[skip ci] Auto-format code with clang-format and black

Author: github-actions[bot]

csrc/include/pmpp/pch.hpp

@@ -1,10 +1,10 @@
 #pragma once
 
+#include <algorithm>
 #include <cuda_runtime.h>
-#include <torch/torch.h>
 #include <torch/python.h>
+#include <torch/torch.h>
 #include <type_traits>
-#include <algorithm>
 
 #include "pmpp/system.hpp"
 #include "pmpp/types/cu_types.cuh"

csrc/lib/ops/alphabetHistogram/op.hpp

@@ -1,7 +1,6 @@
 #pragma once
 #include "pmpp/pch.hpp"
 
-
 namespace pmpp::ops::cpu
 {
 template <typename ScalarT>

csrc/lib/ops/conv2d/op.cuh

@@ -9,7 +9,7 @@ namespace pmpp::ops::cuda
 {
 static constexpr int32_t MAX_CONV2D_KERNEL_SIZE = 9;
 static __constant__ fp32_t
-        KERNEL[MAX_CONV2D_KERNEL_SIZE * MAX_CONV2D_KERNEL_SIZE];
+    KERNEL[MAX_CONV2D_KERNEL_SIZE * MAX_CONV2D_KERNEL_SIZE];
 
 template <typename ScalarT, uint32_t TILE_SIZE = 32>
 __global__ void conv2DKernel(const ScalarT* input, const ScalarT* kernel,

csrc/lib/ops/matmul/op.cuh

@@ -75,8 +75,10 @@ void launchMatmul(const fp32_t* dA, const fp32_t* dB, fp32_t* dC, size_t width)
     PMPP_DEBUG_CUDA_ERR_CHECK(cudaGetLastError());
 }
 
-namespace torch_impl{
-inline auto matmul(const torch::Tensor& A, const torch::Tensor& B) -> torch::Tensor
+namespace torch_impl
+{
+inline auto matmul(const torch::Tensor& A, const torch::Tensor& B)
+    -> torch::Tensor
 {
     torch::Tensor C = torch::empty({A.size(0), B.size(1)}, A.options());
 
@@ -93,5 +95,5 @@ inline auto matmul(const torch::Tensor& A, const torch::Tensor& B) -> torch::Ten
 
     return C;
 }
-}
+}  // namespace torch_impl
 }  // namespace pmpp::ops::cuda

csrc/lib/ops/matmul/op.hpp

@@ -18,7 +18,8 @@ void launchMatmul(const fp32_t* A, const fp32_t* B, fp32_t* C, size_t m)
 
 namespace torch_impl
 {
-inline auto matmul(const torch::Tensor& A, const torch::Tensor& B) -> torch::Tensor
+inline auto matmul(const torch::Tensor& A, const torch::Tensor& B)
+    -> torch::Tensor
 {
     torch::Tensor C = torch::empty({A.size(0), B.size(1)}, A.options());
 
@@ -35,5 +36,5 @@ inline auto matmul(const torch::Tensor& A, const torch::Tensor& B) -> torch::Ten
 
     return C;
 }
-}
+}  // namespace torch_impl
 }  // namespace pmpp::ops::cpu

csrc/lib/ops/reduction/op.hpp

@@ -24,8 +24,10 @@ inline auto mulReduction(const torch::Tensor& in) -> torch::Tensor
 
     switch (in.scalar_type()) {
     case torch::kFloat32: {
-        result = torch::tensor(launchReduction(mutableIn.mutable_data_ptr<fp32_t>(), in.numel(),
-                        std::multiplies<>()), in.options());
+        result =
+            torch::tensor(launchReduction(mutableIn.mutable_data_ptr<fp32_t>(),
+                                          in.numel(), std::multiplies<>()),
+                          in.options());
         break;
     }
     default: {

csrc/lib/ops/torch_impl.hpp

@@ -8,7 +8,7 @@
 #include "./cvtRGBtoGray/op.hpp"
 #include "./matmul/op.cuh"
 #include "./matmul/op.hpp"
-#include "./vecAdd/op.cuh"
-#include "./vecAdd/op.hpp"
+#include "./reduction/op.cuh"
 #include "./reduction/op.hpp"
-#include "./reduction/op.cuh"
+#include "./vecAdd/op.cuh"
+#include "./vecAdd/op.hpp"

csrc/lib/ops/vecAdd/op.cuh

@@ -28,7 +28,8 @@ void launchVecAdd(const fp32_t* d_A, const fp32_t* d_B, fp32_t* d_C, size_t n)
 
 namespace torch_impl
 {
-inline auto vectorAdd(const torch::Tensor& A, const torch::Tensor& B) -> torch::Tensor
+inline auto vectorAdd(const torch::Tensor& A, const torch::Tensor& B)
+    -> torch::Tensor
 {
     torch::Tensor C = torch::empty_like(A);
 
@@ -45,5 +46,5 @@ inline auto vectorAdd(const torch::Tensor& A, const torch::Tensor& B) -> torch::
 
     return C;
 }
-}
+}  // namespace torch_impl
 }  // namespace pmpp::ops::cuda

csrc/lib/ops/vecAdd/op.hpp

@@ -14,7 +14,8 @@ void launchVecAdd(const fp32_t* a, const fp32_t* b, fp32_t* c, size_t n)
 
 namespace torch_impl
 {
-inline auto vectorAdd(const torch::Tensor& A, const torch::Tensor& B) -> torch::Tensor
+inline auto vectorAdd(const torch::Tensor& A, const torch::Tensor& B)
+    -> torch::Tensor
 {
     torch::Tensor C = torch::zeros_like(A);
 
@@ -31,5 +32,5 @@ inline auto vectorAdd(const torch::Tensor& A, const torch::Tensor& B) -> torch::
 
     return C;
 }
-}
+}  // namespace torch_impl
 }  // namespace pmpp::ops::cpu