resort .cu headers, set clang-format not sort include block and consider .cu...

resort .cu headers, set clang-format not sort include block and consider .cu as main source file (#43633)

.pre-commit-config.yaml

@@ -27,7 +27,7 @@ repos:
         files: (api|backward)\.yaml$
 -   repo: local
     hooks:
-    -   id: clang-format-with-version-check
+    -   id: clang-format
         name: clang-format
         description: Format files with ClangFormat.
         entry: bash ./tools/codestyle/clang_format.hook -i

paddle/phi/kernels/gpu/abs_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/abs_grad_kernel.h"
+
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/abs_grad_kernel.h"
 #include "paddle/phi/kernels/impl/abs_grad_kernel_impl.h"
 
 using phi::dtype::complex;

paddle/phi/kernels/gpu/abs_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/abs_kernel.h"
+
 #include <algorithm>
 #include <vector>
 
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/abs_kernel.h"
 #include "paddle/phi/kernels/funcs/complex_functors.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
 

paddle/phi/kernels/gpu/accuracy_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/accuracy_kernel.h"
+
 #include <thrust/execution_policy.h>
 #include <thrust/reduce.h>
 
@@ -20,7 +22,6 @@
 #include "paddle/phi/backends/gpu/gpu_info.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/accuracy_kernel.h"
 
 namespace phi {
 using paddle::platform::PADDLE_CUDA_NUM_THREADS;

paddle/phi/kernels/gpu/activation_grad_kernel.cu

@@ -12,12 +12,13 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/activation_grad_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_device_function.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/activation_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
 #include "paddle/phi/kernels/impl/activation_grad_impl.h"
 

paddle/phi/kernels/gpu/activation_kernel.cu

@@ -12,12 +12,13 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/activation_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_device_function.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/activation_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
 #include "paddle/phi/kernels/impl/activation_grad_impl.h"
 #include "paddle/phi/kernels/impl/activation_impl.h"

paddle/phi/kernels/gpu/adadelta_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/adadelta_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/adadelta_kernel.h"
 #include "paddle/phi/kernels/impl/adadelta_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/adagrad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/adagrad_kernel.h"
+
 #include "paddle/fluid/operators/math/selected_rows_functor.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/adagrad_kernel.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/impl/adagrad_kernel_impl.h"
 

paddle/phi/kernels/gpu/adam_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/adam_kernel.h"
+
 #include <math.h>  // for sqrt in CPU and CUDA
 
 #include <vector>
@@ -22,7 +24,6 @@
 #include "paddle/phi/common/amp_type_traits.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/adam_kernel.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/adam_functors.h"
 #include "paddle/phi/kernels/funcs/for_range.h"

paddle/phi/kernels/gpu/adamax_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/adamax_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/adamax_kernel.h"
 #include "paddle/phi/kernels/impl/adamax_kernel_impl.h"
 
 PD_REGISTER_KERNEL(adamax, GPU, ALL_LAYOUT, phi::AdamaxKernel, float, double) {}

paddle/phi/kernels/gpu/adamw_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/adamw_kernel.h"
+
 #include <math.h>  // for sqrt in CPU and CUDA
 
 #include <vector>
@@ -22,7 +24,6 @@
 #include "paddle/phi/common/amp_type_traits.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/adamw_kernel.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/adam_functors.h"
 #include "paddle/phi/kernels/funcs/for_range.h"

paddle/phi/kernels/gpu/add_n_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/add_n_kernel.h"
+
 #include "paddle/fluid/memory/malloc.h"
 #include "paddle/fluid/memory/memcpy.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/add_n_kernel.h"
 #include "paddle/phi/kernels/funcs/eigen/common.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 

paddle/phi/kernels/gpu/addmm_grad_kernel.cu

@@ -12,9 +12,10 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/addmm_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/addmm_grad_kernel.h"
 #include "paddle/phi/kernels/impl/addmm_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/addmm_kernel.cu

@@ -12,9 +12,10 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/addmm_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/addmm_kernel.h"
 #include "paddle/phi/kernels/impl/addmm_kernel_impl.h"
 
 PD_REGISTER_KERNEL(addmm, GPU, ALL_LAYOUT, phi::AddmmKernel, float, double) {}

paddle/phi/kernels/gpu/allclose_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/allclose_kernel.h"
+
 #include "paddle/phi/core/enforce.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/allclose_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/arange_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/arange_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/arange_kernel.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/range_function.h"
 

paddle/phi/kernels/gpu/arg_min_max_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/arg_min_max_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/arg_min_max_kernel.h"
 
 #if defined(__NVCC__) || defined(__HIPCC__)
 

paddle/phi/kernels/gpu/argsort_kernel.cu

@@ -12,12 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/argsort_kernel.h"
+
 #include <thrust/copy.h>
 #include <thrust/execution_policy.h>
 #include <thrust/sequence.h>
 #include <thrust/sort.h>
-
-#include "paddle/phi/kernels/argsort_kernel.h"
 #ifdef __NVCC__
 #include "cub/cub.cuh"
 #endif

paddle/phi/kernels/gpu/auc_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/auc_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/cpu/cpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/auc_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/bce_loss_grad_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/bce_loss_grad_kernel.h"
+
 #include <algorithm>
 #include <vector>
 
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/hostdevice.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/bce_loss_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
 
 namespace phi {

paddle/phi/kernels/gpu/bce_loss_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/bce_loss_kernel.h"
+
 #include <algorithm>
 #include <vector>
 
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/hostdevice.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/bce_loss_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
 #include "paddle/phi/kernels/primitive/functor_primitives.h"
 

paddle/phi/kernels/gpu/bilinear_tensor_product_grad_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/bilinear_tensor_product_grad_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/bilinear_tensor_product_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(bilinear_tensor_product_grad,

paddle/phi/kernels/gpu/bilinear_tensor_product_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/bilinear_tensor_product_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/bilinear_tensor_product_kernel_impl.h"
 
 PD_REGISTER_KERNEL(bilinear_tensor_product,

paddle/phi/kernels/gpu/bincount_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/bincount_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/bincount_kernel.h"
 #include "paddle/phi/kernels/funcs/eigen/common.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 

paddle/phi/kernels/gpu/broadcast_tensors_grad_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/broadcast_tensors_grad_kernel.h"
+
 #include <vector>
 
 #include "paddle/fluid/framework/tensor_util.h"
@@ -19,7 +21,6 @@
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/enforce.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/broadcast_tensors_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/reduce_function.h"
 #include "paddle/phi/kernels/primitive/functor_primitives.h"
 

paddle/phi/kernels/gpu/broadcast_tensors_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/broadcast_tensors_kernel.h"
+
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/broadcast_tensors_kernel.h"
 #include "paddle/phi/kernels/impl/broadcast_tensors_kernel_impl.h"
 
 PD_REGISTER_KERNEL(broadcast_tensors,

paddle/phi/kernels/gpu/cast_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cast_grad_kernel.h"
+
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/core/visit_type.h"
-#include "paddle/phi/kernels/cast_grad_kernel.h"
 #include "paddle/phi/kernels/gpu/cast_impl.h"
 
 namespace phi {

paddle/phi/kernels/gpu/cast_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cast_kernel.h"
+
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/core/visit_type.h"
-#include "paddle/phi/kernels/cast_kernel.h"
 #include "paddle/phi/kernels/gpu/cast_impl.h"
 
 namespace phi {

paddle/phi/kernels/gpu/channel_shuffle_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/channel_shuffle_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/channel_shuffle_grad_kernel.h"
 #include "paddle/phi/kernels/impl/channel_shuffle_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(channel_shuffle_grad,

paddle/phi/kernels/gpu/channel_shuffle_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/channel_shuffle_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/channel_shuffle_kernel.h"
 #include "paddle/phi/kernels/impl/channel_shuffle_kernel_impl.h"
 
 PD_REGISTER_KERNEL(channel_shuffle,

paddle/phi/kernels/gpu/cholesky_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cholesky_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/cholesky_grad_kernel.h"
 #include "paddle/phi/kernels/impl/cholesky_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/cholesky_kernel.cu

@@ -15,6 +15,8 @@ limitations under the License. */
 #ifndef PADDLE_WITH_HIP
 // HIP not support cusolver
 
+#include "paddle/phi/kernels/cholesky_kernel.h"
+
 #include <thrust/device_vector.h>
 
 #include <algorithm>
@@ -25,7 +27,6 @@ limitations under the License. */
 #include "paddle/phi/backends/dynload/cusolver.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/cholesky_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/clip_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/clip_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/clip_grad_kernel.h"
 #include "paddle/phi/kernels/impl/clip_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(clip_grad,

paddle/phi/kernels/gpu/clip_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/clip_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/clip_kernel.h"
 #include "paddle/phi/kernels/impl/clip_kernel_impl.h"
 
 PD_REGISTER_KERNEL(clip,

paddle/phi/kernels/gpu/complex_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/complex_grad_kernel.h"
+
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/complex_grad_kernel.h"
 #include "paddle/phi/kernels/impl/complex_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(imag_grad,

paddle/phi/kernels/gpu/complex_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/complex_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/complex_kernel.h"
 #include "paddle/phi/kernels/impl/complex_kernel_impl.h"
 
 // See Note [ Why still include the fluid headers? ]

paddle/phi/kernels/gpu/concat_grad_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/concat_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/concat_grad_kernel.h"
 #include "paddle/phi/kernels/impl/concat_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(concat_grad,

paddle/phi/kernels/gpu/concat_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/concat_kernel.h"
+
 #include "paddle/fluid/operators/strided_memcpy.h"
 #include "paddle/fluid/platform/bfloat16.h"
 #include "paddle/fluid/platform/complex.h"
@@ -20,7 +22,6 @@
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/core/lod_utils.h"
-#include "paddle/phi/kernels/concat_kernel.h"
 #include "paddle/phi/kernels/funcs/concat_and_split_functor.h"
 #include "paddle/phi/kernels/funcs/concat_funcs.h"
 

paddle/phi/kernels/gpu/conv_grad_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/conv_grad_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_grad_grad_kernel.h"
 #include "paddle/phi/kernels/impl/conv_grad_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/conv_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/conv_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_grad_kernel.h"
 #include "paddle/phi/kernels/impl/conv_grad_kernel_impl.h"
 
 namespace phi {

paddle/phi/kernels/gpu/conv_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/conv_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_kernel.h"
 #include "paddle/phi/kernels/impl/conv_kernel_impl.h"
 
 namespace phi {

paddle/phi/kernels/gpu/conv_transpose_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/conv_transpose_grad_kernel.h"
+
 #include "paddle/phi/common/layout.h"
 #include "paddle/phi/core/ddim.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_transpose_grad_kernel.h"
 #include "paddle/phi/kernels/cpu/conv_util.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/gpu/depthwise_conv.h"

paddle/phi/kernels/gpu/conv_transpose_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/conv_transpose_kernel.h"
+
 #include "paddle/phi/common/layout.h"
 #include "paddle/phi/core/ddim.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_transpose_kernel.h"
 #include "paddle/phi/kernels/cpu/conv_util.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/gpu/depthwise_conv.h"

paddle/phi/kernels/gpu/copy_kernel.cu

@@ -12,11 +12,12 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/copy_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/data_type.h"
 #include "paddle/phi/core/compat/convert_utils.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/copy_kernel.h"
 
 // See Note [ Why still include the fluid headers? ]
 #include "paddle/fluid/memory/malloc.h"

paddle/phi/kernels/gpu/cross_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cross_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/cross_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/reduce_function.h"
 
 namespace phi {

paddle/phi/kernels/gpu/cross_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cross_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/cross_kernel.h"
 #include "paddle/phi/kernels/funcs/reduce_function.h"
 
 namespace phi {

paddle/phi/kernels/gpu/cum_kernel.cu

@@ -12,12 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cum_kernel.h"
+
 #include <thrust/device_ptr.h>
 #include <thrust/device_vector.h>
 #include <thrust/reverse.h>
 #include <thrust/scan.h>
-
-#include "paddle/phi/kernels/cum_kernel.h"
 #ifdef __NVCC__
 #include <cub/cub.cuh>
 #endif
@@ -271,7 +271,6 @@ void ScanKernel(const Context& dev_ctx,
     return;
   }
 
-
   size_t height = 1;
   size_t width = 1;
   for (size_t i = 0; i <= axis; i++) {
@@ -391,9 +390,5 @@ PD_REGISTER_KERNEL(cumsum,
                    int,
                    int64_t) {}
 
-PD_REGISTER_KERNEL(logcumsumexp,
-                   GPU,
-                   ALL_LAYOUT,
-                   phi::LogcumsumexpKernel,
-                   float,
-                   double) {}
+PD_REGISTER_KERNEL(
+    logcumsumexp, GPU, ALL_LAYOUT, phi::LogcumsumexpKernel, float, double) {}

paddle/phi/kernels/gpu/cumprod_grad_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cumprod_grad_kernel.h"
+
 #include <thrust/transform.h>
 
 #include "paddle/fluid/operators/math/inclusive_scan.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/cumprod_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/complex_functors.h"
 #include "paddle/phi/kernels/funcs/cumprod.h"
 #include "paddle/phi/kernels/funcs/elementwise_functor.h"

paddle/phi/kernels/gpu/cumprod_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cumprod_kernel.h"
+
 #include "paddle/fluid/operators/math/inclusive_scan.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/cumprod_kernel.h"
 #include "paddle/phi/kernels/funcs/complex_functors.h"
 #include "paddle/phi/kernels/funcs/cumprod.h"
 #include "paddle/phi/kernels/funcs/elementwise_functor.h"

paddle/phi/kernels/gpu/deformable_conv_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/deformable_conv_grad_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/deformable_conv_grad_kernel.h"
 #include "paddle/phi/kernels/impl/deformable_conv_grad_kernel_impl.h"
 
 namespace phi {

paddle/phi/kernels/gpu/deformable_conv_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/deformable_conv_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/deformable_conv_kernel.h"
 #include "paddle/phi/kernels/impl/deformable_conv_kernel_impl.h"
 
 PD_REGISTER_KERNEL(deformable_conv,

paddle/phi/kernels/gpu/determinant_grad_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/determinant_grad_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/determinant_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(determinant_grad,

paddle/phi/kernels/gpu/determinant_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/determinant_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/determinant_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/diag_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/diag_kernel.h"
+
 #include <algorithm>
 #include <tuple>
 
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/diag_kernel.h"
 #include "paddle/phi/kernels/funcs/diag_functor.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 

paddle/phi/kernels/gpu/diagonal_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/diagonal_grad_kernel.h"
+
 #include "paddle/fluid/framework/tensor_util.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/diagonal_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/diagonal.h"
 
 namespace phi {

paddle/phi/kernels/gpu/diagonal_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/diagonal_kernel.h"
+
 #include "paddle/fluid/framework/tensor_util.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/diagonal_kernel.h"
 #include "paddle/phi/kernels/funcs/diagonal.h"
 
 namespace phi {

paddle/phi/kernels/gpu/digamma_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/digamma_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/digamma_grad_kernel.h"
 #include "paddle/phi/kernels/impl/digamma_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/digamma_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/digamma_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/scalar.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/digamma_kernel.h"
 #include "paddle/phi/kernels/impl/digamma_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/dist_grad_kernel.cu

@@ -12,9 +12,10 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/dist_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/dist_grad_kernel.h"
 #include "paddle/phi/kernels/impl/dist_grad_kernel_impl.h"
 
 #ifdef PADDLE_WITH_HIP

paddle/phi/kernels/gpu/dist_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/dist_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/dist_kernel.h"
 #include "paddle/phi/kernels/impl/dist_kernel_impl.h"
 
 #ifdef PADDLE_WITH_HIP

paddle/phi/kernels/gpu/dot_grad_kernel.cu

@@ -12,10 +12,11 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/dot_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/dot_grad_kernel.h"
 #include "paddle/phi/kernels/impl/dot_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(dot_grad,

paddle/phi/kernels/gpu/dot_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/dot_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/dot_kernel.h"
 #include "paddle/phi/kernels/funcs/eigen/common.h"
 
 // See Note [ Why still include the fluid headers? ]

paddle/phi/kernels/gpu/dropout_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/dropout_grad_kernel.h"
+
 #include "paddle/fluid/operators/dropout_impl.cu.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/dropout_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/dropout_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/dropout_kernel.h"
+
 #include "paddle/fluid/operators/dropout_impl.cu.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/dropout_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/eigh_grad_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/eigh_grad_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/complex_functors.h"
 #include "paddle/phi/kernels/impl/eigh_grad_kernel_impl.h"
 

paddle/phi/kernels/gpu/eigh_kernel.cu

@@ -15,8 +15,9 @@
 #ifndef PADDLE_WITH_HIP
 // HIP not support cusolver
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/eigh_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/complex_functors.h"
 #include "paddle/phi/kernels/funcs/values_vectors_functor.h"
 

paddle/phi/kernels/gpu/einsum_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/einsum_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/einsum_kernel.h"
 #include "paddle/phi/kernels/impl/einsum_impl.h"
 
 PD_REGISTER_KERNEL(einsum,

paddle/phi/kernels/gpu/elementwise_add_grad_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/elementwise_add_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
-#include "paddle/phi/kernels/elementwise_add_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_functor.h"
 #include "paddle/phi/kernels/gpu/elementwise_grad.h"
 #include "paddle/phi/kernels/impl/elementwise_grad_kernel_impl.h"

paddle/phi/kernels/gpu/elementwise_divide_grad_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/elementwise_divide_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/elementwise_divide_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_functor.h"
 #include "paddle/phi/kernels/gpu/elementwise_grad.h"
 #include "paddle/phi/kernels/impl/elementwise_grad_kernel_impl.h"

paddle/phi/kernels/gpu/elementwise_grad_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/elementwise_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
-#include "paddle/phi/kernels/elementwise_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_functor.h"
 #include "paddle/phi/kernels/gpu/elementwise_grad.h"
 #include "paddle/phi/kernels/impl/elementwise_grad_kernel_impl.h"

paddle/phi/kernels/gpu/elementwise_multiply_grad_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/elementwise_multiply_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/elementwise_multiply_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_functor.h"
 #include "paddle/phi/kernels/gpu/elementwise_grad.h"
 #include "paddle/phi/kernels/impl/elementwise_grad_kernel_impl.h"

paddle/phi/kernels/gpu/embedding_grad_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/embedding_grad_kernel.h"
+
 #include "paddle/fluid/framework/mixed_vector.h"
 #include "paddle/fluid/memory/memcpy.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/data_type.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/embedding_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/eigen/common.h"
 #include "paddle/phi/kernels/funcs/embedding_util.h"
 

paddle/phi/kernels/gpu/embedding_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/embedding_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/data_type.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/embedding_kernel.h"
 #include "paddle/phi/kernels/funcs/eigen/eigen_function.h"
 #include "paddle/phi/kernels/funcs/embedding_util.h"
 

paddle/phi/kernels/gpu/erf_grad_kernel.cu

@@ -12,10 +12,11 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/erf_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/erf_grad_kernel.h"
 #include "paddle/phi/kernels/impl/erf_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(erf_grad,

paddle/phi/kernels/gpu/erf_kernel.cu

@@ -12,10 +12,11 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/erf_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/erf_kernel.h"
 #include "paddle/phi/kernels/impl/erf_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/erfinv_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/erfinv_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/erfinv_grad_kernel.h"
 #include "paddle/phi/kernels/impl/erfinv_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/erfinv_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/erfinv_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/erfinv_kernel.h"
 #include "paddle/phi/kernels/impl/erfinv_kernel_impl.h"
 
 PD_REGISTER_KERNEL(erfinv, GPU, ALL_LAYOUT, phi::ErfinvKernel, float, double) {}

paddle/phi/kernels/gpu/expand_as_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/expand_as_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/expand_as_grad_kernel.h"
 #include "paddle/phi/kernels/impl/expand_as_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(expand_as_grad,

paddle/phi/kernels/gpu/expand_as_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/expand_as_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/expand_as_kernel.h"
 #include "paddle/phi/kernels/impl/expand_as_kernel_impl.h"
 
 PD_REGISTER_KERNEL(expand_as,

paddle/phi/kernels/gpu/expand_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/expand_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/expand_grad_kernel.h"
 #include "paddle/phi/kernels/impl/expand_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(expand_grad,

paddle/phi/kernels/gpu/expand_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/expand_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/scalar.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/expand_kernel.h"
 #include "paddle/phi/kernels/impl/expand_kernel_impl.h"
 
 PD_REGISTER_KERNEL(expand,

paddle/phi/kernels/gpu/eye_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/eye_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/eye_kernel.h"
 #include "paddle/phi/kernels/impl/eye_kernel_impl.h"
 
 PD_REGISTER_KERNEL(eye,

paddle/phi/kernels/gpu/flip_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/flip_kernel.h"
+
 #include "paddle/fluid/memory/malloc.h"
 #include "paddle/fluid/memory/memcpy.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/place.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/flip_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/frobenius_norm_grad_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/frobenius_norm_grad_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/frobenius_norm_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(frobenius_norm_grad,

paddle/phi/kernels/gpu/frobenius_norm_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/frobenius_norm_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/activation_functor.h"
 #include "paddle/phi/kernels/gpu/reduce.h"
 namespace phi {

paddle/phi/kernels/gpu/full_kernel.cu

@@ -12,9 +12,10 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/full_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/full_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
 namespace phi {
 

paddle/phi/kernels/gpu/gather_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/gather_kernel.h"
+
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/gather.cu.h"
-#include "paddle/phi/kernels/gather_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/gather_nd_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/gather_nd_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/eigen/common.h"
 #include "paddle/phi/kernels/funcs/scatter.cu.h"
-#include "paddle/phi/kernels/gather_nd_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/gather_nd_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/gather_nd_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/gather.cu.h"
 #include "paddle/phi/kernels/funcs/scatter.cu.h"
-#include "paddle/phi/kernels/gather_nd_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/gather_tree_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/gather_tree_kernel.h"
+
 #include <algorithm>
 
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/gather_tree_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/gaussian_random_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/gaussian_random_kernel.h"
+
 #include <thrust/random.h>
 
 #include "paddle/fluid/framework/generator.h"
@@ -21,7 +23,6 @@
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/distribution_helper.h"
 #include "paddle/phi/kernels/funcs/index_impl.cu.h"
-#include "paddle/phi/kernels/gaussian_random_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/graph_reindex_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/graph_reindex_kernel.h"
+
 #include <thrust/copy.h>
 #include <thrust/device_vector.h>
 #include <thrust/reduce.h>
@@ -21,7 +23,6 @@
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpu/graph_reindex_funcs.h"
-#include "paddle/phi/kernels/graph_reindex_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/graph_send_recv_grad_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/graph_send_recv_grad_kernel.h"
+
 #include <algorithm>
 #include <vector>
 
@@ -19,7 +21,6 @@
 #include "paddle/phi/core/hostdevice.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpu/graph_send_recv_funcs.h"
-#include "paddle/phi/kernels/graph_send_recv_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/graph_send_recv_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/graph_send_recv_kernel.h"
+
 #include <thrust/device_vector.h>
 #include <thrust/fill.h>
 
@@ -22,7 +24,6 @@
 #include "paddle/phi/core/hostdevice.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpu/graph_send_recv_funcs.h"
-#include "paddle/phi/kernels/graph_send_recv_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/grid_sample_grad_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/grid_sample_grad_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_device_function.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_info.h"
@@ -19,7 +21,6 @@
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/gpu/grid_sample_utils.h"
-#include "paddle/phi/kernels/grid_sample_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/grid_sample_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/grid_sample_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_info.h"
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpu/grid_sample_utils.h"
-#include "paddle/phi/kernels/grid_sample_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/group_norm_grad_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/group_norm_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/layout.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/gpu/group_norm_utils.h"
-#include "paddle/phi/kernels/group_norm_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/group_norm_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/group_norm_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/layout.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/gpu/group_norm_utils.h"
-#include "paddle/phi/kernels/group_norm_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/gumbel_softmax_grad_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gumbel_softmax_grad_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/gumbel_softmax_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(gumbel_softmax_grad,

paddle/phi/kernels/gpu/gumbel_softmax_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/gumbel_softmax_kernel.h"
+
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/axis_utils.h"
-#include "paddle/phi/kernels/gumbel_softmax_kernel.h"
 #include "paddle/phi/kernels/impl/gumbel_softmax_kernel_impl.h"
 
 #if defined(__NVCC__) || defined(__HIPCC__)

paddle/phi/kernels/gpu/histogram_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/histogram_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_launch_config.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
@@ -19,7 +21,6 @@
 #include "paddle/phi/kernels/funcs/eigen/common.h"
 #include "paddle/phi/kernels/funcs/eigen/eigen_function.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/histogram_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/huber_loss_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/huber_loss_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/huber_loss_grad_kernel.h"
 #include "paddle/phi/kernels/impl/huber_loss_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/huber_loss_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/huber_loss_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/huber_loss_kernel.h"
 #include "paddle/phi/kernels/impl/huber_loss_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/increment_kernel.cu

@@ -12,10 +12,11 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/increment_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/increment_kernel_impl.h"
-#include "paddle/phi/kernels/increment_kernel.h"
 
 PD_REGISTER_KERNEL(increment,
                    GPU,

paddle/phi/kernels/gpu/index_sample_grad_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/index_sample_grad_kernel.h"
+
 #include <algorithm>
 #include <vector>
 
@@ -21,7 +23,6 @@
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/index_sample_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/index_sample_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/index_sample_kernel.h"
+
 #include <algorithm>
 #include <vector>
 
@@ -20,7 +22,6 @@
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/index_sample_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/index_select_grad_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/index_select_grad_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_launch_config.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_info.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/core/utils/data_type.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/index_select_grad_kernel.h"
 
 DECLARE_bool(cudnn_deterministic);
 

paddle/phi/kernels/gpu/index_select_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/index_select_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_launch_config.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_info.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/core/utils/data_type.h"
-#include "paddle/phi/kernels/index_select_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/instance_norm_grad_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/instance_norm_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/layout.h"
 #include "paddle/phi/core/kernel_registry.h"
@@ -19,7 +21,6 @@
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/funcs/norm_utils.h"
 #include "paddle/phi/kernels/gpu/instance_norm_utils.h"
-#include "paddle/phi/kernels/instance_norm_grad_kernel.h"
 
 namespace phi {
 template <typename T, int BlockDim>

paddle/phi/kernels/gpu/instance_norm_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/instance_norm_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/layout.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/funcs/norm_utils.h"
 #include "paddle/phi/kernels/gpu/instance_norm_utils.h"
-#include "paddle/phi/kernels/instance_norm_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/interpolate_grad_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/interpolate_grad_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/fluid/platform/fast_divmod.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
@@ -21,7 +23,6 @@
 #include "paddle/phi/kernels/funcs/interpolate_function.h"
 #include "paddle/phi/kernels/funcs/math_cuda_utils.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/interpolate_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/interpolate_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/interpolate_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_device_function.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/fluid/platform/fast_divmod.h"
@@ -20,7 +22,6 @@
 #include "paddle/phi/common/layout.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/interpolate_function.h"
-#include "paddle/phi/kernels/interpolate_kernel.h"
 
 namespace phi {
 using paddle::platform::FastDivMod;

paddle/phi/kernels/gpu/isclose_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/isclose_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/isclose_kernel_impl.h"
-#include "paddle/phi/kernels/isclose_kernel.h"
 
 PD_REGISTER_KERNEL(
     isclose, GPU, ALL_LAYOUT, phi::IscloseKernel, float, double) {}

paddle/phi/kernels/gpu/isfinite_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/isfinite_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/isfinite_kernel_impl.h"
-#include "paddle/phi/kernels/isfinite_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/kldiv_loss_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/kldiv_loss_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/kldiv_loss_grad_kernel_impl.h"
-#include "paddle/phi/kernels/kldiv_loss_grad_kernel.h"
 PD_REGISTER_KERNEL(
     kldiv_loss_grad, GPU, ALL_LAYOUT, phi::KLDivLossGradKernel, float, double) {
 }

paddle/phi/kernels/gpu/kldiv_loss_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/kldiv_loss_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/kldiv_loss_kernel_impl.h"
-#include "paddle/phi/kernels/kldiv_loss_kernel.h"
 PD_REGISTER_KERNEL(
     kldiv_loss, GPU, ALL_LAYOUT, phi::KLDivLossKernel, float, double) {}

paddle/phi/kernels/gpu/kron_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/kron_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/kron_grad_kernel_impl.h"
-#include "paddle/phi/kernels/kron_grad_kernel.h"
 
 PD_REGISTER_KERNEL(kron_grad,
                    GPU,

paddle/phi/kernels/gpu/kron_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/kron_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/kron_kernel_impl.h"
-#include "paddle/phi/kernels/kron_kernel.h"
 
 PD_REGISTER_KERNEL(kron,
                    GPU,

paddle/phi/kernels/gpu/kthvalue_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/kthvalue_grad_kernel.h"
+
 #include "paddle/fluid/operators/top_k_function_cuda.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/kthvalue_grad_kernel.h"
 
 namespace phi {
 static int getBlockSize(int col) {

paddle/phi/kernels/gpu/kthvalue_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/kthvalue_kernel.h"
+
 #include "paddle/fluid/operators/top_k_function_cuda.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/eigen/common.h"
 #include "paddle/phi/kernels/funcs/eigen/eigen_function.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/kthvalue_kernel.h"
 
 namespace phi {
 inline int getBlockSize(int col) {

paddle/phi/kernels/gpu/label_smooth_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/label_smooth_grad_kernel.h"
+
 #include "paddle/fluid/operators/elementwise/elementwise_op_impl.cu.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/label_smooth_grad_kernel.h"
 
 namespace phi {
 template <typename T>

paddle/phi/kernels/gpu/label_smooth_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/label_smooth_kernel.h"
+
 #include <vector>
 
 #include "paddle/fluid/operators/elementwise/elementwise_op_impl.cu.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/label_smooth_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/layer_norm_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/layer_norm_grad_kernel.h"
+
 #include "paddle/fluid/operators/layer_norm_kernel.cu.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/layer_norm_util.h"
-#include "paddle/phi/kernels/layer_norm_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/layer_norm_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/layer_norm_kernel.h"
+
 #include "paddle/fluid/operators/layer_norm_kernel.cu.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/layer_norm_util.h"
-#include "paddle/phi/kernels/layer_norm_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/lerp_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/lerp_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/lerp_grad_kernel_impl.h"
-#include "paddle/phi/kernels/lerp_grad_kernel.h"
 
 PD_REGISTER_KERNEL(
     lerp_grad, GPU, ALL_LAYOUT, phi::LerpGradKernel, float, double) {}

paddle/phi/kernels/gpu/lerp_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/lerp_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/lerp_kernel_impl.h"
-#include "paddle/phi/kernels/lerp_kernel.h"
 
 PD_REGISTER_KERNEL(lerp, GPU, ALL_LAYOUT, phi::LerpKernel, float, double) {}

paddle/phi/kernels/gpu/lgamma_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/lgamma_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/lgamma_grad_kernel_impl.h"
-#include "paddle/phi/kernels/lgamma_grad_kernel.h"
 PD_REGISTER_KERNEL(
     lgamma_grad, GPU, ALL_LAYOUT, phi::LgammaGradKernel, float, double) {}

paddle/phi/kernels/gpu/lgamma_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/lgamma_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
-#include "paddle/phi/kernels/lgamma_kernel.h"
 
 namespace phi {
 template <typename T>

paddle/phi/kernels/gpu/linspace_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/linspace_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/linspace_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/log_loss_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/log_loss_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/log_loss_grad_kernel_impl.h"
-#include "paddle/phi/kernels/log_loss_grad_kernel.h"
 
 PD_REGISTER_KERNEL(
     log_loss_grad, GPU, ALL_LAYOUT, phi::LogLossGradKernel, float) {}

paddle/phi/kernels/gpu/log_loss_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/log_loss_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/log_loss_kernel_impl.h"
-#include "paddle/phi/kernels/log_loss_kernel.h"
 
 PD_REGISTER_KERNEL(log_loss, GPU, ALL_LAYOUT, phi::LogLossKernel, float) {}

paddle/phi/kernels/gpu/log_softmax_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/log_softmax_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpudnn/softmax_gpudnn.h"
-#include "paddle/phi/kernels/log_softmax_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/log_softmax_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/log_softmax_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpudnn/softmax_gpudnn.h"
-#include "paddle/phi/kernels/log_softmax_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/logcumsumexp_grad_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/logcumsumexp_grad_kernel.h"
+
 #include <limits>
 
 #include "paddle/phi/backends/cpu/cpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/logcumsumexp_grad_impl.h"
-#include "paddle/phi/kernels/logcumsumexp_grad_kernel.h"
 
 PD_REGISTER_KERNEL(logcumsumexp_grad,
                    GPU,

paddle/phi/kernels/gpu/logspace_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/logspace_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/data_type_transform.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/logspace_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/logsumexp_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/logsumexp_grad_kernel.h"
+
 #include "paddle/phi/backends/cpu/cpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/logsumexp_grad_kernel_impl.h"
-#include "paddle/phi/kernels/logsumexp_grad_kernel.h"
 
 PD_REGISTER_KERNEL(
     logsumexp_grad, GPU, ALL_LAYOUT, phi::LogsumexpGradKernel, float, double) {}

paddle/phi/kernels/gpu/logsumexp_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/logsumexp_kernel.h"
+
 #include "paddle/phi/backends/cpu/cpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/logsumexp_kernel_impl.h"
-#include "paddle/phi/kernels/logsumexp_kernel.h"
 
 PD_REGISTER_KERNEL(
     logsumexp, GPU, ALL_LAYOUT, phi::LogsumexpKernel, float, double) {}

paddle/phi/kernels/gpu/masked_select_grad_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/masked_select_grad_kernel.h"
+
 #include <thrust/device_ptr.h>
 #include <thrust/device_vector.h>
 #include <thrust/reverse.h>
@@ -19,7 +21,6 @@
 
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/select_impl.cu.h"
-#include "paddle/phi/kernels/masked_select_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/masked_select_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/masked_select_kernel.h"
+
 #include <thrust/device_ptr.h>
 #include <thrust/device_vector.h>
 #include <thrust/reverse.h>
@@ -20,7 +22,6 @@
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/select_impl.cu.h"
-#include "paddle/phi/kernels/masked_select_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/matmul_grad_kernel.cu

@@ -12,10 +12,11 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/matmul_grad_kernel.h"
+
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/matmul_grad_kernel_impl.h"
-#include "paddle/phi/kernels/matmul_grad_kernel.h"
 
 PD_REGISTER_KERNEL(matmul_grad,
                    GPU,

paddle/phi/kernels/gpu/matmul_kernel.cu

@@ -12,11 +12,12 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/matmul_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/matmul_kernel_impl.h"
-#include "paddle/phi/kernels/matmul_kernel.h"
 
 PD_REGISTER_KERNEL(matmul,
                    GPU,

paddle/phi/kernels/gpu/matrix_power_grad_kernel.cu

@@ -12,10 +12,11 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/matrix_power_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/matrix_power_grad_kernel_impl.h"
-#include "paddle/phi/kernels/matrix_power_grad_kernel.h"
 
 PD_REGISTER_KERNEL(matrix_power_grad,
                    GPU,

paddle/phi/kernels/gpu/matrix_power_kernel.cu

@@ -12,10 +12,11 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/matrix_power_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/matrix_power_kernel_impl.h"
-#include "paddle/phi/kernels/matrix_power_kernel.h"
 
 PD_REGISTER_KERNEL(
     matrix_power, GPU, ALL_LAYOUT, phi::MatrixPowerKernel, float, double) {}

paddle/phi/kernels/gpu/matrix_rank_kernel.cu

@@ -15,9 +15,10 @@
 #ifndef PADDLE_WITH_HIP
 // HIP not support cusolver
 
+#include "paddle/phi/kernels/matrix_rank_kernel.h"
+
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/full_kernel.h"
-#include "paddle/phi/kernels/matrix_rank_kernel.h"
 #include "paddle/phi/kernels/matrix_rank_tol_kernel.h"
 
 namespace phi {

paddle/phi/kernels/gpu/matrix_rank_tol_kernel.cu

@@ -15,6 +15,8 @@
 #ifndef PADDLE_WITH_HIP
 // HIP not support cusolver
 
+#include "paddle/phi/kernels/matrix_rank_tol_kernel.h"
+
 #include <algorithm>
 #include <vector>
 
@@ -27,7 +29,6 @@
 #include "paddle/phi/kernels/funcs/broadcast_function.h"
 #include "paddle/phi/kernels/funcs/compare_functors.h"
 #include "paddle/phi/kernels/impl/matrix_rank_kernel_impl.h"
-#include "paddle/phi/kernels/matrix_rank_tol_kernel.h"
 #include "paddle/phi/kernels/reduce_max_kernel.h"
 #include "paddle/phi/kernels/reduce_sum_kernel.h"
 

paddle/phi/kernels/gpu/mean_all_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/mean_all_kernel.h"
+
 #include "paddle/fluid/memory/memcpy.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/reduce_function.h"
-#include "paddle/phi/kernels/mean_all_kernel.h"
 #include "paddle/phi/kernels/primitive/functor_primitives.h"
 
 namespace phi {

paddle/phi/kernels/gpu/mode_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/mode_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/mode.h"
-#include "paddle/phi/kernels/mode_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/mode_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/mode_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/mode.h"
-#include "paddle/phi/kernels/mode_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/momentum_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/momentum_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/momentum_kernel_impl.h"
-#include "paddle/phi/kernels/momentum_kernel.h"
 
 PD_REGISTER_KERNEL(momentum,
                    GPU,

paddle/phi/kernels/gpu/multi_dot_grad_kernel.cu

@@ -12,11 +12,12 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/multi_dot_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/multi_dot_kernel_impl.h"
-#include "paddle/phi/kernels/multi_dot_grad_kernel.h"
 
 using float16 = phi::dtype::float16;
 

paddle/phi/kernels/gpu/multi_dot_kernel.cu

@@ -12,11 +12,12 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/multi_dot_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/multi_dot_kernel_impl.h"
-#include "paddle/phi/kernels/multi_dot_kernel.h"
 
 using float16 = phi::dtype::float16;
 

paddle/phi/kernels/gpu/multiplex_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/multiplex_grad_kernel.h"
+
 #include "paddle/phi/api/lib/utils/tensor_utils.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/eigen/common.h"
-#include "paddle/phi/kernels/multiplex_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/multiplex_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/multiplex_kernel.h"
+
 #include "paddle/phi/api/lib/utils/tensor_utils.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/multiplex_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/mv_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/mv_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/blas/blas.h"
-#include "paddle/phi/kernels/mv_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/mv_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/mv_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/mv_kernel_impl.h"
-#include "paddle/phi/kernels/mv_kernel.h"
 
 PD_REGISTER_KERNEL(mv, GPU, ALL_LAYOUT, phi::MvKernel, float, double) {}

paddle/phi/kernels/gpu/nanmedian_grad_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/nanmedian_grad_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_launch_config.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
@@ -19,7 +21,6 @@
 #include "paddle/phi/core/tensor_meta.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/impl/nanmedian_grad_kernel_impl.h"
-#include "paddle/phi/kernels/nanmedian_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/nanmedian_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/nanmedian_kernel.h"
+
 #include "paddle/fluid/memory/memcpy.h"
 #include "paddle/fluid/platform/device/gpu/gpu_launch_config.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
@@ -19,7 +21,6 @@
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/full_kernel.h"
 #include "paddle/phi/kernels/impl/nanmedian_kernel_impl.h"
-#include "paddle/phi/kernels/nanmedian_kernel.h"
 #include "paddle/phi/kernels/top_k_kernel.h"
 
 namespace phi {

paddle/phi/kernels/gpu/nll_loss_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/nll_loss_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpu/nll_loss.h"
-#include "paddle/phi/kernels/nll_loss_grad_kernel.h"
 
 namespace phi {
 template <typename T, typename Context>

paddle/phi/kernels/gpu/nll_loss_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/nll_loss_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpu/nll_loss.h"
-#include "paddle/phi/kernels/nll_loss_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/norm_grad_kernel.cu

@@ -12,9 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include <algorithm>
-
 #include "paddle/phi/kernels/norm_grad_kernel.h"
+
+#include <algorithm>
 #ifdef __NVCC__
 #include "cub/cub.cuh"
 #endif

paddle/phi/kernels/gpu/norm_kernel.cu

@@ -12,9 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include <algorithm>
-
 #include "paddle/phi/kernels/norm_kernel.h"
+
+#include <algorithm>
 #ifdef __NVCC__
 #include "cub/cub.cuh"
 #endif

paddle/phi/kernels/gpu/one_hot_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/one_hot_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_info.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/one_hot_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/p_norm_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/p_norm_grad_kernel.h"
+
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/funcs/reduce_grad_functions.h"
-#include "paddle/phi/kernels/p_norm_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/p_norm_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/p_norm_kernel.h"
+
 #include "paddle/fluid/operators/elementwise/elementwise_op_impl.cu.h"
 #include "paddle/phi/common/amp_type_traits.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/reduce_function.h"
 #include "paddle/phi/kernels/gpu/reduce.h"
-#include "paddle/phi/kernels/p_norm_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/pad3d_grad_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/pad3d_grad_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_info.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/pad3d_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/pad3d_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/pad3d_kernel.h"
+
 #include <algorithm>
 
 #include "paddle/fluid/platform/device/gpu/gpu_info.h"
@@ -19,7 +21,6 @@
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/pad3d_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/pad_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/pad_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/pad_grad_kernel_impl.h"
-#include "paddle/phi/kernels/pad_grad_kernel.h"
 
 PD_REGISTER_KERNEL(pad_grad,
                    GPU,

paddle/phi/kernels/gpu/pad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/pad_kernel.h"
+
 #include "paddle/fluid/platform/complex.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/pad_kernel_impl.h"
-#include "paddle/phi/kernels/pad_kernel.h"
 
 PD_REGISTER_KERNEL(pad,
                    GPU,

paddle/phi/kernels/gpu/pixel_shuffle_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/pixel_shuffle_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/pixel_shuffle_grad_kernel_impl.h"
-#include "paddle/phi/kernels/pixel_shuffle_grad_kernel.h"
 
 PD_REGISTER_KERNEL(pixel_shuffle_grad,
                    GPU,

paddle/phi/kernels/gpu/pixel_shuffle_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/pixel_shuffle_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/pixel_shuffle_kernel_impl.h"
-#include "paddle/phi/kernels/pixel_shuffle_kernel.h"
 
 PD_REGISTER_KERNEL(
     pixel_shuffle, GPU, ALL_LAYOUT, phi::PixelShuffleKernel, float, double) {}

paddle/phi/kernels/gpu/pixel_unshuffle_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/pixel_unshuffle_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/pixel_unshuffle_grad_kernel_impl.h"
-#include "paddle/phi/kernels/pixel_unshuffle_grad_kernel.h"
 
 PD_REGISTER_KERNEL(pixel_unshuffle_grad,
                    GPU,

paddle/phi/kernels/gpu/pixel_unshuffle_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/pixel_unshuffle_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/pixel_unshuffle_kernel_impl.h"
-#include "paddle/phi/kernels/pixel_unshuffle_kernel.h"
 
 PD_REGISTER_KERNEL(pixel_unshuffle,
                    GPU,

paddle/phi/kernels/gpu/pool_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/pool_grad_kernel.h"
+
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/pool_grad_kernel_impl.h"
-#include "paddle/phi/kernels/pool_grad_kernel.h"
 
 PD_REGISTER_KERNEL(pool2d_grad,
                    GPU,

paddle/phi/kernels/gpu/pool_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/pool_kernel.h"
+
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/pool_kernel_impl.h"
-#include "paddle/phi/kernels/pool_kernel.h"
 
 PD_REGISTER_KERNEL(pool2d,
                    GPU,

paddle/phi/kernels/gpu/prelu_grad_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/prelu_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/core/tensor_meta.h"
 #include "paddle/phi/kernels/empty_kernel.h"
 #include "paddle/phi/kernels/funcs/reduce_function.h"
 #include "paddle/phi/kernels/gpu/prelu_funcs.h"
-#include "paddle/phi/kernels/prelu_grad_kernel.h"
 #include "paddle/phi/kernels/primitive/functor_primitives.h"
 
 namespace phi {

paddle/phi/kernels/gpu/prelu_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/prelu_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpu/prelu_funcs.h"
-#include "paddle/phi/kernels/prelu_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/psroi_pool_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/psroi_pool_kernel.h"
+
 #include <algorithm>
 #include <vector>
 
@@ -19,7 +21,6 @@
 #include "paddle/phi/common/place.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
-#include "paddle/phi/kernels/psroi_pool_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/put_along_axis_grad_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/put_along_axis_grad_kernel.h"
+
 #include "paddle/fluid/framework/convert_utils.h"
 #include "paddle/fluid/operators/gather_scatter_kernel.h"
 #include "paddle/fluid/platform/place.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
-#include "paddle/phi/kernels/put_along_axis_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/put_along_axis_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/put_along_axis_kernel.h"
+
 #include "paddle/fluid/framework/convert_utils.h"
 #include "paddle/fluid/operators/gather_scatter_kernel.h"
 #include "paddle/fluid/platform/place.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
-#include "paddle/phi/kernels/put_along_axis_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/randint_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/randint_kernel.h"
+
 #include <random>
 
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/distribution_helper.h"
-#include "paddle/phi/kernels/randint_kernel.h"
 
 // See Note [ Why still include the fluid headers? ]
 #include "paddle/fluid/memory/memcpy.h"

paddle/phi/kernels/gpu/reduce_max_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/reduce_max_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/reduce_max_grad_kernel_impl.h"
-#include "paddle/phi/kernels/reduce_max_grad_kernel.h"
 
 PD_REGISTER_KERNEL(max_grad,
                    GPU,

paddle/phi/kernels/gpu/reduce_mean_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/reduce_mean_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/reduce_function.h"
 #include "paddle/phi/kernels/gpu/reduce_grad.h"
-#include "paddle/phi/kernels/reduce_mean_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/reduce_min_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/reduce_min_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/reduce_min_grad_kernel_impl.h"
-#include "paddle/phi/kernels/reduce_min_grad_kernel.h"
 
 PD_REGISTER_KERNEL(min_grad,
                    GPU,

paddle/phi/kernels/gpu/reduce_prod_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/reduce_prod_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/reduce_prod_grad_kernel_impl.h"
-#include "paddle/phi/kernels/reduce_prod_grad_kernel.h"
 
 PD_REGISTER_KERNEL(prod_grad,
                    GPU,

paddle/phi/kernels/gpu/reduce_sum_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/reduce_sum_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/reduce_function.h"
 #include "paddle/phi/kernels/gpu/reduce_grad.h"
-#include "paddle/phi/kernels/reduce_sum_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/rmsprop_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/rmsprop_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/rmsprop_kernel_impl.h"
-#include "paddle/phi/kernels/rmsprop_kernel.h"
 
 PD_REGISTER_KERNEL(
     rmsprop, GPU, ALL_LAYOUT, phi::RmspropDenseKernel, float, double) {}

paddle/phi/kernels/gpu/roi_align_grad_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/roi_align_grad_kernel.h"
+
 #include "paddle/fluid/memory/memory.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
@@ -20,7 +22,6 @@
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/empty_kernel.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/roi_align_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/roi_align_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/roi_align_kernel.h"
+
 #include "paddle/fluid/memory/memory.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/common/place.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/roi_align_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/roi_pool_grad_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/roi_pool_grad_kernel.h"
+
 #include "paddle/fluid/memory/memory.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
@@ -19,7 +21,6 @@
 #include "paddle/phi/common/place.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/roi_pool_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/roi_pool_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/roi_pool_kernel.h"
+
 #include "paddle/fluid/memory/memory.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/common/place.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/roi_pool_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/roll_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/roll_grad_kernel.h"
+
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpu/roll_kernel_impl.h"
-#include "paddle/phi/kernels/roll_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/roll_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/roll_kernel.h"
+
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/core/utils/array.h"
 #include "paddle/phi/kernels/gpu/roll_kernel_impl.h"
-#include "paddle/phi/kernels/roll_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/rrelu_grad_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/rrelu_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/core/tensor_meta.h"
@@ -19,7 +21,6 @@
 #include "paddle/phi/kernels/funcs/reduce_function.h"
 #include "paddle/phi/kernels/gpu/prelu_funcs.h"
 #include "paddle/phi/kernels/primitive/functor_primitives.h"
-#include "paddle/phi/kernels/rrelu_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/rrelu_kernel.cu

@@ -12,11 +12,12 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/rrelu_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/distribution_helper.h"
 #include "paddle/phi/kernels/funcs/for_range.h"
-#include "paddle/phi/kernels/rrelu_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/scale_kernel.cu

@@ -12,11 +12,12 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/scale_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
-#include "paddle/phi/kernels/scale_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/scatter_grad_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/scatter_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/gather.cu.h"
 #include "paddle/phi/kernels/funcs/scatter.cu.h"
-#include "paddle/phi/kernels/scatter_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/scatter_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/scatter_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/scatter.cu.h"
-#include "paddle/phi/kernels/scatter_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/scatter_nd_add_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/scatter_nd_add_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/gather.cu.h"
-#include "paddle/phi/kernels/scatter_nd_add_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/scatter_nd_add_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/scatter_nd_add_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/scatter.cu.h"
-#include "paddle/phi/kernels/scatter_nd_add_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/searchsorted_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/searchsorted_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/searchsorted_kernel_impl.h"
-#include "paddle/phi/kernels/searchsorted_kernel.h"
 
 PD_REGISTER_KERNEL(searchsorted,
                    GPU,

paddle/phi/kernels/gpu/segment_pool_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/segment_pool_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/segment_pool_grad_kernel_impl.h"
-#include "paddle/phi/kernels/segment_pool_grad_kernel.h"
 
 PD_REGISTER_KERNEL(segment_pool_grad,
                    GPU,

paddle/phi/kernels/gpu/segment_pool_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/segment_pool_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/segment_pool_kernel_impl.h"
-#include "paddle/phi/kernels/segment_pool_kernel.h"
 
 PD_REGISTER_KERNEL(segment_pool,
                    GPU,

paddle/phi/kernels/gpu/selu_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/selu_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/selu_grad_kernel_impl.h"
-#include "paddle/phi/kernels/selu_grad_kernel.h"
 
 PD_REGISTER_KERNEL(
     selu_grad, GPU, ALL_LAYOUT, phi::SeluGradKernel, float, double) {}

paddle/phi/kernels/gpu/selu_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/selu_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/selu_kernel_impl.h"
-#include "paddle/phi/kernels/selu_kernel.h"
 
 PD_REGISTER_KERNEL(selu, GPU, ALL_LAYOUT, phi::SeluKernel, float, double) {}

paddle/phi/kernels/gpu/set_value_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/set_value_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/set_value_grad_kernel_impl.h"
-#include "paddle/phi/kernels/set_value_grad_kernel.h"
 
 PD_REGISTER_KERNEL(set_value_grad,
                    GPU,

paddle/phi/kernels/gpu/set_value_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/set_value_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/set_value_kernel_impl.h"
-#include "paddle/phi/kernels/set_value_kernel.h"
 
 PD_REGISTER_KERNEL(set_value,
                    GPU,

paddle/phi/kernels/gpu/sgd_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/sgd_kernel.h"
+
 #include "paddle/fluid/framework/mixed_vector.h"
 #include "paddle/fluid/operators/amp/fp16_type_traits.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/backends/gpu/gpu_helper.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/sgd_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/shard_index_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/shard_index_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/shard_index_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/sigmoid_cross_entropy_with_logits_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/kernels/gpu/sigmoid_cross_entropy_with_logits.h"
 #include "paddle/phi/kernels/sigmoid_cross_entropy_with_logits_grad_kernel.h"
 
+#include "paddle/phi/kernels/gpu/sigmoid_cross_entropy_with_logits.h"
+
 namespace phi {
 
 template <typename T>

paddle/phi/kernels/gpu/sigmoid_cross_entropy_with_logits_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/kernels/gpu/sigmoid_cross_entropy_with_logits.h"
 #include "paddle/phi/kernels/sigmoid_cross_entropy_with_logits_kernel.h"
 
+#include "paddle/phi/kernels/gpu/sigmoid_cross_entropy_with_logits.h"
+
 namespace phi {
 
 template <typename T>

paddle/phi/kernels/gpu/size_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/size_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/size_kernel_impl.h"
-#include "paddle/phi/kernels/size_kernel.h"
 
 PD_REGISTER_KERNEL(size,
                    GPU,

paddle/phi/kernels/gpu/softmax_grad_kernel.cu

@@ -12,12 +12,13 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/softmax_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/softmax_grad_kernel_impl.h"
-#include "paddle/phi/kernels/softmax_grad_kernel.h"
 
 PD_REGISTER_KERNEL(softmax_grad,
                    GPU,

paddle/phi/kernels/gpu/softmax_kernel.cu

@@ -12,12 +12,13 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/softmax_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/softmax_kernel_impl.h"
-#include "paddle/phi/kernels/softmax_kernel.h"
 
 PD_REGISTER_KERNEL(softmax,
                    GPU,

paddle/phi/kernels/gpu/split_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/split_kernel.h"
+
 #include "paddle/fluid/operators/strided_memcpy.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/concat_and_split_functor.h"
-#include "paddle/phi/kernels/split_kernel.h"
 namespace phi {
 
 template <typename T, typename Context>

paddle/phi/kernels/gpu/squeeze_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/squeeze_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/squeeze_grad_kernel_impl.h"
-#include "paddle/phi/kernels/squeeze_grad_kernel.h"
 
 PD_REGISTER_KERNEL(squeeze_grad,
                    GPU,

paddle/phi/kernels/gpu/squeeze_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/squeeze_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/squeeze_kernel_impl.h"
-#include "paddle/phi/kernels/squeeze_kernel.h"
 
 PD_REGISTER_KERNEL(squeeze,
                    GPU,

paddle/phi/kernels/gpu/stack_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/stack_grad_kernel.h"
+
 #include "paddle/fluid/memory/memory.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/stack_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/stack_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/stack_kernel.h"
+
 #include "paddle/fluid/memory/memory.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/stack_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/strided_slice_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/strided_slice_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/strided_slice_grad_kernel_impl.h"
-#include "paddle/phi/kernels/strided_slice_grad_kernel.h"
 
 PD_REGISTER_KERNEL(strided_slice_raw_grad,
                    GPU,

paddle/phi/kernels/gpu/strided_slice_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/strided_slice_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/strided_slice_kernel_impl.h"
-#include "paddle/phi/kernels/strided_slice_kernel.h"
 
 PD_REGISTER_KERNEL(strided_slice_raw,
                    GPU,

paddle/phi/kernels/gpu/take_along_axis_grad_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/take_along_axis_grad_kernel.h"
+
 #include "paddle/fluid/framework/convert_utils.h"
 #include "paddle/fluid/operators/gather_scatter_kernel.h"
 #include "paddle/fluid/platform/place.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/take_along_axis_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/take_along_axis_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/take_along_axis_kernel.h"
+
 #include "paddle/fluid/framework/convert_utils.h"
 #include "paddle/fluid/operators/gather_scatter_kernel.h"
 #include "paddle/fluid/platform/place.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/take_along_axis_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/temporal_shift_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/temporal_shift_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/layout.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/temporal_shift_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/temporal_shift_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/temporal_shift_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/layout.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/temporal_shift_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/tile_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/tile_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/tile_grad_kernel_impl.h"
-#include "paddle/phi/kernels/tile_grad_kernel.h"
 
 PD_REGISTER_KERNEL(tile_grad,
                    GPU,

paddle/phi/kernels/gpu/tile_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/tile_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/tile_kernel_impl.h"
-#include "paddle/phi/kernels/tile_kernel.h"
 
 PD_REGISTER_KERNEL(tile,
                    GPU,

paddle/phi/kernels/gpu/top_k_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/top_k_grad_kernel.h"
+
 #include "paddle/fluid/operators/top_k_function_cuda.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/top_k_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/top_k_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/top_k_kernel.h"
+
 #include "paddle/fluid/operators/top_k_function_cuda.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/gather.cu.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
-#include "paddle/phi/kernels/top_k_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/trace_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/trace_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/trace_grad_kernel_impl.h"
-#include "paddle/phi/kernels/trace_grad_kernel.h"
 
 PD_REGISTER_KERNEL(trace_grad,
                    GPU,

paddle/phi/kernels/gpu/trace_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/trace_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/diagonal.h"
 #include "paddle/phi/kernels/funcs/reduce_function.h"
-#include "paddle/phi/kernels/trace_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/transpose_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/transpose_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/transpose_grad_kernel_impl.h"
-#include "paddle/phi/kernels/transpose_grad_kernel.h"
 
 PD_REGISTER_KERNEL(transpose_grad,
                    GPU,

paddle/phi/kernels/gpu/transpose_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/transpose_kernel.h"
+
 #include <vector>
 
 #include "paddle/fluid/framework/gpu_utils.h"
@@ -22,7 +24,6 @@
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/transpose_grad_kernel_impl.h"
-#include "paddle/phi/kernels/transpose_kernel.h"
 
 namespace phi {
 template <typename T, typename Context>

paddle/phi/kernels/gpu/triangular_solve_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/triangular_solve_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/ddim.h"
 #include "paddle/phi/core/kernel_registry.h"
@@ -19,7 +21,6 @@
 #include "paddle/phi/kernels/expand_kernel.h"
 #include "paddle/phi/kernels/funcs/blas/blas.h"
 #include "paddle/phi/kernels/funcs/common_shape.h"
-#include "paddle/phi/kernels/triangular_solve_kernel.h"
 
 // See Note [ Why still include the fluid headers? ]
 #include "paddle/fluid/memory/allocation/allocator.h"

paddle/phi/kernels/gpu/tril_indices_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/tril_indices_kernel.h"
+
 #include <algorithm>
 #include <tuple>
 
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/tril_indices_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/trunc_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/trunc_grad_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/backends/gpu/gpu_info.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/trunc_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/trunc_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/trunc_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/backends/gpu/gpu_info.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/trunc_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/truncated_gaussian_random_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/truncated_gaussian_random_kernel.h"
+
 #include <thrust/device_vector.h>
 #include <thrust/host_vector.h>
 #include <thrust/random.h>
@@ -22,7 +24,6 @@
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/truncated_gaussian_random_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/unbind_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/unbind_kernel.h"
+
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/unbind_kernel_impl.h"
-#include "paddle/phi/kernels/unbind_kernel.h"
 
 PD_REGISTER_KERNEL(unbind,
                    GPU,

paddle/phi/kernels/gpu/unfold_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/unfold_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/unfold_grad_kernel_impl.h"
-#include "paddle/phi/kernels/unfold_grad_kernel.h"
 
 PD_REGISTER_KERNEL(
     unfold_grad, GPU, ALL_LAYOUT, phi::UnfoldGradKernel, float, double) {}

paddle/phi/kernels/gpu/unfold_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/unfold_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/unfold_kernel_impl.h"
-#include "paddle/phi/kernels/unfold_kernel.h"
 
 PD_REGISTER_KERNEL(unfold, GPU, ALL_LAYOUT, phi::UnfoldKernel, float, double) {}

paddle/phi/kernels/gpu/uniform_random_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/uniform_random_kernel.h"
+
 #include <thrust/random.h>
 
 #include "gflags/gflags.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/distribution_helper.h"
 #include "paddle/phi/kernels/funcs/index_impl.cu.h"
-#include "paddle/phi/kernels/uniform_random_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/unique_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/unique_kernel.h"
+
 #include <thrust/adjacent_difference.h>
 #include <thrust/device_vector.h>
 #include <thrust/execution_policy.h>
@@ -28,7 +30,6 @@
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/unique_functor.h"
-#include "paddle/phi/kernels/unique_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/unsqueeze_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/unsqueeze_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/unsqueeze_grad_kernel_impl.h"
-#include "paddle/phi/kernels/unsqueeze_grad_kernel.h"
 
 PD_REGISTER_KERNEL(unsqueeze_grad,
                    GPU,

paddle/phi/kernels/gpu/unsqueeze_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/unsqueeze_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/unsqueeze_kernel_impl.h"
-#include "paddle/phi/kernels/unsqueeze_kernel.h"
 
 PD_REGISTER_KERNEL(unsqueeze,
                    GPU,

paddle/phi/kernels/gpu/unstack_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/unstack_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/unstack_grad_kernel_impl.h"
-#include "paddle/phi/kernels/unstack_grad_kernel.h"
 
 PD_REGISTER_KERNEL(unstack_grad,
                    GPU,

paddle/phi/kernels/gpu/unstack_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/unstack_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/unstack_kernel_impl.h"
-#include "paddle/phi/kernels/unstack_kernel.h"
 
 PD_REGISTER_KERNEL(unstack,
                    GPU,

paddle/phi/kernels/gpu/warpctc_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/warpctc_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/warpctc_grad_kernel_impl.h"
-#include "paddle/phi/kernels/warpctc_grad_kernel.h"
 
 PD_REGISTER_KERNEL(
     warpctc_grad, GPU, ALL_LAYOUT, phi::WarpctcGradKernel, float, double) {}

paddle/phi/kernels/gpu/warpctc_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/warpctc_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/warpctc_kernel_impl.h"
-#include "paddle/phi/kernels/warpctc_kernel.h"
 
 PD_REGISTER_KERNEL(
     warpctc, GPU, ALL_LAYOUT, phi::WarpctcKernel, float, double) {}

paddle/phi/kernels/gpu/where_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/where_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/where_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/where_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/where_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/broadcast_function.h"
 #include "paddle/phi/kernels/funcs/elementwise_functor.h"
-#include "paddle/phi/kernels/where_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/yolo_box_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/yolo_box_kernel.h"
+
 #include "paddle/fluid/memory/malloc.h"
 #include "paddle/fluid/memory/memcpy.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
@@ -19,7 +21,6 @@
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/funcs/yolo_box_util.h"
-#include "paddle/phi/kernels/yolo_box_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpudnn/conv_grad_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/conv_grad_grad_kernel.h"
+
 #include "paddle/fluid/framework/eigen.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_grad_grad_kernel.h"
 #ifdef PADDLE_WITH_HIP
 #include "paddle/fluid/operators/conv_miopen_helper.h"
 #else

paddle/phi/kernels/gpudnn/conv_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/conv_grad_kernel.h"
+
 #include "paddle/fluid/framework/eigen.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_grad_kernel.h"
 #ifdef PADDLE_WITH_HIP
 #include "paddle/fluid/operators/conv_miopen_helper.h"
 #else

paddle/phi/kernels/gpudnn/conv_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/conv_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_kernel.h"
 
 #ifdef PADDLE_WITH_HIP
 #include "paddle/fluid/operators/conv_miopen_helper.h"

paddle/phi/kernels/gpudnn/conv_transpose_grad_kernel.cu

@@ -12,13 +12,14 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/conv_transpose_grad_kernel.h"
+
 #include <algorithm>
 
 #include "paddle/phi/backends/dynload/cudnn.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/ddim.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_transpose_grad_kernel.h"
 #include "paddle/phi/kernels/cpu/conv_util.h"
 #include "paddle/phi/kernels/funcs/batch_norm_utils.h"
 #include "paddle/phi/kernels/funcs/math_function.h"

paddle/phi/kernels/gpudnn/conv_transpose_kernel.cu

@@ -12,13 +12,14 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/conv_transpose_kernel.h"
+
 #include <algorithm>
 
 #include "paddle/phi/backends/dynload/cudnn.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/ddim.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_transpose_kernel.h"
 #include "paddle/phi/kernels/cpu/conv_util.h"
 #include "paddle/phi/kernels/funcs/padding.h"
 #include "paddle/phi/kernels/funcs/slice.h"

paddle/phi/kernels/gpudnn/pool_grad_kernel.cu

@@ -12,13 +12,14 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/pool_grad_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_dnn.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/funcs/pooling.h"
 #include "paddle/phi/kernels/gpudnn/pool_gpudnn.h"
-#include "paddle/phi/kernels/pool_grad_kernel.h"
 #include "paddle/phi/kernels/pool_kernel.h"
 
 #ifdef PADDLE_WITH_HIP

paddle/phi/kernels/gpudnn/pool_kernel.cu

@@ -12,12 +12,13 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/pool_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_dnn.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/funcs/pooling.h"
 #include "paddle/phi/kernels/gpudnn/pool_gpudnn.h"
-#include "paddle/phi/kernels/pool_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpudnn/softmax_grad_kernel.cu

@@ -12,10 +12,11 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/softmax_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpudnn/softmax_gpudnn.h"
-#include "paddle/phi/kernels/softmax_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpudnn/softmax_kernel.cu

@@ -12,10 +12,11 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/softmax_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpudnn/softmax_gpudnn.h"
-#include "paddle/phi/kernels/softmax_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/selected_rows/gpu/adam_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/selected_rows/adam_kernel.h"
+
 #include "paddle/fluid/framework/tensor_util.h"
 #include "paddle/fluid/operators/math/selected_rows_functor.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
@@ -21,7 +23,6 @@
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/adam_functors.h"
 #include "paddle/phi/kernels/funcs/for_range.h"
-#include "paddle/phi/kernels/selected_rows/adam_kernel.h"
 
 namespace phi {
 namespace sr {

paddle/phi/kernels/selected_rows/gpu/adamw_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/selected_rows/adamw_kernel.h"
+
 #include <math.h>  // for sqrt in CPU and CUDA
 
 #include <vector>
@@ -25,7 +27,6 @@
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/adam_functors.h"
 #include "paddle/phi/kernels/funcs/for_range.h"
-#include "paddle/phi/kernels/selected_rows/adamw_kernel.h"
 
 namespace phi {
 namespace sr {

paddle/phi/kernels/selected_rows/gpu/clip_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/selected_rows/clip_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/selected_rows/clip_kernel.h"
 #include "paddle/phi/kernels/selected_rows/impl/clip_kernel_impl.h"
 
 PD_REGISTER_KERNEL(clip_sr,

paddle/phi/kernels/sparse/gpu/coalesced_kernel.cu

@@ -12,6 +12,8 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/sparse/coalesced_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_info.h"
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/core/kernel_registry.h"
@@ -20,7 +22,6 @@ limitations under the License. */
 #include "paddle/phi/kernels/funcs/sparse/flatten_indices.cu.h"
 #include "paddle/phi/kernels/funcs/sparse/scatter.cu.h"
 #include "paddle/phi/kernels/funcs/sparse/utils.cu.h"
-#include "paddle/phi/kernels/sparse/coalesced_kernel.h"
 
 namespace phi {
 namespace sparse {

paddle/phi/kernels/sparse/gpu/convolution_grad_kernel.cu

@@ -12,6 +12,8 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/sparse/convolution_grad_kernel.h"
+
 #include "glog/logging.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/backends/gpu/gpu_info.h"
@@ -23,7 +25,6 @@ limitations under the License. */
 #include "paddle/phi/kernels/funcs/blas/blas.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/funcs/scatter.cu.h"
-#include "paddle/phi/kernels/sparse/convolution_grad_kernel.h"
 #include "paddle/phi/kernels/sparse/gpu/convolution.cu.h"
 
 namespace phi {

paddle/phi/kernels/sparse/gpu/convolution_kernel.cu

@@ -12,6 +12,8 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/sparse/convolution_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/core/tensor_meta.h"
@@ -19,7 +21,6 @@ limitations under the License. */
 #include "paddle/phi/kernels/funcs/blas/blas.h"
 #include "paddle/phi/kernels/funcs/scatter.cu.h"
 #include "paddle/phi/kernels/funcs/sparse/scatter.cu.h"
-#include "paddle/phi/kernels/sparse/convolution_kernel.h"
 #include "paddle/phi/kernels/sparse/gpu/convolution.cu.h"
 
 namespace phi {

paddle/phi/kernels/sparse/gpu/full_kernel.cu

@@ -12,12 +12,13 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/sparse/full_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
-#include "paddle/phi/kernels/sparse/full_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/sparse/gpu/matmul_grad_kernel.cu

@@ -12,6 +12,8 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/sparse/matmul_grad_kernel.h"
+
 #include <vector>
 
 #include "paddle/phi/backends/gpu/gpu_context.h"
@@ -19,7 +21,6 @@ limitations under the License. */
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/empty_kernel.h"
 #include "paddle/phi/kernels/funcs/sparse/sparse_blas.h"
-#include "paddle/phi/kernels/sparse/matmul_grad_kernel.h"
 #include "paddle/phi/kernels/transpose_kernel.h"
 
 namespace phi {

paddle/phi/kernels/sparse/gpu/matmul_kernel.cu

@@ -12,6 +12,8 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/sparse/matmul_kernel.h"
+
 #include <vector>
 
 #include "paddle/phi/backends/gpu/gpu_context.h"
@@ -24,7 +26,6 @@ limitations under the License. */
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/empty_kernel.h"
 #include "paddle/phi/kernels/funcs/sparse/sparse_blas.h"
-#include "paddle/phi/kernels/sparse/matmul_kernel.h"
 
 namespace phi {
 namespace sparse {

paddle/phi/kernels/sparse/gpu/sparse_mask_kernel.cu

@@ -12,6 +12,8 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/sparse/sparse_mask_kernel.h"
+
 #include <thrust/binary_search.h>
 
 #include "paddle/phi/backends/gpu/gpu_info.h"
@@ -24,7 +26,6 @@ limitations under the License. */
 #include "paddle/phi/kernels/empty_kernel.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/funcs/sparse/flatten_indices.cu.h"
-#include "paddle/phi/kernels/sparse/sparse_mask_kernel.h"
 
 namespace phi {
 namespace sparse {

paddle/phi/kernels/sparse/gpu/sparse_pool_grad_kernel.cu

@@ -12,6 +12,8 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/sparse/sparse_pool_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/backends/gpu/gpu_info.h"
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
@@ -22,7 +24,6 @@ limitations under the License. */
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/funcs/pooling.h"
 #include "paddle/phi/kernels/funcs/sparse/convolution.h"
-#include "paddle/phi/kernels/sparse/sparse_pool_grad_kernel.h"
 
 namespace phi {
 namespace sparse {

paddle/phi/kernels/sparse/gpu/sparse_pool_kernel.cu

@@ -12,13 +12,14 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/sparse/sparse_pool_kernel.h"
+
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/core/tensor_meta.h"
 #include "paddle/phi/core/visit_type.h"
 #include "paddle/phi/kernels/funcs/pooling.h"
 #include "paddle/phi/kernels/funcs/sparse/convolution.h"
 #include "paddle/phi/kernels/sparse/gpu/convolution.cu.h"
-#include "paddle/phi/kernels/sparse/sparse_pool_kernel.h"
 
 namespace phi {
 namespace sparse {

paddle/phi/kernels/sparse/gpu/sparse_utils_kernel.cu

@@ -12,6 +12,8 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/sparse/sparse_utils_kernel.h"
+
 #include <thrust/execution_policy.h>
 #include <thrust/remove.h>
 
@@ -23,7 +25,6 @@ limitations under the License. */
 #include "paddle/phi/core/visit_type.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/funcs/sparse/common_shape.h"
-#include "paddle/phi/kernels/sparse/sparse_utils_kernel.h"
 
 namespace phi {
 namespace sparse {

paddle/phi/kernels/strings/gpu/strings_copy_kernel.cu

@@ -12,6 +12,8 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/strings/strings_copy_kernel.h"
+
 #include "paddle/phi/backends/all_context.h"
 #include "paddle/phi/backends/gpu/gpu_helper.h"
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
@@ -20,7 +22,6 @@ limitations under the License. */
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/empty_kernel.h"
 #include "paddle/phi/kernels/strings/gpu/copy_utils.h"
-#include "paddle/phi/kernels/strings/strings_copy_kernel.h"
 
 using pstring = ::phi::dtype::pstring;
 

paddle/phi/kernels/strings/gpu/strings_lower_upper_kernel.cu

@@ -9,11 +9,12 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
+#include "paddle/phi/kernels/strings/strings_lower_upper_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/common/pstring.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/strings/strings_lower_upper_kernel.h"
 #include "paddle/phi/kernels/strings/unicode.h"
 
 using pstring = ::phi::dtype::pstring;