koichi12 commited on
Commit
7ccf023
·
verified ·
1 Parent(s): aeeeb0f

Add files using upload-large-folder tool

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_adaptive_avg_pool2d_native.h +25 -0
  2. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_amp_update_scale_cpu_dispatch.h +23 -0
  3. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_batch_norm_with_update_ops.h +50 -0
  4. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_cast_Long.h +30 -0
  5. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_coalesced.h +39 -0
  6. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_convert_indices_from_coo_to_csr_meta.h +27 -0
  7. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_ctc_loss_backward_compositeexplicitautograd_dispatch.h +24 -0
  8. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_dim_arange.h +30 -0
  9. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_efficientzerotensor.h +113 -0
  10. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_embedding_bag_forward_only_compositeexplicitautograd_dispatch.h +24 -0
  11. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_fake_quantize_learnable_per_channel_affine.h +39 -0
  12. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_fake_quantize_learnable_per_channel_affine_cpu_dispatch.h +23 -0
  13. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_foreach_tanh_compositeexplicitautograd_dispatch.h +26 -0
  14. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_histogramdd_from_bin_tensors_compositeexplicitautograd_dispatch.h +24 -0
  15. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_linalg_svd_native.h +23 -0
  16. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_local_scalar_dense.h +30 -0
  17. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_mixed_dtypes_linear_ops.h +28 -0
  18. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_scaled_dot_product_flash_attention_backward.h +47 -0
  19. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_sobol_engine_initialize_state_ops.h +28 -0
  20. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_sparse_log_softmax_ops.h +61 -0
  21. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_sparse_softmax.h +49 -0
  22. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_unpack_dual_ops.h +28 -0
  23. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/_weight_int4pack_mm_ops.h +28 -0
  24. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/abs.h +44 -0
  25. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/adaptive_max_pool1d_compositeimplicitautograd_dispatch.h +23 -0
  26. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/adaptive_max_pool2d_backward.h +39 -0
  27. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/addcdiv_meta_dispatch.h +26 -0
  28. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/batch_norm_gather_stats_native.h +22 -0
  29. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/bilinear_native.h +21 -0
  30. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/bitwise_or_cuda_dispatch.h +26 -0
  31. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/conv1d_ops.h +39 -0
  32. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/gather_backward_ops.h +28 -0
  33. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/hardswish_cpu_dispatch.h +26 -0
  34. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/hspmm_ops.h +39 -0
  35. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/hypot_meta_dispatch.h +26 -0
  36. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/inverse_ops.h +39 -0
  37. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/isneginf_cuda_dispatch.h +25 -0
  38. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_lu_factor_compositeimplicitautograd_dispatch.h +25 -0
  39. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_matrix_exp_cuda_dispatch.h +23 -0
  40. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_slogdet_compositeimplicitautograd_dispatch.h +25 -0
  41. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_solve_triangular_cuda_dispatch.h +25 -0
  42. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/lu_unpack.h +39 -0
  43. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/mean_compositeimplicitautograd_dispatch.h +25 -0
  44. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/msort_ops.h +39 -0
  45. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/multi_margin_loss_native.h +24 -0
  46. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/nanmedian_cpu_dispatch.h +25 -0
  47. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/native_batch_norm_backward_cuda_dispatch.h +23 -0
  48. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/new_empty_strided_compositeexplicitautogradnonfunctional_dispatch.h +26 -0
  49. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/one_hot_compositeimplicitautograd_dispatch.h +23 -0
  50. .venv/lib/python3.11/site-packages/torch/include/ATen/ops/one_hot_native.h +21 -0
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_adaptive_avg_pool2d_native.h ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from NativeFunction.h
4
+
5
+ #include <c10/core/Scalar.h>
6
+ #include <c10/core/Storage.h>
7
+ #include <c10/core/TensorOptions.h>
8
+ #include <c10/util/Deprecated.h>
9
+ #include <optional>
10
+ #include <c10/core/QScheme.h>
11
+ #include <ATen/core/Reduction.h>
12
+ #include <ATen/core/Tensor.h>
13
+ #include <tuple>
14
+ #include <vector>
15
+
16
+
17
+ namespace at {
18
+ namespace native {
19
+ TORCH_API at::Tensor & _adaptive_avg_pool2d_out_symint(const at::Tensor & self, c10::SymIntArrayRef output_size, at::Tensor & out);
20
+ TORCH_API at::Tensor adaptive_avg_pool2d_cpu(const at::Tensor & self, at::IntArrayRef output_size);
21
+ TORCH_API at::Tensor adaptive_avg_pool2d_cuda(const at::Tensor & self, at::IntArrayRef output_size);
22
+ TORCH_API at::Tensor adaptive_avg_pool2d_quantized_cpu(const at::Tensor & self, at::IntArrayRef output_size);
23
+ TORCH_API at::Tensor adaptive_avg_pool2d_quantized_cuda(const at::Tensor & self, at::IntArrayRef output_size);
24
+ } // namespace native
25
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_amp_update_scale_cpu_dispatch.h ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace cpu {
19
+
20
+ TORCH_API at::Tensor & _amp_update_scale_(at::Tensor & self, at::Tensor & growth_tracker, const at::Tensor & found_inf, double scale_growth_factor, double scale_backoff_factor, int64_t growth_interval);
21
+
22
+ } // namespace cpu
23
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_batch_norm_with_update_ops.h ADDED
@@ -0,0 +1,50 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Operator.h
4
+
5
+ #include <tuple>
6
+ #include <vector>
7
+
8
+ // Forward declarations of any types needed in the operator signatures.
9
+ // We can't directly include these classes because it will cause circular include dependencies.
10
+ // This file is included by TensorBody.h, which defines the Tensor class.
11
+ #include <ATen/core/ATen_fwd.h>
12
+
13
+ namespace at {
14
+ namespace _ops {
15
+
16
+
17
+ struct TORCH_API _batch_norm_with_update {
18
+ using schema = ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor> (const at::Tensor &, const ::std::optional<at::Tensor> &, const ::std::optional<at::Tensor> &, at::Tensor &, at::Tensor &, double, double);
19
+ using ptr_schema = schema*;
20
+ // See Note [static constexpr char* members for windows NVCC]
21
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_batch_norm_with_update")
22
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
23
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_batch_norm_with_update(Tensor input, Tensor? weight, Tensor? bias, Tensor(a!) running_mean, Tensor(b!) running_var, float momentum, float eps) -> (Tensor, Tensor, Tensor, Tensor)")
24
+ static ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor> call(const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & bias, at::Tensor & running_mean, at::Tensor & running_var, double momentum, double eps);
25
+ static ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & bias, at::Tensor & running_mean, at::Tensor & running_var, double momentum, double eps);
26
+ };
27
+
28
+ struct TORCH_API _batch_norm_with_update_out {
29
+ using schema = ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> (const at::Tensor &, const ::std::optional<at::Tensor> &, const ::std::optional<at::Tensor> &, at::Tensor &, at::Tensor &, double, double, at::Tensor &, at::Tensor &, at::Tensor &, at::Tensor &);
30
+ using ptr_schema = schema*;
31
+ // See Note [static constexpr char* members for windows NVCC]
32
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_batch_norm_with_update")
33
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
34
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_batch_norm_with_update.out(Tensor input, Tensor? weight, Tensor? bias, Tensor(a!) running_mean, Tensor(b!) running_var, float momentum, float eps, *, Tensor(d!) out, Tensor(e!) save_mean, Tensor(f!) save_invstd, Tensor(g!) reserve) -> (Tensor(d!), Tensor(e!), Tensor(f!), Tensor(g!))")
35
+ static ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> call(const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & bias, at::Tensor & running_mean, at::Tensor & running_var, double momentum, double eps, at::Tensor & out, at::Tensor & save_mean, at::Tensor & save_invstd, at::Tensor & reserve);
36
+ static ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & bias, at::Tensor & running_mean, at::Tensor & running_var, double momentum, double eps, at::Tensor & out, at::Tensor & save_mean, at::Tensor & save_invstd, at::Tensor & reserve);
37
+ };
38
+
39
+ struct TORCH_API _batch_norm_with_update_functional {
40
+ using schema = ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,at::Tensor,at::Tensor> (const at::Tensor &, const ::std::optional<at::Tensor> &, const ::std::optional<at::Tensor> &, const at::Tensor &, const at::Tensor &, double, double);
41
+ using ptr_schema = schema*;
42
+ // See Note [static constexpr char* members for windows NVCC]
43
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_batch_norm_with_update_functional")
44
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
45
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_batch_norm_with_update_functional(Tensor input, Tensor? weight, Tensor? bias, Tensor running_mean, Tensor running_var, float momentum, float eps) -> (Tensor, Tensor, Tensor, Tensor, Tensor running_mean_out, Tensor running_var_out)")
46
+ static ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,at::Tensor,at::Tensor> call(const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & bias, const at::Tensor & running_mean, const at::Tensor & running_var, double momentum, double eps);
47
+ static ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,at::Tensor,at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & bias, const at::Tensor & running_mean, const at::Tensor & running_var, double momentum, double eps);
48
+ };
49
+
50
+ }} // namespace at::_ops
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_cast_Long.h ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Function.h
4
+
5
+ #include <ATen/Context.h>
6
+ #include <ATen/DeviceGuard.h>
7
+ #include <ATen/TensorUtils.h>
8
+ #include <ATen/TracerMode.h>
9
+ #include <ATen/core/Generator.h>
10
+ #include <ATen/core/Reduction.h>
11
+ #include <ATen/core/Tensor.h>
12
+ #include <c10/core/Scalar.h>
13
+ #include <c10/core/Storage.h>
14
+ #include <c10/core/TensorOptions.h>
15
+ #include <c10/util/Deprecated.h>
16
+ #include <optional>
17
+
18
+
19
+
20
+ #include <ATen/ops/_cast_Long_ops.h>
21
+
22
+ namespace at {
23
+
24
+
25
+ // aten::_cast_Long(Tensor self, bool non_blocking=False) -> Tensor
26
+ inline at::Tensor _cast_Long(const at::Tensor & self, bool non_blocking=false) {
27
+ return at::_ops::_cast_Long::call(self, non_blocking);
28
+ }
29
+
30
+ }
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_coalesced.h ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Function.h
4
+
5
+ #include <ATen/Context.h>
6
+ #include <ATen/DeviceGuard.h>
7
+ #include <ATen/TensorUtils.h>
8
+ #include <ATen/TracerMode.h>
9
+ #include <ATen/core/Generator.h>
10
+ #include <ATen/core/Reduction.h>
11
+ #include <ATen/core/Tensor.h>
12
+ #include <c10/core/Scalar.h>
13
+ #include <c10/core/Storage.h>
14
+ #include <c10/core/TensorOptions.h>
15
+ #include <c10/util/Deprecated.h>
16
+ #include <optional>
17
+
18
+
19
+
20
+ #include <ATen/ops/_coalesced_ops.h>
21
+
22
+ namespace at {
23
+
24
+
25
+ // aten::_coalesced.out(Tensor self, bool coalesced, *, Tensor(a!) out) -> Tensor(a!)
26
+ inline at::Tensor & _coalesced_out(at::Tensor & out, const at::Tensor & self, bool coalesced) {
27
+ return at::_ops::_coalesced_out::call(self, coalesced, out);
28
+ }
29
+ // aten::_coalesced.out(Tensor self, bool coalesced, *, Tensor(a!) out) -> Tensor(a!)
30
+ inline at::Tensor & _coalesced_outf(const at::Tensor & self, bool coalesced, at::Tensor & out) {
31
+ return at::_ops::_coalesced_out::call(self, coalesced, out);
32
+ }
33
+
34
+ // aten::_coalesced(Tensor self, bool coalesced) -> Tensor
35
+ inline at::Tensor _coalesced(const at::Tensor & self, bool coalesced) {
36
+ return at::_ops::_coalesced::call(self, coalesced);
37
+ }
38
+
39
+ }
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_convert_indices_from_coo_to_csr_meta.h ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from NativeMetaFunction.h
4
+
5
+ #include <c10/core/Scalar.h>
6
+ #include <c10/core/Storage.h>
7
+ #include <c10/core/TensorOptions.h>
8
+ #include <c10/util/Deprecated.h>
9
+ #include <optional>
10
+ #include <c10/core/QScheme.h>
11
+ #include <ATen/core/Reduction.h>
12
+ #include <ATen/TensorIterator.h>
13
+ #include <ATen/TensorMeta.h>
14
+ #include <tuple>
15
+ #include <vector>
16
+
17
+ namespace at {
18
+ namespace meta {
19
+
20
+ struct TORCH_API structured__convert_indices_from_coo_to_csr : public at::impl::MetaBase {
21
+
22
+
23
+ void meta(const at::Tensor & self, int64_t size, bool out_int32);
24
+ };
25
+
26
+ } // namespace native
27
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_ctc_loss_backward_compositeexplicitautograd_dispatch.h ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace compositeexplicitautograd {
19
+
20
+ TORCH_API at::Tensor & _ctc_loss_backward_out(at::Tensor & out, const at::Tensor & grad, const at::Tensor & log_probs, const at::Tensor & targets, at::IntArrayRef input_lengths, at::IntArrayRef target_lengths, const at::Tensor & neg_log_likelihood, const at::Tensor & log_alpha, int64_t blank, bool zero_infinity=false);
21
+ TORCH_API at::Tensor & _ctc_loss_backward_outf(const at::Tensor & grad, const at::Tensor & log_probs, const at::Tensor & targets, at::IntArrayRef input_lengths, at::IntArrayRef target_lengths, const at::Tensor & neg_log_likelihood, const at::Tensor & log_alpha, int64_t blank, bool zero_infinity, at::Tensor & out);
22
+
23
+ } // namespace compositeexplicitautograd
24
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_dim_arange.h ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Function.h
4
+
5
+ #include <ATen/Context.h>
6
+ #include <ATen/DeviceGuard.h>
7
+ #include <ATen/TensorUtils.h>
8
+ #include <ATen/TracerMode.h>
9
+ #include <ATen/core/Generator.h>
10
+ #include <ATen/core/Reduction.h>
11
+ #include <ATen/core/Tensor.h>
12
+ #include <c10/core/Scalar.h>
13
+ #include <c10/core/Storage.h>
14
+ #include <c10/core/TensorOptions.h>
15
+ #include <c10/util/Deprecated.h>
16
+ #include <optional>
17
+
18
+
19
+
20
+ #include <ATen/ops/_dim_arange_ops.h>
21
+
22
+ namespace at {
23
+
24
+
25
+ // aten::_dim_arange(Tensor like, int dim) -> Tensor
26
+ inline at::Tensor _dim_arange(const at::Tensor & like, int64_t dim) {
27
+ return at::_ops::_dim_arange::call(like, dim);
28
+ }
29
+
30
+ }
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_efficientzerotensor.h ADDED
@@ -0,0 +1,113 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Function.h
4
+
5
+ #include <ATen/Context.h>
6
+ #include <ATen/DeviceGuard.h>
7
+ #include <ATen/TensorUtils.h>
8
+ #include <ATen/TracerMode.h>
9
+ #include <ATen/core/Generator.h>
10
+ #include <ATen/core/Reduction.h>
11
+ #include <ATen/core/Tensor.h>
12
+ #include <c10/core/Scalar.h>
13
+ #include <c10/core/Storage.h>
14
+ #include <c10/core/TensorOptions.h>
15
+ #include <c10/util/Deprecated.h>
16
+ #include <optional>
17
+
18
+
19
+
20
+ #include <ATen/ops/_efficientzerotensor_ops.h>
21
+
22
+ namespace at {
23
+
24
+
25
+ // aten::_efficientzerotensor(SymInt[] size, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor
26
+ inline at::Tensor _efficientzerotensor(at::IntArrayRef size, at::TensorOptions options={}) {
27
+ return at::_ops::_efficientzerotensor::call(c10::fromIntArrayRefSlow(size), c10::optTypeMetaToScalarType(options.dtype_opt()), options.layout_opt(), options.device_opt(), options.pinned_memory_opt());
28
+ }
29
+ namespace symint {
30
+ template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
31
+ at::Tensor _efficientzerotensor(at::IntArrayRef size, at::TensorOptions options={}) {
32
+ return at::_ops::_efficientzerotensor::call(c10::fromIntArrayRefSlow(size), c10::optTypeMetaToScalarType(options.dtype_opt()), options.layout_opt(), options.device_opt(), options.pinned_memory_opt());
33
+ }
34
+ }
35
+
36
+ // aten::_efficientzerotensor(SymInt[] size, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor
37
+ inline at::Tensor _efficientzerotensor(at::IntArrayRef size, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory) {
38
+ return at::_ops::_efficientzerotensor::call(c10::fromIntArrayRefSlow(size), dtype, layout, device, pin_memory);
39
+ }
40
+ namespace symint {
41
+ template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
42
+ at::Tensor _efficientzerotensor(at::IntArrayRef size, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory) {
43
+ return at::_ops::_efficientzerotensor::call(c10::fromIntArrayRefSlow(size), dtype, layout, device, pin_memory);
44
+ }
45
+ }
46
+
47
+ // aten::_efficientzerotensor(SymInt[] size, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor
48
+ inline at::Tensor _efficientzerotensor_symint(c10::SymIntArrayRef size, at::TensorOptions options={}) {
49
+ return at::_ops::_efficientzerotensor::call(size, c10::optTypeMetaToScalarType(options.dtype_opt()), options.layout_opt(), options.device_opt(), options.pinned_memory_opt());
50
+ }
51
+ namespace symint {
52
+ template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
53
+ at::Tensor _efficientzerotensor(c10::SymIntArrayRef size, at::TensorOptions options={}) {
54
+ return at::_ops::_efficientzerotensor::call(size, c10::optTypeMetaToScalarType(options.dtype_opt()), options.layout_opt(), options.device_opt(), options.pinned_memory_opt());
55
+ }
56
+ }
57
+
58
+ // aten::_efficientzerotensor(SymInt[] size, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor
59
+ inline at::Tensor _efficientzerotensor_symint(c10::SymIntArrayRef size, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory) {
60
+ return at::_ops::_efficientzerotensor::call(size, dtype, layout, device, pin_memory);
61
+ }
62
+ namespace symint {
63
+ template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
64
+ at::Tensor _efficientzerotensor(c10::SymIntArrayRef size, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory) {
65
+ return at::_ops::_efficientzerotensor::call(size, dtype, layout, device, pin_memory);
66
+ }
67
+ }
68
+
69
+ // aten::_efficientzerotensor.out(SymInt[] size, *, Tensor(a!) out) -> Tensor(a!)
70
+ inline at::Tensor & _efficientzerotensor_out(at::Tensor & out, at::IntArrayRef size) {
71
+ return at::_ops::_efficientzerotensor_out::call(c10::fromIntArrayRefSlow(size), out);
72
+ }
73
+ namespace symint {
74
+ template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
75
+ at::Tensor & _efficientzerotensor_out(at::Tensor & out, at::IntArrayRef size) {
76
+ return at::_ops::_efficientzerotensor_out::call(c10::fromIntArrayRefSlow(size), out);
77
+ }
78
+ }
79
+
80
+ // aten::_efficientzerotensor.out(SymInt[] size, *, Tensor(a!) out) -> Tensor(a!)
81
+ inline at::Tensor & _efficientzerotensor_outf(at::IntArrayRef size, at::Tensor & out) {
82
+ return at::_ops::_efficientzerotensor_out::call(c10::fromIntArrayRefSlow(size), out);
83
+ }
84
+ namespace symint {
85
+ template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
86
+ at::Tensor & _efficientzerotensor_outf(at::IntArrayRef size, at::Tensor & out) {
87
+ return at::_ops::_efficientzerotensor_out::call(c10::fromIntArrayRefSlow(size), out);
88
+ }
89
+ }
90
+
91
+ // aten::_efficientzerotensor.out(SymInt[] size, *, Tensor(a!) out) -> Tensor(a!)
92
+ inline at::Tensor & _efficientzerotensor_symint_out(at::Tensor & out, c10::SymIntArrayRef size) {
93
+ return at::_ops::_efficientzerotensor_out::call(size, out);
94
+ }
95
+ namespace symint {
96
+ template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
97
+ at::Tensor & _efficientzerotensor_out(at::Tensor & out, c10::SymIntArrayRef size) {
98
+ return at::_ops::_efficientzerotensor_out::call(size, out);
99
+ }
100
+ }
101
+
102
+ // aten::_efficientzerotensor.out(SymInt[] size, *, Tensor(a!) out) -> Tensor(a!)
103
+ inline at::Tensor & _efficientzerotensor_symint_outf(c10::SymIntArrayRef size, at::Tensor & out) {
104
+ return at::_ops::_efficientzerotensor_out::call(size, out);
105
+ }
106
+ namespace symint {
107
+ template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
108
+ at::Tensor & _efficientzerotensor_outf(c10::SymIntArrayRef size, at::Tensor & out) {
109
+ return at::_ops::_efficientzerotensor_out::call(size, out);
110
+ }
111
+ }
112
+
113
+ }
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_embedding_bag_forward_only_compositeexplicitautograd_dispatch.h ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace compositeexplicitautograd {
19
+
20
+ TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> _embedding_bag_forward_only_out(at::Tensor & out0, at::Tensor & out1, at::Tensor & out2, at::Tensor & out3, const at::Tensor & weight, const at::Tensor & indices, const at::Tensor & offsets, bool scale_grad_by_freq=false, int64_t mode=0, bool sparse=false, const ::std::optional<at::Tensor> & per_sample_weights={}, bool include_last_offset=false, int64_t padding_idx=-1);
21
+ TORCH_API ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &,at::Tensor &> _embedding_bag_forward_only_outf(const at::Tensor & weight, const at::Tensor & indices, const at::Tensor & offsets, bool scale_grad_by_freq, int64_t mode, bool sparse, const ::std::optional<at::Tensor> & per_sample_weights, bool include_last_offset, int64_t padding_idx, at::Tensor & out0, at::Tensor & out1, at::Tensor & out2, at::Tensor & out3);
22
+
23
+ } // namespace compositeexplicitautograd
24
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_fake_quantize_learnable_per_channel_affine.h ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Function.h
4
+
5
+ #include <ATen/Context.h>
6
+ #include <ATen/DeviceGuard.h>
7
+ #include <ATen/TensorUtils.h>
8
+ #include <ATen/TracerMode.h>
9
+ #include <ATen/core/Generator.h>
10
+ #include <ATen/core/Reduction.h>
11
+ #include <ATen/core/Tensor.h>
12
+ #include <c10/core/Scalar.h>
13
+ #include <c10/core/Storage.h>
14
+ #include <c10/core/TensorOptions.h>
15
+ #include <c10/util/Deprecated.h>
16
+ #include <optional>
17
+
18
+
19
+
20
+ #include <ATen/ops/_fake_quantize_learnable_per_channel_affine_ops.h>
21
+
22
+ namespace at {
23
+
24
+
25
+ // aten::_fake_quantize_learnable_per_channel_affine(Tensor self, Tensor scale, Tensor zero_point, int axis, int quant_min, int quant_max, float grad_factor=1.0) -> Tensor
26
+ inline 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) {
27
+ return at::_ops::_fake_quantize_learnable_per_channel_affine::call(self, scale, zero_point, axis, quant_min, quant_max, grad_factor);
28
+ }
29
+
30
+ // aten::_fake_quantize_learnable_per_channel_affine.out(Tensor self, Tensor scale, Tensor zero_point, int axis, int quant_min, int quant_max, float grad_factor=1.0, *, Tensor(a!) out) -> Tensor(a!)
31
+ inline at::Tensor & _fake_quantize_learnable_per_channel_affine_out(at::Tensor & 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=1.0) {
32
+ return at::_ops::_fake_quantize_learnable_per_channel_affine_out::call(self, scale, zero_point, axis, quant_min, quant_max, grad_factor, out);
33
+ }
34
+ // aten::_fake_quantize_learnable_per_channel_affine.out(Tensor self, Tensor scale, Tensor zero_point, int axis, int quant_min, int quant_max, float grad_factor=1.0, *, Tensor(a!) out) -> Tensor(a!)
35
+ inline at::Tensor & _fake_quantize_learnable_per_channel_affine_outf(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) {
36
+ return at::_ops::_fake_quantize_learnable_per_channel_affine_out::call(self, scale, zero_point, axis, quant_min, quant_max, grad_factor, out);
37
+ }
38
+
39
+ }
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_fake_quantize_learnable_per_channel_affine_cpu_dispatch.h ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace cpu {
19
+
20
+ 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);
21
+
22
+ } // namespace cpu
23
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_foreach_tanh_compositeexplicitautograd_dispatch.h ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace compositeexplicitautograd {
19
+
20
+ TORCH_API ::std::vector<at::Tensor> _foreach_tanh(at::TensorList self);
21
+ TORCH_API void _foreach_tanh_out(at::TensorList out, at::TensorList self);
22
+ TORCH_API void _foreach_tanh_outf(at::TensorList self, at::TensorList out);
23
+ TORCH_API void _foreach_tanh_(at::TensorList self);
24
+
25
+ } // namespace compositeexplicitautograd
26
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_histogramdd_from_bin_tensors_compositeexplicitautograd_dispatch.h ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace compositeexplicitautograd {
19
+
20
+ TORCH_API at::Tensor & _histogramdd_from_bin_tensors_out(at::Tensor & out, const at::Tensor & self, at::TensorList bins, const ::std::optional<at::Tensor> & weight={}, bool density=false);
21
+ TORCH_API at::Tensor & _histogramdd_from_bin_tensors_outf(const at::Tensor & self, at::TensorList bins, const ::std::optional<at::Tensor> & weight, bool density, at::Tensor & out);
22
+
23
+ } // namespace compositeexplicitautograd
24
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_linalg_svd_native.h ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from NativeFunction.h
4
+
5
+ #include <c10/core/Scalar.h>
6
+ #include <c10/core/Storage.h>
7
+ #include <c10/core/TensorOptions.h>
8
+ #include <c10/util/Deprecated.h>
9
+ #include <optional>
10
+ #include <c10/core/QScheme.h>
11
+ #include <ATen/core/Reduction.h>
12
+ #include <ATen/core/Tensor.h>
13
+ #include <tuple>
14
+ #include <vector>
15
+ #include <ATen/ops/_linalg_svd_meta.h>
16
+
17
+ namespace at {
18
+ namespace native {
19
+ struct TORCH_API structured__linalg_svd_out : public at::meta::structured__linalg_svd {
20
+ void impl(const at::Tensor & A, bool full_matrices, bool compute_uv, ::std::optional<c10::string_view> driver, const at::Tensor & U, const at::Tensor & S, const at::Tensor & Vh);
21
+ };
22
+ } // namespace native
23
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_local_scalar_dense.h ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Function.h
4
+
5
+ #include <ATen/Context.h>
6
+ #include <ATen/DeviceGuard.h>
7
+ #include <ATen/TensorUtils.h>
8
+ #include <ATen/TracerMode.h>
9
+ #include <ATen/core/Generator.h>
10
+ #include <ATen/core/Reduction.h>
11
+ #include <ATen/core/Tensor.h>
12
+ #include <c10/core/Scalar.h>
13
+ #include <c10/core/Storage.h>
14
+ #include <c10/core/TensorOptions.h>
15
+ #include <c10/util/Deprecated.h>
16
+ #include <optional>
17
+
18
+
19
+
20
+ #include <ATen/ops/_local_scalar_dense_ops.h>
21
+
22
+ namespace at {
23
+
24
+
25
+ // aten::_local_scalar_dense(Tensor self) -> Scalar
26
+ inline at::Scalar _local_scalar_dense(const at::Tensor & self) {
27
+ return at::_ops::_local_scalar_dense::call(self);
28
+ }
29
+
30
+ }
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_mixed_dtypes_linear_ops.h ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Operator.h
4
+
5
+ #include <tuple>
6
+ #include <vector>
7
+
8
+ // Forward declarations of any types needed in the operator signatures.
9
+ // We can't directly include these classes because it will cause circular include dependencies.
10
+ // This file is included by TensorBody.h, which defines the Tensor class.
11
+ #include <ATen/core/ATen_fwd.h>
12
+
13
+ namespace at {
14
+ namespace _ops {
15
+
16
+
17
+ struct TORCH_API _mixed_dtypes_linear {
18
+ using schema = at::Tensor (const at::Tensor &, const at::Tensor &, const at::Tensor &, const ::std::optional<at::Tensor> &, ::std::optional<c10::string_view>);
19
+ using ptr_schema = schema*;
20
+ // See Note [static constexpr char* members for windows NVCC]
21
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_mixed_dtypes_linear")
22
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
23
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_mixed_dtypes_linear(Tensor input, Tensor weight, Tensor scale, *, Tensor? bias=None, str? activation=None) -> Tensor")
24
+ static at::Tensor call(const at::Tensor & input, const at::Tensor & weight, const at::Tensor & scale, const ::std::optional<at::Tensor> & bias, ::std::optional<c10::string_view> activation);
25
+ static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & input, const at::Tensor & weight, const at::Tensor & scale, const ::std::optional<at::Tensor> & bias, ::std::optional<c10::string_view> activation);
26
+ };
27
+
28
+ }} // namespace at::_ops
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_scaled_dot_product_flash_attention_backward.h ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Function.h
4
+
5
+ #include <ATen/Context.h>
6
+ #include <ATen/DeviceGuard.h>
7
+ #include <ATen/TensorUtils.h>
8
+ #include <ATen/TracerMode.h>
9
+ #include <ATen/core/Generator.h>
10
+ #include <ATen/core/Reduction.h>
11
+ #include <ATen/core/Tensor.h>
12
+ #include <c10/core/Scalar.h>
13
+ #include <c10/core/Storage.h>
14
+ #include <c10/core/TensorOptions.h>
15
+ #include <c10/util/Deprecated.h>
16
+ #include <optional>
17
+
18
+
19
+
20
+ #include <ATen/ops/_scaled_dot_product_flash_attention_backward_ops.h>
21
+
22
+ namespace at {
23
+
24
+
25
+ // aten::_scaled_dot_product_flash_attention_backward(Tensor grad_out, Tensor query, Tensor key, Tensor value, Tensor out, Tensor logsumexp, Tensor cum_seq_q, Tensor cum_seq_k, SymInt max_q, SymInt max_k, float dropout_p, bool is_causal, Tensor philox_seed, Tensor philox_offset, *, float? scale=None) -> (Tensor grad_query, Tensor grad_key, Tensor grad_value)
26
+ inline ::std::tuple<at::Tensor,at::Tensor,at::Tensor> _scaled_dot_product_flash_attention_backward(const at::Tensor & grad_out, const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const at::Tensor & out, const at::Tensor & logsumexp, const at::Tensor & cum_seq_q, const at::Tensor & cum_seq_k, int64_t max_q, int64_t max_k, double dropout_p, bool is_causal, const at::Tensor & philox_seed, const at::Tensor & philox_offset, ::std::optional<double> scale=::std::nullopt) {
27
+ return at::_ops::_scaled_dot_product_flash_attention_backward::call(grad_out, query, key, value, out, logsumexp, cum_seq_q, cum_seq_k, max_q, max_k, dropout_p, is_causal, philox_seed, philox_offset, scale);
28
+ }
29
+ namespace symint {
30
+ template <typename T, typename = std::enable_if_t<std::is_same<T, int64_t>::value>>
31
+ ::std::tuple<at::Tensor,at::Tensor,at::Tensor> _scaled_dot_product_flash_attention_backward(const at::Tensor & grad_out, const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const at::Tensor & out, const at::Tensor & logsumexp, const at::Tensor & cum_seq_q, const at::Tensor & cum_seq_k, int64_t max_q, int64_t max_k, double dropout_p, bool is_causal, const at::Tensor & philox_seed, const at::Tensor & philox_offset, ::std::optional<double> scale=::std::nullopt) {
32
+ return at::_ops::_scaled_dot_product_flash_attention_backward::call(grad_out, query, key, value, out, logsumexp, cum_seq_q, cum_seq_k, max_q, max_k, dropout_p, is_causal, philox_seed, philox_offset, scale);
33
+ }
34
+ }
35
+
36
+ // aten::_scaled_dot_product_flash_attention_backward(Tensor grad_out, Tensor query, Tensor key, Tensor value, Tensor out, Tensor logsumexp, Tensor cum_seq_q, Tensor cum_seq_k, SymInt max_q, SymInt max_k, float dropout_p, bool is_causal, Tensor philox_seed, Tensor philox_offset, *, float? scale=None) -> (Tensor grad_query, Tensor grad_key, Tensor grad_value)
37
+ inline ::std::tuple<at::Tensor,at::Tensor,at::Tensor> _scaled_dot_product_flash_attention_backward_symint(const at::Tensor & grad_out, const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const at::Tensor & out, const at::Tensor & logsumexp, const at::Tensor & cum_seq_q, const at::Tensor & cum_seq_k, c10::SymInt max_q, c10::SymInt max_k, double dropout_p, bool is_causal, const at::Tensor & philox_seed, const at::Tensor & philox_offset, ::std::optional<double> scale=::std::nullopt) {
38
+ return at::_ops::_scaled_dot_product_flash_attention_backward::call(grad_out, query, key, value, out, logsumexp, cum_seq_q, cum_seq_k, max_q, max_k, dropout_p, is_causal, philox_seed, philox_offset, scale);
39
+ }
40
+ namespace symint {
41
+ template <typename T, typename = std::enable_if_t<std::is_same<T, c10::SymInt>::value>>
42
+ ::std::tuple<at::Tensor,at::Tensor,at::Tensor> _scaled_dot_product_flash_attention_backward(const at::Tensor & grad_out, const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const at::Tensor & out, const at::Tensor & logsumexp, const at::Tensor & cum_seq_q, const at::Tensor & cum_seq_k, c10::SymInt max_q, c10::SymInt max_k, double dropout_p, bool is_causal, const at::Tensor & philox_seed, const at::Tensor & philox_offset, ::std::optional<double> scale=::std::nullopt) {
43
+ return at::_ops::_scaled_dot_product_flash_attention_backward::call(grad_out, query, key, value, out, logsumexp, cum_seq_q, cum_seq_k, max_q, max_k, dropout_p, is_causal, philox_seed, philox_offset, scale);
44
+ }
45
+ }
46
+
47
+ }
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_sobol_engine_initialize_state_ops.h ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Operator.h
4
+
5
+ #include <tuple>
6
+ #include <vector>
7
+
8
+ // Forward declarations of any types needed in the operator signatures.
9
+ // We can't directly include these classes because it will cause circular include dependencies.
10
+ // This file is included by TensorBody.h, which defines the Tensor class.
11
+ #include <ATen/core/ATen_fwd.h>
12
+
13
+ namespace at {
14
+ namespace _ops {
15
+
16
+
17
+ struct TORCH_API _sobol_engine_initialize_state_ {
18
+ using schema = at::Tensor & (at::Tensor &, int64_t);
19
+ using ptr_schema = schema*;
20
+ // See Note [static constexpr char* members for windows NVCC]
21
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_sobol_engine_initialize_state_")
22
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
23
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_sobol_engine_initialize_state_(Tensor(a!) self, int dimension) -> Tensor(a!)")
24
+ static at::Tensor & call(at::Tensor & self, int64_t dimension);
25
+ static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, at::Tensor & self, int64_t dimension);
26
+ };
27
+
28
+ }} // namespace at::_ops
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_sparse_log_softmax_ops.h ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Operator.h
4
+
5
+ #include <tuple>
6
+ #include <vector>
7
+
8
+ // Forward declarations of any types needed in the operator signatures.
9
+ // We can't directly include these classes because it will cause circular include dependencies.
10
+ // This file is included by TensorBody.h, which defines the Tensor class.
11
+ #include <ATen/core/ATen_fwd.h>
12
+
13
+ namespace at {
14
+ namespace _ops {
15
+
16
+
17
+ struct TORCH_API _sparse_log_softmax_int {
18
+ using schema = at::Tensor (const at::Tensor &, int64_t, ::std::optional<at::ScalarType>);
19
+ using ptr_schema = schema*;
20
+ // See Note [static constexpr char* members for windows NVCC]
21
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_sparse_log_softmax")
22
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "int")
23
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_sparse_log_softmax.int(Tensor self, int dim, ScalarType? dtype=None) -> Tensor")
24
+ static at::Tensor call(const at::Tensor & self, int64_t dim, ::std::optional<at::ScalarType> dtype);
25
+ static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, int64_t dim, ::std::optional<at::ScalarType> dtype);
26
+ };
27
+
28
+ struct TORCH_API _sparse_log_softmax_Dimname {
29
+ using schema = at::Tensor (const at::Tensor &, at::Dimname, ::std::optional<at::ScalarType>);
30
+ using ptr_schema = schema*;
31
+ // See Note [static constexpr char* members for windows NVCC]
32
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_sparse_log_softmax")
33
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "Dimname")
34
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_sparse_log_softmax.Dimname(Tensor self, Dimname dim, *, ScalarType? dtype=None) -> Tensor")
35
+ static at::Tensor call(const at::Tensor & self, at::Dimname dim, ::std::optional<at::ScalarType> dtype);
36
+ static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, at::Dimname dim, ::std::optional<at::ScalarType> dtype);
37
+ };
38
+
39
+ struct TORCH_API _sparse_log_softmax {
40
+ using schema = at::Tensor (const at::Tensor &, int64_t, bool);
41
+ using ptr_schema = schema*;
42
+ // See Note [static constexpr char* members for windows NVCC]
43
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_sparse_log_softmax")
44
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
45
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_sparse_log_softmax(Tensor self, int dim, bool half_to_float) -> Tensor")
46
+ static at::Tensor call(const at::Tensor & self, int64_t dim, bool half_to_float);
47
+ static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, int64_t dim, bool half_to_float);
48
+ };
49
+
50
+ struct TORCH_API _sparse_log_softmax_out {
51
+ using schema = at::Tensor & (const at::Tensor &, int64_t, bool, at::Tensor &);
52
+ using ptr_schema = schema*;
53
+ // See Note [static constexpr char* members for windows NVCC]
54
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_sparse_log_softmax")
55
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
56
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_sparse_log_softmax.out(Tensor self, int dim, bool half_to_float, *, Tensor(a!) out) -> Tensor(a!)")
57
+ static at::Tensor & call(const at::Tensor & self, int64_t dim, bool half_to_float, at::Tensor & out);
58
+ static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, int64_t dim, bool half_to_float, at::Tensor & out);
59
+ };
60
+
61
+ }} // namespace at::_ops
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_sparse_softmax.h ADDED
@@ -0,0 +1,49 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Function.h
4
+
5
+ #include <ATen/Context.h>
6
+ #include <ATen/DeviceGuard.h>
7
+ #include <ATen/TensorUtils.h>
8
+ #include <ATen/TracerMode.h>
9
+ #include <ATen/core/Generator.h>
10
+ #include <ATen/core/Reduction.h>
11
+ #include <ATen/core/Tensor.h>
12
+ #include <c10/core/Scalar.h>
13
+ #include <c10/core/Storage.h>
14
+ #include <c10/core/TensorOptions.h>
15
+ #include <c10/util/Deprecated.h>
16
+ #include <optional>
17
+
18
+
19
+
20
+ #include <ATen/ops/_sparse_softmax_ops.h>
21
+
22
+ namespace at {
23
+
24
+
25
+ // aten::_sparse_softmax.int(Tensor self, int dim, ScalarType? dtype=None) -> Tensor
26
+ inline at::Tensor _sparse_softmax(const at::Tensor & self, int64_t dim, ::std::optional<at::ScalarType> dtype=::std::nullopt) {
27
+ return at::_ops::_sparse_softmax_int::call(self, dim, dtype);
28
+ }
29
+
30
+ // aten::_sparse_softmax.Dimname(Tensor self, Dimname dim, *, ScalarType? dtype=None) -> Tensor
31
+ inline at::Tensor _sparse_softmax(const at::Tensor & self, at::Dimname dim, ::std::optional<at::ScalarType> dtype=::std::nullopt) {
32
+ return at::_ops::_sparse_softmax_Dimname::call(self, dim, dtype);
33
+ }
34
+
35
+ // aten::_sparse_softmax(Tensor self, int dim, bool half_to_float) -> Tensor
36
+ inline at::Tensor _sparse_softmax(const at::Tensor & self, int64_t dim, bool half_to_float) {
37
+ return at::_ops::_sparse_softmax::call(self, dim, half_to_float);
38
+ }
39
+
40
+ // aten::_sparse_softmax.out(Tensor self, int dim, bool half_to_float, *, Tensor(a!) out) -> Tensor(a!)
41
+ inline at::Tensor & _sparse_softmax_out(at::Tensor & out, const at::Tensor & self, int64_t dim, bool half_to_float) {
42
+ return at::_ops::_sparse_softmax_out::call(self, dim, half_to_float, out);
43
+ }
44
+ // aten::_sparse_softmax.out(Tensor self, int dim, bool half_to_float, *, Tensor(a!) out) -> Tensor(a!)
45
+ inline at::Tensor & _sparse_softmax_outf(const at::Tensor & self, int64_t dim, bool half_to_float, at::Tensor & out) {
46
+ return at::_ops::_sparse_softmax_out::call(self, dim, half_to_float, out);
47
+ }
48
+
49
+ }
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_unpack_dual_ops.h ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Operator.h
4
+
5
+ #include <tuple>
6
+ #include <vector>
7
+
8
+ // Forward declarations of any types needed in the operator signatures.
9
+ // We can't directly include these classes because it will cause circular include dependencies.
10
+ // This file is included by TensorBody.h, which defines the Tensor class.
11
+ #include <ATen/core/ATen_fwd.h>
12
+
13
+ namespace at {
14
+ namespace _ops {
15
+
16
+
17
+ struct TORCH_API _unpack_dual {
18
+ using schema = ::std::tuple<at::Tensor,at::Tensor> (const at::Tensor &, int64_t);
19
+ using ptr_schema = schema*;
20
+ // See Note [static constexpr char* members for windows NVCC]
21
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_unpack_dual")
22
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
23
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_unpack_dual(Tensor(a) dual, int level) -> (Tensor(a) primal, Tensor tangent)")
24
+ static ::std::tuple<at::Tensor,at::Tensor> call(const at::Tensor & dual, int64_t level);
25
+ static ::std::tuple<at::Tensor,at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & dual, int64_t level);
26
+ };
27
+
28
+ }} // namespace at::_ops
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/_weight_int4pack_mm_ops.h ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Operator.h
4
+
5
+ #include <tuple>
6
+ #include <vector>
7
+
8
+ // Forward declarations of any types needed in the operator signatures.
9
+ // We can't directly include these classes because it will cause circular include dependencies.
10
+ // This file is included by TensorBody.h, which defines the Tensor class.
11
+ #include <ATen/core/ATen_fwd.h>
12
+
13
+ namespace at {
14
+ namespace _ops {
15
+
16
+
17
+ struct TORCH_API _weight_int4pack_mm {
18
+ using schema = at::Tensor (const at::Tensor &, const at::Tensor &, int64_t, const at::Tensor &);
19
+ using ptr_schema = schema*;
20
+ // See Note [static constexpr char* members for windows NVCC]
21
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::_weight_int4pack_mm")
22
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
23
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "_weight_int4pack_mm(Tensor self, Tensor mat2, int qGroupSize, Tensor qScaleAndZeros) -> Tensor")
24
+ static at::Tensor call(const at::Tensor & self, const at::Tensor & mat2, int64_t qGroupSize, const at::Tensor & qScaleAndZeros);
25
+ static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Tensor & mat2, int64_t qGroupSize, const at::Tensor & qScaleAndZeros);
26
+ };
27
+
28
+ }} // namespace at::_ops
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/abs.h ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Function.h
4
+
5
+ #include <ATen/Context.h>
6
+ #include <ATen/DeviceGuard.h>
7
+ #include <ATen/TensorUtils.h>
8
+ #include <ATen/TracerMode.h>
9
+ #include <ATen/core/Generator.h>
10
+ #include <ATen/core/Reduction.h>
11
+ #include <ATen/core/Tensor.h>
12
+ #include <c10/core/Scalar.h>
13
+ #include <c10/core/Storage.h>
14
+ #include <c10/core/TensorOptions.h>
15
+ #include <c10/util/Deprecated.h>
16
+ #include <optional>
17
+
18
+
19
+
20
+ #include <ATen/ops/abs_ops.h>
21
+
22
+ namespace at {
23
+
24
+
25
+ // aten::abs(Tensor self) -> Tensor
26
+ inline at::Tensor abs(const at::Tensor & self) {
27
+ return at::_ops::abs::call(self);
28
+ }
29
+
30
+ // aten::abs_(Tensor(a!) self) -> Tensor(a!)
31
+ inline at::Tensor & abs_(at::Tensor & self) {
32
+ return at::_ops::abs_::call(self);
33
+ }
34
+
35
+ // aten::abs.out(Tensor self, *, Tensor(a!) out) -> Tensor(a!)
36
+ inline at::Tensor & abs_out(at::Tensor & out, const at::Tensor & self) {
37
+ return at::_ops::abs_out::call(self, out);
38
+ }
39
+ // aten::abs.out(Tensor self, *, Tensor(a!) out) -> Tensor(a!)
40
+ inline at::Tensor & abs_outf(const at::Tensor & self, at::Tensor & out) {
41
+ return at::_ops::abs_out::call(self, out);
42
+ }
43
+
44
+ }
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/adaptive_max_pool1d_compositeimplicitautograd_dispatch.h ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace compositeimplicitautograd {
19
+
20
+ TORCH_API ::std::tuple<at::Tensor,at::Tensor> adaptive_max_pool1d(const at::Tensor & self, at::IntArrayRef output_size);
21
+
22
+ } // namespace compositeimplicitautograd
23
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/adaptive_max_pool2d_backward.h ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Function.h
4
+
5
+ #include <ATen/Context.h>
6
+ #include <ATen/DeviceGuard.h>
7
+ #include <ATen/TensorUtils.h>
8
+ #include <ATen/TracerMode.h>
9
+ #include <ATen/core/Generator.h>
10
+ #include <ATen/core/Reduction.h>
11
+ #include <ATen/core/Tensor.h>
12
+ #include <c10/core/Scalar.h>
13
+ #include <c10/core/Storage.h>
14
+ #include <c10/core/TensorOptions.h>
15
+ #include <c10/util/Deprecated.h>
16
+ #include <optional>
17
+
18
+
19
+
20
+ #include <ATen/ops/adaptive_max_pool2d_backward_ops.h>
21
+
22
+ namespace at {
23
+
24
+
25
+ // aten::adaptive_max_pool2d_backward.grad_input(Tensor grad_output, Tensor self, Tensor indices, *, Tensor(a!) grad_input) -> Tensor(a!)
26
+ inline at::Tensor & adaptive_max_pool2d_backward_out(at::Tensor & grad_input, const at::Tensor & grad_output, const at::Tensor & self, const at::Tensor & indices) {
27
+ return at::_ops::adaptive_max_pool2d_backward_grad_input::call(grad_output, self, indices, grad_input);
28
+ }
29
+ // aten::adaptive_max_pool2d_backward.grad_input(Tensor grad_output, Tensor self, Tensor indices, *, Tensor(a!) grad_input) -> Tensor(a!)
30
+ inline at::Tensor & adaptive_max_pool2d_backward_outf(const at::Tensor & grad_output, const at::Tensor & self, const at::Tensor & indices, at::Tensor & grad_input) {
31
+ return at::_ops::adaptive_max_pool2d_backward_grad_input::call(grad_output, self, indices, grad_input);
32
+ }
33
+
34
+ // aten::adaptive_max_pool2d_backward(Tensor grad_output, Tensor self, Tensor indices) -> Tensor
35
+ inline at::Tensor adaptive_max_pool2d_backward(const at::Tensor & grad_output, const at::Tensor & self, const at::Tensor & indices) {
36
+ return at::_ops::adaptive_max_pool2d_backward::call(grad_output, self, indices);
37
+ }
38
+
39
+ }
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/addcdiv_meta_dispatch.h ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace meta {
19
+
20
+ TORCH_API at::Tensor addcdiv(const at::Tensor & self, const at::Tensor & tensor1, const at::Tensor & tensor2, const at::Scalar & value=1);
21
+ TORCH_API at::Tensor & addcdiv_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & tensor1, const at::Tensor & tensor2, const at::Scalar & value=1);
22
+ TORCH_API at::Tensor & addcdiv_outf(const at::Tensor & self, const at::Tensor & tensor1, const at::Tensor & tensor2, const at::Scalar & value, at::Tensor & out);
23
+ TORCH_API at::Tensor & addcdiv_(at::Tensor & self, const at::Tensor & tensor1, const at::Tensor & tensor2, const at::Scalar & value=1);
24
+
25
+ } // namespace meta
26
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/batch_norm_gather_stats_native.h ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from NativeFunction.h
4
+
5
+ #include <c10/core/Scalar.h>
6
+ #include <c10/core/Storage.h>
7
+ #include <c10/core/TensorOptions.h>
8
+ #include <c10/util/Deprecated.h>
9
+ #include <optional>
10
+ #include <c10/core/QScheme.h>
11
+ #include <ATen/core/Reduction.h>
12
+ #include <ATen/core/Tensor.h>
13
+ #include <tuple>
14
+ #include <vector>
15
+
16
+
17
+ namespace at {
18
+ namespace native {
19
+ TORCH_API ::std::tuple<at::Tensor &,at::Tensor &> batch_norm_gather_stats_out(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);
20
+ TORCH_API ::std::tuple<at::Tensor,at::Tensor> batch_norm_gather_stats_cuda(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);
21
+ } // namespace native
22
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/bilinear_native.h ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from NativeFunction.h
4
+
5
+ #include <c10/core/Scalar.h>
6
+ #include <c10/core/Storage.h>
7
+ #include <c10/core/TensorOptions.h>
8
+ #include <c10/util/Deprecated.h>
9
+ #include <optional>
10
+ #include <c10/core/QScheme.h>
11
+ #include <ATen/core/Reduction.h>
12
+ #include <ATen/core/Tensor.h>
13
+ #include <tuple>
14
+ #include <vector>
15
+
16
+
17
+ namespace at {
18
+ namespace native {
19
+ TORCH_API at::Tensor bilinear(const at::Tensor & input1, const at::Tensor & input2, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias={});
20
+ } // namespace native
21
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/bitwise_or_cuda_dispatch.h ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace cuda {
19
+
20
+ TORCH_API at::Tensor bitwise_or(const at::Tensor & self, const at::Tensor & other);
21
+ TORCH_API at::Tensor & bitwise_or_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & other);
22
+ TORCH_API at::Tensor & bitwise_or_outf(const at::Tensor & self, const at::Tensor & other, at::Tensor & out);
23
+ TORCH_API at::Tensor & bitwise_or_(at::Tensor & self, const at::Tensor & other);
24
+
25
+ } // namespace cuda
26
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/conv1d_ops.h ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Operator.h
4
+
5
+ #include <tuple>
6
+ #include <vector>
7
+
8
+ // Forward declarations of any types needed in the operator signatures.
9
+ // We can't directly include these classes because it will cause circular include dependencies.
10
+ // This file is included by TensorBody.h, which defines the Tensor class.
11
+ #include <ATen/core/ATen_fwd.h>
12
+
13
+ namespace at {
14
+ namespace _ops {
15
+
16
+
17
+ struct TORCH_API conv1d {
18
+ using schema = at::Tensor (const at::Tensor &, const at::Tensor &, const ::std::optional<at::Tensor> &, c10::SymIntArrayRef, c10::SymIntArrayRef, c10::SymIntArrayRef, c10::SymInt);
19
+ using ptr_schema = schema*;
20
+ // See Note [static constexpr char* members for windows NVCC]
21
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::conv1d")
22
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
23
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "conv1d(Tensor input, Tensor weight, Tensor? bias=None, SymInt[1] stride=1, SymInt[1] padding=0, SymInt[1] dilation=1, SymInt groups=1) -> Tensor")
24
+ static at::Tensor call(const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::SymIntArrayRef padding, c10::SymIntArrayRef dilation, c10::SymInt groups);
25
+ static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::SymIntArrayRef padding, c10::SymIntArrayRef dilation, c10::SymInt groups);
26
+ };
27
+
28
+ struct TORCH_API conv1d_padding {
29
+ using schema = at::Tensor (const at::Tensor &, const at::Tensor &, const ::std::optional<at::Tensor> &, c10::SymIntArrayRef, c10::string_view, c10::SymIntArrayRef, c10::SymInt);
30
+ using ptr_schema = schema*;
31
+ // See Note [static constexpr char* members for windows NVCC]
32
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::conv1d")
33
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "padding")
34
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "conv1d.padding(Tensor input, Tensor weight, Tensor? bias=None, SymInt[1] stride=1, str padding=\"valid\", SymInt[1] dilation=1, SymInt groups=1) -> Tensor")
35
+ static at::Tensor call(const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::string_view padding, c10::SymIntArrayRef dilation, c10::SymInt groups);
36
+ static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::string_view padding, c10::SymIntArrayRef dilation, c10::SymInt groups);
37
+ };
38
+
39
+ }} // namespace at::_ops
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/gather_backward_ops.h ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Operator.h
4
+
5
+ #include <tuple>
6
+ #include <vector>
7
+
8
+ // Forward declarations of any types needed in the operator signatures.
9
+ // We can't directly include these classes because it will cause circular include dependencies.
10
+ // This file is included by TensorBody.h, which defines the Tensor class.
11
+ #include <ATen/core/ATen_fwd.h>
12
+
13
+ namespace at {
14
+ namespace _ops {
15
+
16
+
17
+ struct TORCH_API gather_backward {
18
+ using schema = at::Tensor (const at::Tensor &, const at::Tensor &, int64_t, const at::Tensor &, bool);
19
+ using ptr_schema = schema*;
20
+ // See Note [static constexpr char* members for windows NVCC]
21
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::gather_backward")
22
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
23
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "gather_backward(Tensor grad, Tensor self, int dim, Tensor index, bool sparse_grad) -> Tensor")
24
+ static at::Tensor call(const at::Tensor & grad, const at::Tensor & self, int64_t dim, const at::Tensor & index, bool sparse_grad);
25
+ static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & grad, const at::Tensor & self, int64_t dim, const at::Tensor & index, bool sparse_grad);
26
+ };
27
+
28
+ }} // namespace at::_ops
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/hardswish_cpu_dispatch.h ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace cpu {
19
+
20
+ TORCH_API at::Tensor hardswish(const at::Tensor & self);
21
+ TORCH_API at::Tensor & hardswish_out(at::Tensor & out, const at::Tensor & self);
22
+ TORCH_API at::Tensor & hardswish_outf(const at::Tensor & self, at::Tensor & out);
23
+ TORCH_API at::Tensor & hardswish_(at::Tensor & self);
24
+
25
+ } // namespace cpu
26
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/hspmm_ops.h ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Operator.h
4
+
5
+ #include <tuple>
6
+ #include <vector>
7
+
8
+ // Forward declarations of any types needed in the operator signatures.
9
+ // We can't directly include these classes because it will cause circular include dependencies.
10
+ // This file is included by TensorBody.h, which defines the Tensor class.
11
+ #include <ATen/core/ATen_fwd.h>
12
+
13
+ namespace at {
14
+ namespace _ops {
15
+
16
+
17
+ struct TORCH_API hspmm_out {
18
+ using schema = at::Tensor & (const at::Tensor &, const at::Tensor &, at::Tensor &);
19
+ using ptr_schema = schema*;
20
+ // See Note [static constexpr char* members for windows NVCC]
21
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::hspmm")
22
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
23
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "hspmm.out(Tensor mat1, Tensor mat2, *, Tensor(a!) out) -> Tensor(a!)")
24
+ static at::Tensor & call(const at::Tensor & mat1, const at::Tensor & mat2, at::Tensor & out);
25
+ static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & mat1, const at::Tensor & mat2, at::Tensor & out);
26
+ };
27
+
28
+ struct TORCH_API hspmm {
29
+ using schema = at::Tensor (const at::Tensor &, const at::Tensor &);
30
+ using ptr_schema = schema*;
31
+ // See Note [static constexpr char* members for windows NVCC]
32
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::hspmm")
33
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
34
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "hspmm(Tensor mat1, Tensor mat2) -> Tensor")
35
+ static at::Tensor call(const at::Tensor & mat1, const at::Tensor & mat2);
36
+ static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & mat1, const at::Tensor & mat2);
37
+ };
38
+
39
+ }} // namespace at::_ops
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/hypot_meta_dispatch.h ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace meta {
19
+
20
+ TORCH_API at::Tensor hypot(const at::Tensor & self, const at::Tensor & other);
21
+ TORCH_API at::Tensor & hypot_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & other);
22
+ TORCH_API at::Tensor & hypot_outf(const at::Tensor & self, const at::Tensor & other, at::Tensor & out);
23
+ TORCH_API at::Tensor & hypot_(at::Tensor & self, const at::Tensor & other);
24
+
25
+ } // namespace meta
26
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/inverse_ops.h ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Operator.h
4
+
5
+ #include <tuple>
6
+ #include <vector>
7
+
8
+ // Forward declarations of any types needed in the operator signatures.
9
+ // We can't directly include these classes because it will cause circular include dependencies.
10
+ // This file is included by TensorBody.h, which defines the Tensor class.
11
+ #include <ATen/core/ATen_fwd.h>
12
+
13
+ namespace at {
14
+ namespace _ops {
15
+
16
+
17
+ struct TORCH_API inverse {
18
+ using schema = at::Tensor (const at::Tensor &);
19
+ using ptr_schema = schema*;
20
+ // See Note [static constexpr char* members for windows NVCC]
21
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::inverse")
22
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
23
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "inverse(Tensor self) -> Tensor")
24
+ static at::Tensor call(const at::Tensor & self);
25
+ static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self);
26
+ };
27
+
28
+ struct TORCH_API inverse_out {
29
+ using schema = at::Tensor & (const at::Tensor &, at::Tensor &);
30
+ using ptr_schema = schema*;
31
+ // See Note [static constexpr char* members for windows NVCC]
32
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::inverse")
33
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
34
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "inverse.out(Tensor self, *, Tensor(a!) out) -> Tensor(a!)")
35
+ static at::Tensor & call(const at::Tensor & self, at::Tensor & out);
36
+ static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, at::Tensor & out);
37
+ };
38
+
39
+ }} // namespace at::_ops
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/isneginf_cuda_dispatch.h ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace cuda {
19
+
20
+ TORCH_API at::Tensor isneginf(const at::Tensor & self);
21
+ TORCH_API at::Tensor & isneginf_out(at::Tensor & out, const at::Tensor & self);
22
+ TORCH_API at::Tensor & isneginf_outf(const at::Tensor & self, at::Tensor & out);
23
+
24
+ } // namespace cuda
25
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_lu_factor_compositeimplicitautograd_dispatch.h ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace compositeimplicitautograd {
19
+
20
+ TORCH_API ::std::tuple<at::Tensor,at::Tensor> linalg_lu_factor(const at::Tensor & A, bool pivot=true);
21
+ TORCH_API ::std::tuple<at::Tensor &,at::Tensor &> linalg_lu_factor_out(at::Tensor & LU, at::Tensor & pivots, const at::Tensor & A, bool pivot=true);
22
+ TORCH_API ::std::tuple<at::Tensor &,at::Tensor &> linalg_lu_factor_outf(const at::Tensor & A, bool pivot, at::Tensor & LU, at::Tensor & pivots);
23
+
24
+ } // namespace compositeimplicitautograd
25
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_matrix_exp_cuda_dispatch.h ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace cuda {
19
+
20
+ TORCH_API at::Tensor linalg_matrix_exp(const at::Tensor & self);
21
+
22
+ } // namespace cuda
23
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_slogdet_compositeimplicitautograd_dispatch.h ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace compositeimplicitautograd {
19
+
20
+ TORCH_API ::std::tuple<at::Tensor,at::Tensor> linalg_slogdet(const at::Tensor & A);
21
+ TORCH_API ::std::tuple<at::Tensor &,at::Tensor &> linalg_slogdet_out(at::Tensor & sign, at::Tensor & logabsdet, const at::Tensor & A);
22
+ TORCH_API ::std::tuple<at::Tensor &,at::Tensor &> linalg_slogdet_outf(const at::Tensor & A, at::Tensor & sign, at::Tensor & logabsdet);
23
+
24
+ } // namespace compositeimplicitautograd
25
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/linalg_solve_triangular_cuda_dispatch.h ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace cuda {
19
+
20
+ TORCH_API at::Tensor linalg_solve_triangular(const at::Tensor & self, const at::Tensor & B, bool upper, bool left=true, bool unitriangular=false);
21
+ TORCH_API at::Tensor & linalg_solve_triangular_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & B, bool upper, bool left=true, bool unitriangular=false);
22
+ TORCH_API at::Tensor & linalg_solve_triangular_outf(const at::Tensor & self, const at::Tensor & B, bool upper, bool left, bool unitriangular, at::Tensor & out);
23
+
24
+ } // namespace cuda
25
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/lu_unpack.h ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Function.h
4
+
5
+ #include <ATen/Context.h>
6
+ #include <ATen/DeviceGuard.h>
7
+ #include <ATen/TensorUtils.h>
8
+ #include <ATen/TracerMode.h>
9
+ #include <ATen/core/Generator.h>
10
+ #include <ATen/core/Reduction.h>
11
+ #include <ATen/core/Tensor.h>
12
+ #include <c10/core/Scalar.h>
13
+ #include <c10/core/Storage.h>
14
+ #include <c10/core/TensorOptions.h>
15
+ #include <c10/util/Deprecated.h>
16
+ #include <optional>
17
+
18
+
19
+
20
+ #include <ATen/ops/lu_unpack_ops.h>
21
+
22
+ namespace at {
23
+
24
+
25
+ // aten::lu_unpack(Tensor LU_data, Tensor LU_pivots, bool unpack_data=True, bool unpack_pivots=True) -> (Tensor P, Tensor L, Tensor U)
26
+ inline ::std::tuple<at::Tensor,at::Tensor,at::Tensor> lu_unpack(const at::Tensor & LU_data, const at::Tensor & LU_pivots, bool unpack_data=true, bool unpack_pivots=true) {
27
+ return at::_ops::lu_unpack::call(LU_data, LU_pivots, unpack_data, unpack_pivots);
28
+ }
29
+
30
+ // aten::lu_unpack.out(Tensor LU_data, Tensor LU_pivots, bool unpack_data=True, bool unpack_pivots=True, *, Tensor(a!) P, Tensor(b!) L, Tensor(c!) U) -> (Tensor(a!) P, Tensor(b!) L, Tensor(c!) U)
31
+ inline ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> lu_unpack_out(at::Tensor & P, at::Tensor & L, at::Tensor & U, const at::Tensor & LU_data, const at::Tensor & LU_pivots, bool unpack_data=true, bool unpack_pivots=true) {
32
+ return at::_ops::lu_unpack_out::call(LU_data, LU_pivots, unpack_data, unpack_pivots, P, L, U);
33
+ }
34
+ // aten::lu_unpack.out(Tensor LU_data, Tensor LU_pivots, bool unpack_data=True, bool unpack_pivots=True, *, Tensor(a!) P, Tensor(b!) L, Tensor(c!) U) -> (Tensor(a!) P, Tensor(b!) L, Tensor(c!) U)
35
+ inline ::std::tuple<at::Tensor &,at::Tensor &,at::Tensor &> lu_unpack_outf(const at::Tensor & LU_data, const at::Tensor & LU_pivots, bool unpack_data, bool unpack_pivots, at::Tensor & P, at::Tensor & L, at::Tensor & U) {
36
+ return at::_ops::lu_unpack_out::call(LU_data, LU_pivots, unpack_data, unpack_pivots, P, L, U);
37
+ }
38
+
39
+ }
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/mean_compositeimplicitautograd_dispatch.h ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace compositeimplicitautograd {
19
+
20
+ TORCH_API at::Tensor mean(const at::Tensor & self, at::DimnameList dim, bool keepdim=false, ::std::optional<at::ScalarType> dtype=::std::nullopt);
21
+ TORCH_API at::Tensor & mean_out(at::Tensor & out, const at::Tensor & self, at::DimnameList dim, bool keepdim=false, ::std::optional<at::ScalarType> dtype=::std::nullopt);
22
+ TORCH_API at::Tensor & mean_outf(const at::Tensor & self, at::DimnameList dim, bool keepdim, ::std::optional<at::ScalarType> dtype, at::Tensor & out);
23
+
24
+ } // namespace compositeimplicitautograd
25
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/msort_ops.h ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from Operator.h
4
+
5
+ #include <tuple>
6
+ #include <vector>
7
+
8
+ // Forward declarations of any types needed in the operator signatures.
9
+ // We can't directly include these classes because it will cause circular include dependencies.
10
+ // This file is included by TensorBody.h, which defines the Tensor class.
11
+ #include <ATen/core/ATen_fwd.h>
12
+
13
+ namespace at {
14
+ namespace _ops {
15
+
16
+
17
+ struct TORCH_API msort_out {
18
+ using schema = at::Tensor & (const at::Tensor &, at::Tensor &);
19
+ using ptr_schema = schema*;
20
+ // See Note [static constexpr char* members for windows NVCC]
21
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::msort")
22
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "out")
23
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "msort.out(Tensor self, *, Tensor(a!) out) -> Tensor(a!)")
24
+ static at::Tensor & call(const at::Tensor & self, at::Tensor & out);
25
+ static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, at::Tensor & out);
26
+ };
27
+
28
+ struct TORCH_API msort {
29
+ using schema = at::Tensor (const at::Tensor &);
30
+ using ptr_schema = schema*;
31
+ // See Note [static constexpr char* members for windows NVCC]
32
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(name, "aten::msort")
33
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(overload_name, "")
34
+ STATIC_CONSTEXPR_STR_INL_EXCEPT_WIN_CUDA(schema_str, "msort(Tensor self) -> Tensor")
35
+ static at::Tensor call(const at::Tensor & self);
36
+ static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self);
37
+ };
38
+
39
+ }} // namespace at::_ops
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/multi_margin_loss_native.h ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from NativeFunction.h
4
+
5
+ #include <c10/core/Scalar.h>
6
+ #include <c10/core/Storage.h>
7
+ #include <c10/core/TensorOptions.h>
8
+ #include <c10/util/Deprecated.h>
9
+ #include <optional>
10
+ #include <c10/core/QScheme.h>
11
+ #include <ATen/core/Reduction.h>
12
+ #include <ATen/core/Tensor.h>
13
+ #include <tuple>
14
+ #include <vector>
15
+
16
+
17
+ namespace at {
18
+ namespace native {
19
+ TORCH_API at::Tensor multi_margin_loss_cpu(const at::Tensor & self, const at::Tensor & target, const at::Scalar & p=1, const at::Scalar & margin=1, const ::std::optional<at::Tensor> & weight={}, int64_t reduction=at::Reduction::Mean);
20
+ TORCH_API at::Tensor & multi_margin_loss_cpu_out(const at::Tensor & self, const at::Tensor & target, const at::Scalar & p, const at::Scalar & margin, const ::std::optional<at::Tensor> & weight, int64_t reduction, at::Tensor & out);
21
+ TORCH_API at::Tensor multi_margin_loss_cuda(const at::Tensor & self, const at::Tensor & target, const at::Scalar & p=1, const at::Scalar & margin=1, const ::std::optional<at::Tensor> & weight={}, int64_t reduction=at::Reduction::Mean);
22
+ TORCH_API at::Tensor & multi_margin_loss_cuda_out(const at::Tensor & self, const at::Tensor & target, const at::Scalar & p, const at::Scalar & margin, const ::std::optional<at::Tensor> & weight, int64_t reduction, at::Tensor & out);
23
+ } // namespace native
24
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/nanmedian_cpu_dispatch.h ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace cpu {
19
+
20
+ TORCH_API at::Tensor nanmedian(const at::Tensor & self);
21
+ TORCH_API ::std::tuple<at::Tensor &,at::Tensor &> nanmedian_out(at::Tensor & values, at::Tensor & indices, const at::Tensor & self, int64_t dim, bool keepdim=false);
22
+ TORCH_API ::std::tuple<at::Tensor &,at::Tensor &> nanmedian_outf(const at::Tensor & self, int64_t dim, bool keepdim, at::Tensor & values, at::Tensor & indices);
23
+
24
+ } // namespace cpu
25
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/native_batch_norm_backward_cuda_dispatch.h ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace cuda {
19
+
20
+ TORCH_API ::std::tuple<at::Tensor,at::Tensor,at::Tensor> native_batch_norm_backward(const at::Tensor & grad_out, const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & running_mean, const ::std::optional<at::Tensor> & running_var, const ::std::optional<at::Tensor> & save_mean, const ::std::optional<at::Tensor> & save_invstd, bool train, double eps, ::std::array<bool,3> output_mask);
21
+
22
+ } // namespace cuda
23
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/new_empty_strided_compositeexplicitautogradnonfunctional_dispatch.h ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace compositeexplicitautogradnonfunctional {
19
+
20
+ TORCH_API at::Tensor new_empty_strided(const at::Tensor & self, at::IntArrayRef size, at::IntArrayRef stride, at::TensorOptions options={});
21
+ TORCH_API at::Tensor new_empty_strided(const at::Tensor & self, at::IntArrayRef size, at::IntArrayRef stride, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory);
22
+ TORCH_API at::Tensor new_empty_strided_symint(const at::Tensor & self, c10::SymIntArrayRef size, c10::SymIntArrayRef stride, at::TensorOptions options={});
23
+ TORCH_API at::Tensor new_empty_strided_symint(const at::Tensor & self, c10::SymIntArrayRef size, c10::SymIntArrayRef stride, ::std::optional<at::ScalarType> dtype, ::std::optional<at::Layout> layout, ::std::optional<at::Device> device, ::std::optional<bool> pin_memory);
24
+
25
+ } // namespace compositeexplicitautogradnonfunctional
26
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/one_hot_compositeimplicitautograd_dispatch.h ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+ // @generated by torchgen/gen.py from DispatchKeyFunction.h
3
+
4
+ // NB: The implementing C++ file is RegisterDispatchKey.cpp
5
+
6
+ // The only #includes we need are for custom classes that have defaults in the C++ API
7
+ #include <c10/core/MemoryFormat.h>
8
+ #include <c10/core/Scalar.h>
9
+ #include <ATen/core/Reduction.h>
10
+
11
+ // Forward declarations of any types needed in the operator signatures.
12
+ // We can't directly include these classes because it will cause circular include dependencies.
13
+ // This file is included by TensorBody.h, which defines the Tensor class.
14
+ #include <ATen/core/ATen_fwd.h>
15
+
16
+ namespace at {
17
+
18
+ namespace compositeimplicitautograd {
19
+
20
+ TORCH_API at::Tensor one_hot(const at::Tensor & self, int64_t num_classes=-1);
21
+
22
+ } // namespace compositeimplicitautograd
23
+ } // namespace at
.venv/lib/python3.11/site-packages/torch/include/ATen/ops/one_hot_native.h ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #pragma once
2
+
3
+ // @generated by torchgen/gen.py from NativeFunction.h
4
+
5
+ #include <c10/core/Scalar.h>
6
+ #include <c10/core/Storage.h>
7
+ #include <c10/core/TensorOptions.h>
8
+ #include <c10/util/Deprecated.h>
9
+ #include <optional>
10
+ #include <c10/core/QScheme.h>
11
+ #include <ATen/core/Reduction.h>
12
+ #include <ATen/core/Tensor.h>
13
+ #include <tuple>
14
+ #include <vector>
15
+
16
+
17
+ namespace at {
18
+ namespace native {
19
+ TORCH_API at::Tensor one_hot(const at::Tensor & self, int64_t num_classes=-1);
20
+ } // namespace native
21
+ } // namespace at