Add files using upload-large-folder tool

.gitattributes

@@ -80,3 +80,4 @@ tuning-competition-baseline/.venv/lib/python3.11/site-packages/nvidia/cudnn/lib/
 tuning-competition-baseline/.venv/lib/python3.11/site-packages/pip/_vendor/distlib/t64-arm.exe filter=lfs diff=lfs merge=lfs -text
 tuning-competition-baseline/.venv/lib/python3.11/site-packages/nvidia/cuda_runtime/lib/libcudart.so.11.0 filter=lfs diff=lfs merge=lfs -text
 tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/_inductor/__pycache__/cudagraph_trees.cpython-311.pyc filter=lfs diff=lfs merge=lfs -text
+tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/nn/parallel/__pycache__/distributed.cpython-311.pyc filter=lfs diff=lfs merge=lfs -text

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_convert_indices_from_coo_to_csr_ops.h

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

tuning-competition-baseline/.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

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_cudnn_rnn_backward_cuda_dispatch.h

@@ -0,0 +1,24 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API ::std::tuple<at::Tensor,at::Tensor,at::Tensor,::std::vector<at::Tensor>> _cudnn_rnn_backward(const at::Tensor & input, at::TensorList weight, int64_t weight_stride0, const at::Tensor & weight_buf, const at::Tensor & hx, const c10::optional<at::Tensor> & cx, const at::Tensor & output, const c10::optional<at::Tensor> & grad_output, const c10::optional<at::Tensor> & grad_hy, const c10::optional<at::Tensor> & grad_cy, int64_t mode, int64_t hidden_size, int64_t proj_size, int64_t num_layers, bool batch_first, double dropout, bool train, bool bidirectional, at::IntArrayRef batch_sizes, const c10::optional<at::Tensor> & dropout_state, const at::Tensor & reserve, ::std::array<bool,4> output_mask);
+TORCH_API ::std::tuple<at::Tensor,at::Tensor,at::Tensor,::std::vector<at::Tensor>> _cudnn_rnn_backward_symint(const at::Tensor & input, at::TensorList weight, int64_t weight_stride0, const at::Tensor & weight_buf, const at::Tensor & hx, const c10::optional<at::Tensor> & cx, const at::Tensor & output, const c10::optional<at::Tensor> & grad_output, const c10::optional<at::Tensor> & grad_hy, const c10::optional<at::Tensor> & grad_cy, int64_t mode, c10::SymInt hidden_size, c10::SymInt proj_size, int64_t num_layers, bool batch_first, double dropout, bool train, bool bidirectional, c10::SymIntArrayRef batch_sizes, const c10::optional<at::Tensor> & dropout_state, const at::Tensor & reserve, ::std::array<bool,4> output_mask);
+
+} // namespace cuda
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_debug_has_internal_overlap_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 int64_t _debug_has_internal_overlap(const at::Tensor & self);
+
+} // namespace compositeimplicitautograd
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_fake_quantize_learnable_per_channel_affine_native.h

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

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_foreach_ceil.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 <c10/util/Optional.h>
+
+
+
+#include <ATen/ops/_foreach_ceil_ops.h>
+
+namespace at {
+
+
+// aten::_foreach_ceil(Tensor[] self) -> Tensor[]
+inline ::std::vector<at::Tensor> _foreach_ceil(at::TensorList self) {
+    return at::_ops::_foreach_ceil::call(self);
+}
+
+// aten::_foreach_ceil_(Tensor(a!)[] self) -> ()
+inline void _foreach_ceil_(at::TensorList self) {
+    return at::_ops::_foreach_ceil_::call(self);
+}
+
+// aten::_foreach_ceil.out(Tensor[] self, *, Tensor(a!)[] out) -> ()
+inline void _foreach_ceil_out(at::TensorList out, at::TensorList self) {
+    return at::_ops::_foreach_ceil_out::call(self, out);
+}
+// aten::_foreach_ceil.out(Tensor[] self, *, Tensor(a!)[] out) -> ()
+inline void _foreach_ceil_outf(at::TensorList self, at::TensorList out) {
+    return at::_ops::_foreach_ceil_out::call(self, out);
+}
+
+}

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_foreach_pow_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 void _foreach_pow_out(at::TensorList out, at::TensorList self, at::TensorList exponent);
+TORCH_API void _foreach_pow_outf(at::TensorList self, at::TensorList exponent, at::TensorList out);
+TORCH_API void _foreach_pow_out(at::TensorList out, at::TensorList self, const at::Scalar & exponent);
+TORCH_API void _foreach_pow_outf(at::TensorList self, const at::Scalar & exponent, at::TensorList out);
+TORCH_API void _foreach_pow_out(at::TensorList out, at::TensorList self, at::ArrayRef<at::Scalar> exponent);
+TORCH_API void _foreach_pow_outf(at::TensorList self, at::ArrayRef<at::Scalar> exponent, at::TensorList out);
+
+} // namespace compositeexplicitautograd
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_fused_adamw_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 void _fused_adamw_(at::TensorList self, at::TensorList grads, at::TensorList exp_avgs, at::TensorList exp_avg_sqs, at::TensorList max_exp_avg_sqs, at::TensorList state_steps, double lr, double beta1, double beta2, double weight_decay, double eps, bool amsgrad, bool maximize, const c10::optional<at::Tensor> & grad_scale={}, const c10::optional<at::Tensor> & found_inf={});
+TORCH_API void _fused_adamw_(at::TensorList self, at::TensorList grads, at::TensorList exp_avgs, at::TensorList exp_avg_sqs, at::TensorList max_exp_avg_sqs, at::TensorList state_steps, const at::Tensor & lr, double beta1, double beta2, double weight_decay, double eps, bool amsgrad, bool maximize, const c10::optional<at::Tensor> & grad_scale={}, const c10::optional<at::Tensor> & found_inf={});
+
+} // namespace cuda
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_fused_adamw_ops.h

@@ -0,0 +1,83 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _fused_adamw_ {
+  using schema = void (at::TensorList, at::TensorList, at::TensorList, at::TensorList, at::TensorList, at::TensorList, double, double, double, double, double, bool, bool, const c10::optional<at::Tensor> &, const c10::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::_fused_adamw_")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_fused_adamw_(Tensor(a!)[] self, Tensor(b!)[] grads, Tensor(c!)[] exp_avgs, Tensor(d!)[] exp_avg_sqs, Tensor(e!)[] max_exp_avg_sqs, Tensor[] state_steps, *, float lr, float beta1, float beta2, float weight_decay, float eps, bool amsgrad, bool maximize, Tensor? grad_scale=None, Tensor? found_inf=None) -> ()")
+  static void call(at::TensorList self, at::TensorList grads, at::TensorList exp_avgs, at::TensorList exp_avg_sqs, at::TensorList max_exp_avg_sqs, at::TensorList state_steps, double lr, double beta1, double beta2, double weight_decay, double eps, bool amsgrad, bool maximize, const c10::optional<at::Tensor> & grad_scale, const c10::optional<at::Tensor> & found_inf);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::TensorList grads, at::TensorList exp_avgs, at::TensorList exp_avg_sqs, at::TensorList max_exp_avg_sqs, at::TensorList state_steps, double lr, double beta1, double beta2, double weight_decay, double eps, bool amsgrad, bool maximize, const c10::optional<at::Tensor> & grad_scale, const c10::optional<at::Tensor> & found_inf);
+};
+
+struct TORCH_API _fused_adamw__tensor_lr {
+  using schema = void (at::TensorList, at::TensorList, at::TensorList, at::TensorList, at::TensorList, at::TensorList, const at::Tensor &, double, double, double, double, bool, bool, const c10::optional<at::Tensor> &, const c10::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::_fused_adamw_")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "tensor_lr")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_fused_adamw_.tensor_lr(Tensor(a!)[] self, Tensor(b!)[] grads, Tensor(c!)[] exp_avgs, Tensor(d!)[] exp_avg_sqs, Tensor(e!)[] max_exp_avg_sqs, Tensor[] state_steps, *, Tensor lr, float beta1, float beta2, float weight_decay, float eps, bool amsgrad, bool maximize, Tensor? grad_scale=None, Tensor? found_inf=None) -> ()")
+  static void call(at::TensorList self, at::TensorList grads, at::TensorList exp_avgs, at::TensorList exp_avg_sqs, at::TensorList max_exp_avg_sqs, at::TensorList state_steps, const at::Tensor & lr, double beta1, double beta2, double weight_decay, double eps, bool amsgrad, bool maximize, const c10::optional<at::Tensor> & grad_scale, const c10::optional<at::Tensor> & found_inf);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::TensorList grads, at::TensorList exp_avgs, at::TensorList exp_avg_sqs, at::TensorList max_exp_avg_sqs, at::TensorList state_steps, const at::Tensor & lr, double beta1, double beta2, double weight_decay, double eps, bool amsgrad, bool maximize, const c10::optional<at::Tensor> & grad_scale, const c10::optional<at::Tensor> & found_inf);
+};
+
+struct TORCH_API _fused_adamw_out {
+  using schema = void (at::TensorList, at::TensorList, at::TensorList, at::TensorList, at::TensorList, at::TensorList, double, double, double, double, double, bool, bool, const c10::optional<at::Tensor> &, const c10::optional<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::_fused_adamw")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_fused_adamw.out(Tensor[] self, Tensor(b!)[] grads, Tensor(c!)[] exp_avgs, Tensor(d!)[] exp_avg_sqs, Tensor(e!)[] max_exp_avg_sqs, Tensor[] state_steps, *, float lr, float beta1, float beta2, float weight_decay, float eps, bool amsgrad, bool maximize, Tensor? grad_scale=None, Tensor? found_inf=None, Tensor(a!)[] out) -> ()")
+  static void call(at::TensorList self, at::TensorList grads, at::TensorList exp_avgs, at::TensorList exp_avg_sqs, at::TensorList max_exp_avg_sqs, at::TensorList state_steps, double lr, double beta1, double beta2, double weight_decay, double eps, bool amsgrad, bool maximize, const c10::optional<at::Tensor> & grad_scale, const c10::optional<at::Tensor> & found_inf, at::TensorList out);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::TensorList grads, at::TensorList exp_avgs, at::TensorList exp_avg_sqs, at::TensorList max_exp_avg_sqs, at::TensorList state_steps, double lr, double beta1, double beta2, double weight_decay, double eps, bool amsgrad, bool maximize, const c10::optional<at::Tensor> & grad_scale, const c10::optional<at::Tensor> & found_inf, at::TensorList out);
+};
+
+struct TORCH_API _fused_adamw {
+  using schema = ::std::tuple<::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>> (at::TensorList, at::TensorList, at::TensorList, at::TensorList, at::TensorList, at::TensorList, double, double, double, double, double, bool, bool, const c10::optional<at::Tensor> &, const c10::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::_fused_adamw")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_fused_adamw(Tensor[] self, Tensor[] grads, Tensor[] exp_avgs, Tensor[] exp_avg_sqs, Tensor[] max_exp_avg_sqs, Tensor[] state_steps, *, float lr, float beta1, float beta2, float weight_decay, float eps, bool amsgrad, bool maximize, Tensor? grad_scale=None, Tensor? found_inf=None) -> (Tensor[] self_out, Tensor[] grads_out, Tensor[] exp_avgs_out, Tensor[] exp_avg_sqs_out, Tensor[] max_exp_avg_sqs_out)")
+  static ::std::tuple<::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>> call(at::TensorList self, at::TensorList grads, at::TensorList exp_avgs, at::TensorList exp_avg_sqs, at::TensorList max_exp_avg_sqs, at::TensorList state_steps, double lr, double beta1, double beta2, double weight_decay, double eps, bool amsgrad, bool maximize, const c10::optional<at::Tensor> & grad_scale, const c10::optional<at::Tensor> & found_inf);
+  static ::std::tuple<::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>> redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::TensorList grads, at::TensorList exp_avgs, at::TensorList exp_avg_sqs, at::TensorList max_exp_avg_sqs, at::TensorList state_steps, double lr, double beta1, double beta2, double weight_decay, double eps, bool amsgrad, bool maximize, const c10::optional<at::Tensor> & grad_scale, const c10::optional<at::Tensor> & found_inf);
+};
+
+struct TORCH_API _fused_adamw_tensor_lr_out {
+  using schema = void (at::TensorList, at::TensorList, at::TensorList, at::TensorList, at::TensorList, at::TensorList, const at::Tensor &, double, double, double, double, bool, bool, const c10::optional<at::Tensor> &, const c10::optional<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::_fused_adamw")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "tensor_lr_out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_fused_adamw.tensor_lr_out(Tensor[] self, Tensor(b!)[] grads, Tensor(c!)[] exp_avgs, Tensor(d!)[] exp_avg_sqs, Tensor(e!)[] max_exp_avg_sqs, Tensor[] state_steps, *, Tensor lr, float beta1, float beta2, float weight_decay, float eps, bool amsgrad, bool maximize, Tensor? grad_scale=None, Tensor? found_inf=None, Tensor(a!)[] out) -> ()")
+  static void call(at::TensorList self, at::TensorList grads, at::TensorList exp_avgs, at::TensorList exp_avg_sqs, at::TensorList max_exp_avg_sqs, at::TensorList state_steps, const at::Tensor & lr, double beta1, double beta2, double weight_decay, double eps, bool amsgrad, bool maximize, const c10::optional<at::Tensor> & grad_scale, const c10::optional<at::Tensor> & found_inf, at::TensorList out);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::TensorList grads, at::TensorList exp_avgs, at::TensorList exp_avg_sqs, at::TensorList max_exp_avg_sqs, at::TensorList state_steps, const at::Tensor & lr, double beta1, double beta2, double weight_decay, double eps, bool amsgrad, bool maximize, const c10::optional<at::Tensor> & grad_scale, const c10::optional<at::Tensor> & found_inf, at::TensorList out);
+};
+
+struct TORCH_API _fused_adamw_tensor_lr {
+  using schema = ::std::tuple<::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>> (at::TensorList, at::TensorList, at::TensorList, at::TensorList, at::TensorList, at::TensorList, const at::Tensor &, double, double, double, double, bool, bool, const c10::optional<at::Tensor> &, const c10::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::_fused_adamw")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "tensor_lr")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_fused_adamw.tensor_lr(Tensor[] self, Tensor[] grads, Tensor[] exp_avgs, Tensor[] exp_avg_sqs, Tensor[] max_exp_avg_sqs, Tensor[] state_steps, *, Tensor lr, float beta1, float beta2, float weight_decay, float eps, bool amsgrad, bool maximize, Tensor? grad_scale=None, Tensor? found_inf=None) -> (Tensor[] self_out, Tensor[] grads_out, Tensor[] exp_avgs_out, Tensor[] exp_avg_sqs_out, Tensor[] max_exp_avg_sqs_out)")
+  static ::std::tuple<::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>> call(at::TensorList self, at::TensorList grads, at::TensorList exp_avgs, at::TensorList exp_avg_sqs, at::TensorList max_exp_avg_sqs, at::TensorList state_steps, const at::Tensor & lr, double beta1, double beta2, double weight_decay, double eps, bool amsgrad, bool maximize, const c10::optional<at::Tensor> & grad_scale, const c10::optional<at::Tensor> & found_inf);
+  static ::std::tuple<::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>,::std::vector<at::Tensor>> redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::TensorList grads, at::TensorList exp_avgs, at::TensorList exp_avg_sqs, at::TensorList max_exp_avg_sqs, at::TensorList state_steps, const at::Tensor & lr, double beta1, double beta2, double weight_decay, double eps, bool amsgrad, bool maximize, const c10::optional<at::Tensor> & grad_scale, const c10::optional<at::Tensor> & found_inf);
+};
+
+}} // namespace at::_ops

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_fused_sdp_choice_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 _fused_sdp_choice {
+  using schema = int64_t (const at::Tensor &, const at::Tensor &, const at::Tensor &, const c10::optional<at::Tensor> &, double, bool, c10::optional<double>);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_fused_sdp_choice")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_fused_sdp_choice(Tensor query, Tensor key, Tensor value, Tensor? attn_mask=None, float dropout_p=0.0, bool is_causal=False, *, float? scale=None) -> int")
+  static int64_t call(const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const c10::optional<at::Tensor> & attn_mask, double dropout_p, bool is_causal, c10::optional<double> scale);
+  static int64_t redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const c10::optional<at::Tensor> & attn_mask, double dropout_p, bool is_causal, c10::optional<double> scale);
+};
+
+}} // namespace at::_ops

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_fused_sgd_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 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 c10::optional<at::Tensor> & grad_scale={}, const c10::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 c10::optional<at::Tensor> & grad_scale={}, const c10::optional<at::Tensor> & found_inf={});
+
+} // namespace cuda
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_native_multi_head_attention_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 &> _native_multi_head_attention_out(at::Tensor & out0, at::Tensor & out1, const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, int64_t embed_dim, int64_t num_head, const at::Tensor & qkv_weight, const at::Tensor & qkv_bias, const at::Tensor & proj_weight, const at::Tensor & proj_bias, const c10::optional<at::Tensor> & mask={}, bool need_weights=true, bool average_attn_weights=true, c10::optional<int64_t> mask_type=c10::nullopt);
+TORCH_API ::std::tuple<at::Tensor &,at::Tensor &> _native_multi_head_attention_outf(const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, int64_t embed_dim, int64_t num_head, const at::Tensor & qkv_weight, const at::Tensor & qkv_bias, const at::Tensor & proj_weight, const at::Tensor & proj_bias, const c10::optional<at::Tensor> & mask, bool need_weights, bool average_attn_weights, c10::optional<int64_t> mask_type, at::Tensor & out0, at::Tensor & out1);
+
+} // namespace compositeexplicitautograd
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_nested_tensor_from_mask_left_aligned.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 <c10/util/Optional.h>
+
+
+
+#include <ATen/ops/_nested_tensor_from_mask_left_aligned_ops.h>
+
+namespace at {
+
+
+// aten::_nested_tensor_from_mask_left_aligned(Tensor t, Tensor mask) -> bool
+inline bool _nested_tensor_from_mask_left_aligned(const at::Tensor & t, const at::Tensor & mask) {
+    return at::_ops::_nested_tensor_from_mask_left_aligned::call(t, mask);
+}
+
+}

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_pin_memory_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 & _pin_memory_out(at::Tensor & out, const at::Tensor & self, c10::optional<at::Device> device=c10::nullopt);
+TORCH_API at::Tensor & _pin_memory_outf(const at::Tensor & self, c10::optional<at::Device> device, at::Tensor & out);
+
+} // namespace compositeexplicitautograd
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_reshape_alias_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 <c10/util/Optional.h>
+#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 _reshape_alias(const at::Tensor & self, at::IntArrayRef size, at::IntArrayRef stride);
+} // namespace native
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_segment_reduce_backward.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 <c10/util/Optional.h>
+
+
+
+#include <ATen/ops/_segment_reduce_backward_ops.h>
+
+namespace at {
+
+
+// aten::_segment_reduce_backward(Tensor grad, Tensor output, Tensor data, str reduce, *, Tensor? lengths=None, Tensor? offsets=None, int axis=0, Scalar? initial=None) -> Tensor
+inline at::Tensor _segment_reduce_backward(const at::Tensor & grad, const at::Tensor & output, const at::Tensor & data, c10::string_view reduce, const c10::optional<at::Tensor> & lengths={}, const c10::optional<at::Tensor> & offsets={}, int64_t axis=0, const c10::optional<at::Scalar> & initial=c10::nullopt) {
+    return at::_ops::_segment_reduce_backward::call(grad, output, data, reduce, lengths, offsets, axis, initial);
+}
+
+// aten::_segment_reduce_backward.out(Tensor grad, Tensor output, Tensor data, str reduce, *, Tensor? lengths=None, Tensor? offsets=None, int axis=0, Scalar? initial=None, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & _segment_reduce_backward_out(at::Tensor & out, const at::Tensor & grad, const at::Tensor & output, const at::Tensor & data, c10::string_view reduce, const c10::optional<at::Tensor> & lengths={}, const c10::optional<at::Tensor> & offsets={}, int64_t axis=0, const c10::optional<at::Scalar> & initial=c10::nullopt) {
+    return at::_ops::_segment_reduce_backward_out::call(grad, output, data, reduce, lengths, offsets, axis, initial, out);
+}
+// aten::_segment_reduce_backward.out(Tensor grad, Tensor output, Tensor data, str reduce, *, Tensor? lengths=None, Tensor? offsets=None, int axis=0, Scalar? initial=None, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & _segment_reduce_backward_outf(const at::Tensor & grad, const at::Tensor & output, const at::Tensor & data, c10::string_view reduce, const c10::optional<at::Tensor> & lengths, const c10::optional<at::Tensor> & offsets, int64_t axis, const c10::optional<at::Scalar> & initial, at::Tensor & out) {
+    return at::_ops::_segment_reduce_backward_out::call(grad, output, data, reduce, lengths, offsets, axis, initial, out);
+}
+
+}

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_softmax_cuda_dispatch.h

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

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_sparse_sparse_matmul_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_sparse_matmul_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & other);
+TORCH_API at::Tensor & _sparse_sparse_matmul_outf(const at::Tensor & self, const at::Tensor & other, at::Tensor & out);
+
+} // namespace compositeexplicitautograd
+} // namespace at

tuning-competition-baseline/.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 &, c10::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, c10::optional<int64_t> dense_dim);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, c10::optional<int64_t> dense_dim);
+};
+
+struct TORCH_API _to_sparse_csr_out {
+  using schema = at::Tensor & (const at::Tensor &, c10::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, c10::optional<int64_t> dense_dim, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, c10::optional<int64_t> dense_dim, at::Tensor & out);
+};
+
+}} // namespace at::_ops

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_values_copy_native.h

@@ -0,0 +1,22 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <c10/util/Optional.h>
+#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 & _values_copy_out(const at::Tensor & self, at::Tensor & out);
+TORCH_API at::Tensor _values_copy(const at::Tensor & self);
+} // namespace native
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/addcmul_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 addcmul(const at::Tensor & self, const at::Tensor & tensor1, const at::Tensor & tensor2, const at::Scalar & value=1);
+TORCH_API at::Tensor & addcmul_(at::Tensor & self, const at::Tensor & tensor1, const at::Tensor & tensor2, const at::Scalar & value=1);
+
+} // namespace compositeexplicitautogradnonfunctional
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/align_to_compositeimplicitautograd_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 compositeimplicitautograd {
+
+TORCH_API at::Tensor align_to(const at::Tensor & self, at::DimnameList names);
+TORCH_API at::Tensor align_to(const at::Tensor & self, at::DimnameList order, int64_t ellipsis_idx);
+
+} // namespace compositeimplicitautograd
+} // namespace at

tuning-competition-baseline/.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 <c10/util/Optional.h>
+
+
+
+#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 c10::optional<at::Tensor> & running_mean, const c10::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 c10::optional<at::Tensor> & running_mean, const c10::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 c10::optional<at::Tensor> & running_mean, const c10::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);
+}
+
+}

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/bincount_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 bincount {
+  using schema = at::Tensor (const at::Tensor &, const c10::optional<at::Tensor> &, int64_t);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::bincount")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "bincount(Tensor self, Tensor? weights=None, int minlength=0) -> Tensor")
+  static at::Tensor call(const at::Tensor & self, const c10::optional<at::Tensor> & weights, int64_t minlength);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const c10::optional<at::Tensor> & weights, int64_t minlength);
+};
+
+struct TORCH_API bincount_out {
+  using schema = at::Tensor & (const at::Tensor &, const c10::optional<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::bincount")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
+  STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "bincount.out(Tensor self, Tensor? weights=None, int minlength=0, *, Tensor(a!) out) -> Tensor(a!)")
+  static at::Tensor & call(const at::Tensor & self, const c10::optional<at::Tensor> & weights, int64_t minlength, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const c10::optional<at::Tensor> & weights, int64_t minlength, at::Tensor & out);
+};
+
+}} // namespace at::_ops

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/bitwise_and_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 bitwise_and(const at::Tensor & self, const at::Tensor & other);
+TORCH_API at::Tensor & bitwise_and_(at::Tensor & self, const at::Tensor & other);
+
+} // namespace compositeexplicitautogradnonfunctional
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/cholesky_inverse_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 cholesky_inverse(const at::Tensor & self, bool upper=false);
+TORCH_API at::Tensor & cholesky_inverse_out(at::Tensor & out, const at::Tensor & self, bool upper=false);
+TORCH_API at::Tensor & cholesky_inverse_outf(const at::Tensor & self, bool upper, at::Tensor & out);
+
+} // namespace cpu
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/conv1d_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 <c10/util/Optional.h>
+#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 conv1d_symint(const at::Tensor & input, const at::Tensor & weight, const c10::optional<at::Tensor> & bias={}, c10::SymIntArrayRef stride=c10::SymInt(1), c10::SymIntArrayRef padding=c10::SymInt(0), c10::SymIntArrayRef dilation=c10::SymInt(1), c10::SymInt groups=1);
+TORCH_API at::Tensor conv1d_padding_symint(const at::Tensor & input, const at::Tensor & weight, const c10::optional<at::Tensor> & bias={}, c10::SymIntArrayRef stride=c10::SymInt(1), c10::string_view padding="valid", c10::SymIntArrayRef dilation=c10::SymInt(1), c10::SymInt groups=1);
+} // namespace native
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/conv_transpose3d_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 <c10/util/Optional.h>
+#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 conv_transpose3d_symint(const at::Tensor & input, const at::Tensor & weight, const c10::optional<at::Tensor> & bias={}, c10::SymIntArrayRef stride=c10::SymInt(1), c10::SymIntArrayRef padding=c10::SymInt(0), c10::SymIntArrayRef output_padding=c10::SymInt(0), c10::SymInt groups=1, c10::SymIntArrayRef dilation=c10::SymInt(1));
+} // namespace native
+} // namespace at

tuning-competition-baseline/.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 <c10/util/Optional.h>
+
+
+
+#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 c10::optional<at::Tensor> & fweights={}, const c10::optional<at::Tensor> & aweights={}) {
+    return at::_ops::cov::call(self, correction, fweights, aweights);
+}
+
+}

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/cudnn_batch_norm_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,at::Tensor,at::Tensor> cudnn_batch_norm(const at::Tensor & input, const at::Tensor & weight, const c10::optional<at::Tensor> & bias, const c10::optional<at::Tensor> & running_mean, const c10::optional<at::Tensor> & running_var, bool training, double exponential_average_factor, double epsilon);
+
+} // namespace cuda
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/cudnn_convolution_transpose_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 <c10/util/Optional.h>
+#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 & cudnn_convolution_transpose_out_symint(const at::Tensor & self, const at::Tensor & weight, c10::SymIntArrayRef padding, c10::SymIntArrayRef output_padding, c10::SymIntArrayRef stride, c10::SymIntArrayRef dilation, c10::SymInt groups, bool benchmark, bool deterministic, bool allow_tf32, at::Tensor & out);
+TORCH_API at::Tensor cudnn_convolution_transpose(const at::Tensor & self, const at::Tensor & weight, at::IntArrayRef padding, at::IntArrayRef output_padding, at::IntArrayRef stride, at::IntArrayRef dilation, int64_t groups, bool benchmark, bool deterministic, bool allow_tf32);
+} // namespace native
+} // namespace at

tuning-competition-baseline/.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

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/fft_hfftn.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 <c10/util/Optional.h>
+
+
+
+#include <ATen/ops/fft_hfftn_ops.h>
+
+namespace at {
+
+
+// aten::fft_hfftn(Tensor self, SymInt[1]? s=None, int[1]? dim=None, str? norm=None) -> Tensor
+inline at::Tensor fft_hfftn(const at::Tensor & self, at::OptionalIntArrayRef s=c10::nullopt, at::OptionalIntArrayRef dim=c10::nullopt, c10::optional<c10::string_view> norm=c10::nullopt) {
+    return at::_ops::fft_hfftn::call(self, s.has_value() ? c10::make_optional(c10::fromIntArrayRefSlow(*s)) : c10::nullopt, dim, norm);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
+  at::Tensor fft_hfftn(const at::Tensor & self, at::OptionalIntArrayRef s=c10::nullopt, at::OptionalIntArrayRef dim=c10::nullopt, c10::optional<c10::string_view> norm=c10::nullopt) {
+    return at::_ops::fft_hfftn::call(self, s.has_value() ? c10::make_optional(c10::fromIntArrayRefSlow(*s)) : c10::nullopt, dim, norm);
+  }
+}
+
+// aten::fft_hfftn(Tensor self, SymInt[1]? s=None, int[1]? dim=None, str? norm=None) -> Tensor
+inline at::Tensor fft_hfftn_symint(const at::Tensor & self, at::OptionalSymIntArrayRef s=c10::nullopt, at::OptionalIntArrayRef dim=c10::nullopt, c10::optional<c10::string_view> norm=c10::nullopt) {
+    return at::_ops::fft_hfftn::call(self, s, dim, norm);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
+  at::Tensor fft_hfftn(const at::Tensor & self, at::OptionalSymIntArrayRef s=c10::nullopt, at::OptionalIntArrayRef dim=c10::nullopt, c10::optional<c10::string_view> norm=c10::nullopt) {
+    return at::_ops::fft_hfftn::call(self, s, dim, norm);
+  }
+}
+
+// aten::fft_hfftn.out(Tensor self, SymInt[1]? s=None, int[1]? dim=None, str? norm=None, *, Tensor(a!) out) -> Tensor(a!)
+inline const at::Tensor & fft_hfftn_out(const at::Tensor & out, const at::Tensor & self, at::OptionalIntArrayRef s=c10::nullopt, at::OptionalIntArrayRef dim=c10::nullopt, c10::optional<c10::string_view> norm=c10::nullopt) {
+    return at::_ops::fft_hfftn_out::call(self, s.has_value() ? c10::make_optional(c10::fromIntArrayRefSlow(*s)) : c10::nullopt, dim, norm, out);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
+  const at::Tensor & fft_hfftn_out(const at::Tensor & out, const at::Tensor & self, at::OptionalIntArrayRef s=c10::nullopt, at::OptionalIntArrayRef dim=c10::nullopt, c10::optional<c10::string_view> norm=c10::nullopt) {
+    return at::_ops::fft_hfftn_out::call(self, s.has_value() ? c10::make_optional(c10::fromIntArrayRefSlow(*s)) : c10::nullopt, dim, norm, out);
+  }
+}
+
+// aten::fft_hfftn.out(Tensor self, SymInt[1]? s=None, int[1]? dim=None, str? norm=None, *, Tensor(a!) out) -> Tensor(a!)
+inline const at::Tensor & fft_hfftn_outf(const at::Tensor & self, at::OptionalIntArrayRef s, at::OptionalIntArrayRef dim, c10::optional<c10::string_view> norm, const at::Tensor & out) {
+    return at::_ops::fft_hfftn_out::call(self, s.has_value() ? c10::make_optional(c10::fromIntArrayRefSlow(*s)) : c10::nullopt, dim, norm, out);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
+  const at::Tensor & fft_hfftn_outf(const at::Tensor & self, at::OptionalIntArrayRef s, at::OptionalIntArrayRef dim, c10::optional<c10::string_view> norm, const at::Tensor & out) {
+    return at::_ops::fft_hfftn_out::call(self, s.has_value() ? c10::make_optional(c10::fromIntArrayRefSlow(*s)) : c10::nullopt, dim, norm, out);
+  }
+}
+
+// aten::fft_hfftn.out(Tensor self, SymInt[1]? s=None, int[1]? dim=None, str? norm=None, *, Tensor(a!) out) -> Tensor(a!)
+inline const at::Tensor & fft_hfftn_symint_out(const at::Tensor & out, const at::Tensor & self, at::OptionalSymIntArrayRef s=c10::nullopt, at::OptionalIntArrayRef dim=c10::nullopt, c10::optional<c10::string_view> norm=c10::nullopt) {
+    return at::_ops::fft_hfftn_out::call(self, s, dim, norm, out);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
+  const at::Tensor & fft_hfftn_out(const at::Tensor & out, const at::Tensor & self, at::OptionalSymIntArrayRef s=c10::nullopt, at::OptionalIntArrayRef dim=c10::nullopt, c10::optional<c10::string_view> norm=c10::nullopt) {
+    return at::_ops::fft_hfftn_out::call(self, s, dim, norm, out);
+  }
+}
+
+// aten::fft_hfftn.out(Tensor self, SymInt[1]? s=None, int[1]? dim=None, str? norm=None, *, Tensor(a!) out) -> Tensor(a!)
+inline const at::Tensor & fft_hfftn_symint_outf(const at::Tensor & self, at::OptionalSymIntArrayRef s, at::OptionalIntArrayRef dim, c10::optional<c10::string_view> norm, const at::Tensor & out) {
+    return at::_ops::fft_hfftn_out::call(self, s, dim, norm, out);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
+  const at::Tensor & fft_hfftn_outf(const at::Tensor & self, at::OptionalSymIntArrayRef s, at::OptionalIntArrayRef dim, c10::optional<c10::string_view> norm, const at::Tensor & out) {
+    return at::_ops::fft_hfftn_out::call(self, s, dim, norm, out);
+  }
+}
+
+}

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/fft_hfftn_ops.h

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

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/fractional_max_pool2d_backward_cuda_dispatch.h

@@ -0,0 +1,25 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API at::Tensor fractional_max_pool2d_backward(const at::Tensor & grad_output, const at::Tensor & self, at::IntArrayRef kernel_size, at::IntArrayRef output_size, const at::Tensor & indices);
+TORCH_API at::Tensor & fractional_max_pool2d_backward_out(at::Tensor & grad_input, const at::Tensor & grad_output, const at::Tensor & self, at::IntArrayRef kernel_size, at::IntArrayRef output_size, const at::Tensor & indices);
+TORCH_API at::Tensor & fractional_max_pool2d_backward_outf(const at::Tensor & grad_output, const at::Tensor & self, at::IntArrayRef kernel_size, at::IntArrayRef output_size, const at::Tensor & indices, at::Tensor & grad_input);
+
+} // namespace cuda
+} // namespace at

tuning-competition-baseline/.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 <c10/util/Optional.h>
+#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

tuning-competition-baseline/.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

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/hstack_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 <c10/util/Optional.h>
+#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 hstack(at::TensorList tensors);
+TORCH_API at::Tensor & hstack_out(at::TensorList tensors, at::Tensor & out);
+} // namespace native
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/isneginf_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 isneginf(const at::Tensor & self);
+TORCH_API at::Tensor & isneginf_out(at::Tensor & out, const at::Tensor & self);
+TORCH_API at::Tensor & isneginf_outf(const at::Tensor & self, at::Tensor & out);
+
+} // namespace cpu
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/less.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 <c10/util/Optional.h>
+
+
+
+#include <ATen/ops/less_ops.h>
+
+namespace at {
+
+
+// aten::less.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & less_out(at::Tensor & out, const at::Tensor & self, const at::Scalar & other) {
+    return at::_ops::less_Scalar_out::call(self, other, out);
+}
+// aten::less.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & less_outf(const at::Tensor & self, const at::Scalar & other, at::Tensor & out) {
+    return at::_ops::less_Scalar_out::call(self, other, out);
+}
+
+// aten::less.Scalar(Tensor self, Scalar other) -> Tensor
+inline at::Tensor less(const at::Tensor & self, const at::Scalar & other) {
+    return at::_ops::less_Scalar::call(self, other);
+}
+
+// aten::less.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & less_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & other) {
+    return at::_ops::less_Tensor_out::call(self, other, out);
+}
+// aten::less.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & less_outf(const at::Tensor & self, const at::Tensor & other, at::Tensor & out) {
+    return at::_ops::less_Tensor_out::call(self, other, out);
+}
+
+// aten::less.Tensor(Tensor self, Tensor other) -> Tensor
+inline at::Tensor less(const at::Tensor & self, const at::Tensor & other) {
+    return at::_ops::less_Tensor::call(self, other);
+}
+
+}

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/less_equal_ops.h

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

tuning-competition-baseline/.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

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_qr_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 <c10/util/Optional.h>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+#include <ATen/ops/linalg_qr_meta.h>
+
+namespace at {
+namespace native {
+struct TORCH_API structured_linalg_qr_out : public at::meta::structured_linalg_qr {
+void impl(const at::Tensor & A, c10::string_view mode, const at::Tensor & Q, const at::Tensor & R);
+};
+} // namespace native
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_svdvals.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 <c10/util/Optional.h>
+
+
+
+#include <ATen/ops/linalg_svdvals_ops.h>
+
+namespace at {
+
+
+// aten::linalg_svdvals(Tensor A, *, str? driver=None) -> Tensor
+inline at::Tensor linalg_svdvals(const at::Tensor & A, c10::optional<c10::string_view> driver=c10::nullopt) {
+    return at::_ops::linalg_svdvals::call(A, driver);
+}
+
+// aten::linalg_svdvals.out(Tensor A, *, str? driver=None, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & linalg_svdvals_out(at::Tensor & out, const at::Tensor & A, c10::optional<c10::string_view> driver=c10::nullopt) {
+    return at::_ops::linalg_svdvals_out::call(A, driver, out);
+}
+// aten::linalg_svdvals.out(Tensor A, *, str? driver=None, Tensor(a!) out) -> Tensor(a!)
+inline at::Tensor & linalg_svdvals_outf(const at::Tensor & A, c10::optional<c10::string_view> driver, at::Tensor & out) {
+    return at::_ops::linalg_svdvals_out::call(A, driver, out);
+}
+
+}

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_tensorsolve_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 <c10/util/Optional.h>
+#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 linalg_tensorsolve(const at::Tensor & self, const at::Tensor & other, at::OptionalIntArrayRef dims=c10::nullopt);
+TORCH_API at::Tensor & linalg_tensorsolve_out(const at::Tensor & self, const at::Tensor & other, at::OptionalIntArrayRef dims, at::Tensor & out);
+} // namespace native
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_vector_norm_cuda_dispatch.h

@@ -0,0 +1,25 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API at::Tensor linalg_vector_norm(const at::Tensor & self, const at::Scalar & ord=2, at::OptionalIntArrayRef dim=c10::nullopt, bool keepdim=false, c10::optional<at::ScalarType> dtype=c10::nullopt);
+TORCH_API at::Tensor & linalg_vector_norm_out(at::Tensor & out, const at::Tensor & self, const at::Scalar & ord=2, at::OptionalIntArrayRef dim=c10::nullopt, bool keepdim=false, c10::optional<at::ScalarType> dtype=c10::nullopt);
+TORCH_API at::Tensor & linalg_vector_norm_outf(const at::Tensor & self, const at::Scalar & ord, at::OptionalIntArrayRef dim, bool keepdim, c10::optional<at::ScalarType> dtype, at::Tensor & out);
+
+} // namespace cuda
+} // namespace at

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/logical_xor_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 & logical_xor_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & other);
+TORCH_API at::Tensor & logical_xor_outf(const at::Tensor & self, const at::Tensor & other, at::Tensor & out);
+
+} // namespace cpu
+} // namespace at

tuning-competition-baseline/.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 <c10/util/Optional.h>
+#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

tuning-competition-baseline/.venv/lib/python3.11/site-packages/torch/include/ATen/ops/max_pool2d_with_indices.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 <c10/util/Optional.h>
+
+
+
+#include <ATen/ops/max_pool2d_with_indices_ops.h>
+
+namespace at {
+
+
+// aten::max_pool2d_with_indices.out(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(b!) indices) -> (Tensor(a!), Tensor(b!))
+inline ::std::tuple<at::Tensor &,at::Tensor &> max_pool2d_with_indices_out(at::Tensor & out, at::Tensor & indices, const at::Tensor & self, at::IntArrayRef kernel_size, at::IntArrayRef stride={}, at::IntArrayRef padding=0, at::IntArrayRef dilation=1, bool ceil_mode=false) {
+    return at::_ops::max_pool2d_with_indices_out::call(self, kernel_size, stride, padding, dilation, ceil_mode, out, indices);
+}
+// aten::max_pool2d_with_indices.out(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(b!) indices) -> (Tensor(a!), Tensor(b!))
+inline ::std::tuple<at::Tensor &,at::Tensor &> max_pool2d_with_indices_outf(const at::Tensor & self, at::IntArrayRef kernel_size, at::IntArrayRef stride, at::IntArrayRef padding, at::IntArrayRef dilation, bool ceil_mode, at::Tensor & out, at::Tensor & indices) {
+    return at::_ops::max_pool2d_with_indices_out::call(self, kernel_size, stride, padding, dilation, ceil_mode, out, indices);
+}
+
+// aten::max_pool2d_with_indices(Tensor self, int[2] kernel_size, int[2] stride=[], int[2] padding=0, int[2] dilation=1, bool ceil_mode=False) -> (Tensor, Tensor)
+inline ::std::tuple<at::Tensor,at::Tensor> max_pool2d_with_indices(const at::Tensor & self, at::IntArrayRef kernel_size, at::IntArrayRef stride={}, at::IntArrayRef padding=0, at::IntArrayRef dilation=1, bool ceil_mode=false) {
+    return at::_ops::max_pool2d_with_indices::call(self, kernel_size, stride, padding, dilation, ceil_mode);
+}
+
+}