);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::batch_norm_backward_reduce")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "batch_norm_backward_reduce(Tensor grad_out, Tensor input, Tensor mean, Tensor invstd, Tensor? weight, bool input_g, bool weight_g, bool bias_g) -> (Tensor, Tensor, Tensor, Tensor)")
+  static ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor> call(const at::Tensor & grad_out, const at::Tensor & input, const at::Tensor & mean, const at::Tensor & invstd, const ::std::optional<at::Tensor> & weight, bool input_g, bool weight_g, bool bias_g);
+  static ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & grad_out, const at::Tensor & input, const at::Tensor & mean, const at::Tensor & invstd, const ::std::optional<at::Tensor> & weight, bool input_g, bool weight_g, bool bias_g);
+};
+
+struct TORCH_API batch_norm_backward_reduce_out {
+  using schema = ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> (const at::Tensor &, const at::Tensor &, const at::Tensor &, const at::Tensor &, const ::std::optional<at::Tensor> &, bool, bool, bool, at::Tensor &, at::Tensor &, at::Tensor &, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::batch_norm_backward_reduce")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "batch_norm_backward_reduce.out(Tensor grad_out, Tensor input, Tensor mean, Tensor invstd, Tensor? weight, bool input_g, bool weight_g, bool bias_g, *, Tensor(a!) out0, Tensor(b!) out1, Tensor(c!) out2, Tensor(d!) out3) -> (Tensor(a!), Tensor(b!), Tensor(c!), Tensor(d!))")
+  static ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> call(const at::Tensor & grad_out, const at::Tensor & input, const at::Tensor & mean, const at::Tensor & invstd, const ::std::optional<at::Tensor> & weight, bool input_g, bool weight_g, bool bias_g, at::Tensor & out0, at::Tensor & out1, at::Tensor & out2, at::Tensor & out3);
+  static ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & grad_out, const at::Tensor & input, const at::Tensor & mean, const at::Tensor & invstd, const ::std::optional<at::Tensor> & weight, bool input_g, bool weight_g, bool bias_g, at::Tensor & out0, at::Tensor & out1, at::Tensor & out2, at::Tensor & out3);
+};
+
+}} // namespace at::_ops

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/batch_norm_gather_stats.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/batch_norm_gather_stats_ops.h>
+
+namespace at {
+
+
+// aten::batch_norm_gather_stats(Tensor input, Tensor mean, Tensor invstd, Tensor? running_mean, Tensor? running_var, float momentum, float eps, int count) -> (Tensor, Tensor)
+inline ::std::tuple<at::Tensor,at::Tensor> batch_norm_gather_stats(const at::Tensor & input, const at::Tensor & mean, const at::Tensor & invstd, const ::std::optional<at::Tensor> & running_mean, const ::std::optional<at::Tensor> & running_var, double momentum, double eps, int64_t count) {
+    return at::_ops::batch_norm_gather_stats::call(input, mean, invstd, running_mean, running_var, momentum, eps, count);
+}
+
+// aten::batch_norm_gather_stats.out(Tensor input, Tensor mean, Tensor invstd, Tensor? running_mean, Tensor? running_var, float momentum, float eps, int count, *, Tensor(a!) out0, Tensor(b!) out1) -> (Tensor(a!), Tensor(b!))
+inline ::std::tuple<at::Tensor &,at::Tensor &> batch_norm_gather_stats_out(at::Tensor & out0, at::Tensor & out1, const at::Tensor & input, const at::Tensor & mean, const at::Tensor & invstd, const ::std::optional<at::Tensor> & running_mean, const ::std::optional<at::Tensor> & running_var, double momentum, double eps, int64_t count) {
+    return at::_ops::batch_norm_gather_stats_out::call(input, mean, invstd, running_mean, running_var, momentum, eps, count, out0, out1);
+}
+// aten::batch_norm_gather_stats.out(Tensor input, Tensor mean, Tensor invstd, Tensor? running_mean, Tensor? running_var, float momentum, float eps, int count, *, Tensor(a!) out0, Tensor(b!) out1) -> (Tensor(a!), Tensor(b!))
+inline ::std::tuple<at::Tensor &,at::Tensor &> batch_norm_gather_stats_outf(const at::Tensor & input, const at::Tensor & mean, const at::Tensor & invstd, const ::std::optional<at::Tensor> & running_mean, const ::std::optional<at::Tensor> & running_var, double momentum, double eps, int64_t count, at::Tensor & out0, at::Tensor & out1) {
+    return at::_ops::batch_norm_gather_stats_out::call(input, mean, invstd, running_mean, running_var, momentum, eps, count, out0, out1);
+}
+
+}

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

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

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/column_stack_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 column_stack(at::TensorList tensors);
+TORCH_API at::Tensor & column_stack_out(at::TensorList tensors, at::Tensor & out);
+} // namespace native
+} // namespace at

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

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeimplicitautograd {
+
+TORCH_API at::Tensor cosine_similarity(const at::Tensor & x1, const at::Tensor & x2, int64_t dim=1, double eps=1e-08);
+
+} // namespace compositeimplicitautograd
+} // namespace at

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

@@ -0,0 +1,30 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/cov_ops.h>
+
+namespace at {
+
+
+// aten::cov(Tensor self, *, int correction=1, Tensor? fweights=None, Tensor? aweights=None) -> Tensor
+inline at::Tensor cov(const at::Tensor & self, int64_t correction=1, const ::std::optional<at::Tensor> & fweights={}, const ::std::optional<at::Tensor> & aweights={}) {
+    return at::_ops::cov::call(self, correction, fweights, aweights);
+}
+
+}

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

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

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/exp_meta_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 meta {
+
+TORCH_API at::Tensor exp(const at::Tensor & self);
+TORCH_API at::Tensor & exp_out(at::Tensor & out, const at::Tensor & self);
+TORCH_API at::Tensor & exp_outf(const at::Tensor & self, at::Tensor & out);
+TORCH_API at::Tensor & exp_(at::Tensor & self);
+
+} // namespace meta
+} // namespace at

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

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/fmax_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 fmax(const at::Tensor & self, const at::Tensor & other);
+TORCH_API at::Tensor & fmax_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & other);
+TORCH_API at::Tensor & fmax_outf(const at::Tensor & self, const at::Tensor & other, at::Tensor & out);
+
+} // namespace meta
+} // namespace at

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

@@ -0,0 +1,28 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+#include <ATen/ops/gelu_backward_meta.h>
+
+namespace at {
+namespace native {
+struct TORCH_API structured_gelu_backward_out_cpu : public at::meta::structured_gelu_backward {
+void impl(const at::Tensor & grad_output, const at::Tensor & self, c10::string_view approximate, const at::Tensor & grad_input);
+};
+struct TORCH_API structured_gelu_backward_out_cuda : public at::meta::structured_gelu_backward {
+void impl(const at::Tensor & grad_output, const at::Tensor & self, c10::string_view approximate, const at::Tensor & grad_input);
+};
+TORCH_API at::Tensor gelu_backwards_nested(const at::Tensor & grad_output, const at::Tensor & self, c10::string_view approximate="none");
+TORCH_API at::Tensor mkldnn_gelu_backward(const at::Tensor & grad_output, const at::Tensor & self, c10::string_view approximate="none");
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/geqrf_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> geqrf(const at::Tensor & self);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &> geqrf_out(at::Tensor & a, at::Tensor & tau, const at::Tensor & self);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &> geqrf_outf(const at::Tensor & self, at::Tensor & a, at::Tensor & tau);
+
+} // namespace cuda
+} // namespace at

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

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

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/gru.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/gru_ops.h>
+
+namespace at {
+
+
+// aten::gru.input(Tensor input, Tensor hx, Tensor[] params, bool has_biases, int num_layers, float dropout, bool train, bool bidirectional, bool batch_first) -> (Tensor, Tensor)
+inline ::std::tuple<at::Tensor,at::Tensor> gru(const at::Tensor & input, const at::Tensor & hx, at::TensorList params, bool has_biases, int64_t num_layers, double dropout, bool train, bool bidirectional, bool batch_first) {
+    return at::_ops::gru_input::call(input, hx, params, has_biases, num_layers, dropout, train, bidirectional, batch_first);
+}
+
+// aten::gru.data(Tensor data, Tensor batch_sizes, Tensor hx, Tensor[] params, bool has_biases, int num_layers, float dropout, bool train, bool bidirectional) -> (Tensor, Tensor)
+inline ::std::tuple<at::Tensor,at::Tensor> gru(const at::Tensor & data, const at::Tensor & batch_sizes, const at::Tensor & hx, at::TensorList params, bool has_biases, int64_t num_layers, double dropout, bool train, bool bidirectional) {
+    return at::_ops::gru_data::call(data, batch_sizes, hx, params, has_biases, num_layers, dropout, train, bidirectional);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/hardshrink_backward_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 hardshrink_backward(const at::Tensor & grad_out, const at::Tensor & self, const at::Scalar & lambd);
+TORCH_API at::Tensor & hardshrink_backward_out(at::Tensor & grad_input, const at::Tensor & grad_out, const at::Tensor & self, const at::Scalar & lambd);
+TORCH_API at::Tensor & hardshrink_backward_outf(const at::Tensor & grad_out, const at::Tensor & self, const at::Scalar & lambd, at::Tensor & grad_input);
+
+} // namespace meta
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/is_set_to_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 bool is_set_to(const at::Tensor & self, const at::Tensor & tensor);
+} // namespace native
+} // namespace at

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

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

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

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

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

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

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_lu.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/linalg_lu_ops.h>
+
+namespace at {
+
+
+// aten::linalg_lu(Tensor A, *, bool pivot=True) -> (Tensor P, Tensor L, Tensor U)
+inline ::std::tuple<at::Tensor,at::Tensor,at::Tensor> linalg_lu(const at::Tensor & A, bool pivot=true) {
+    return at::_ops::linalg_lu::call(A, pivot);
+}
+
+// aten::linalg_lu.out(Tensor A, *, bool pivot=True, Tensor(a!) P, Tensor(b!) L, Tensor(c!) U) -> (Tensor(a!) P, Tensor(b!) L, Tensor(c!) U)
+inline ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> linalg_lu_out(at::Tensor & P, at::Tensor & L, at::Tensor & U, const at::Tensor & A, bool pivot=true) {
+    return at::_ops::linalg_lu_out::call(A, pivot, P, L, U);
+}
+// aten::linalg_lu.out(Tensor A, *, bool pivot=True, Tensor(a!) P, Tensor(b!) L, Tensor(c!) U) -> (Tensor(a!) P, Tensor(b!) L, Tensor(c!) U)
+inline ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> linalg_lu_outf(const at::Tensor & A, bool pivot, at::Tensor & P, at::Tensor & L, at::Tensor & U) {
+    return at::_ops::linalg_lu_out::call(A, pivot, P, L, U);
+}
+
+}

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_matmul_compositeimplicitautograd_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 compositeimplicitautograd {
+
+TORCH_API at::Tensor linalg_matmul(const at::Tensor & self, const at::Tensor & other);
+TORCH_API at::Tensor & linalg_matmul_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & other);
+TORCH_API at::Tensor & linalg_matmul_outf(const at::Tensor & self, const at::Tensor & other, at::Tensor & out);
+
+} // namespace compositeimplicitautograd
+} // namespace at

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

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/lshift_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>
+
+
+namespace at {
+namespace native {
+TORCH_API at::Tensor & __lshift___Scalar_out(const at::Tensor & self, const at::Scalar & other, at::Tensor & out);
+TORCH_API at::Tensor __lshift__(const at::Tensor & self, const at::Scalar & other);
+TORCH_API at::Tensor & __ilshift__(at::Tensor & self, const at::Scalar & other);
+TORCH_API at::Tensor & __lshift___Tensor_out(const at::Tensor & self, const at::Tensor & other, at::Tensor & out);
+TORCH_API at::Tensor __lshift__(const at::Tensor & self, const at::Tensor & other);
+TORCH_API at::Tensor & __ilshift__(at::Tensor & self, const at::Tensor & other);
+} // namespace native
+} // namespace at

.venv/lib/python3.11/site-packages/torch/include/ATen/ops/max_pool2d_backward_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 max_pool2d_backward {
+  using schema = at::Tensor (const at::Tensor &, const at::Tensor &, at::IntArrayRef, at::IntArrayRef, at::IntArrayRef, at::IntArrayRef, bool);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::max_pool2d_backward")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "max_pool2d_backward(Tensor grad_output, Tensor self, int[2] kernel_size, int[2] stride=[], int[2] padding=0, int[2] dilation=1, bool ceil_mode=False) -> Tensor")
+  static at::Tensor call(const at::Tensor & grad_output, const at::Tensor & self, at::IntArrayRef kernel_size, at::IntArrayRef stride, at::IntArrayRef padding, at::IntArrayRef dilation, bool ceil_mode);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & grad_output, const at::Tensor & self, at::IntArrayRef kernel_size, at::IntArrayRef stride, at::IntArrayRef padding, at::IntArrayRef dilation, bool ceil_mode);
+};
+
+struct TORCH_API max_pool2d_backward_out {
+  using schema = at::Tensor & (const at::Tensor &, const at::Tensor &, at::IntArrayRef, at::IntArrayRef, at::IntArrayRef, at::IntArrayRef, 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::max_pool2d_backward")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "max_pool2d_backward.out(Tensor grad_output, Tensor self, int[2] kernel_size, int[2] stride=[], int[2] padding=0, int[2] dilation=1, bool ceil_mode=False, *, Tensor(a!) out) -> Tensor(a!)")
+  static at::Tensor & call(const at::Tensor & grad_output, const at::Tensor & self, at::IntArrayRef kernel_size, at::IntArrayRef stride, at::IntArrayRef padding, at::IntArrayRef dilation, bool ceil_mode, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & grad_output, const at::Tensor & self, at::IntArrayRef kernel_size, at::IntArrayRef stride, at::IntArrayRef padding, at::IntArrayRef dilation, bool ceil_mode, at::Tensor & out);
+};
+
+}} // namespace at::_ops

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

@@ -0,0 +1,23 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+#include <ATen/ops/maximum_meta.h>
+
+namespace at {
+namespace native {
+struct TORCH_API structured_maximum_out : public at::meta::structured_maximum {
+void impl(const at::Tensor & self, const at::Tensor & other, const at::Tensor & out);
+};
+} // namespace native
+} // namespace at