提交 d23a1712 编写于 作者: L liuruilong

format files

上级 7b95cd02
...@@ -17,8 +17,8 @@ limitations under the License. */ ...@@ -17,8 +17,8 @@ limitations under the License. */
#pragma once #pragma once
#include <string> #include <string>
#include <vector>
#include <utility> #include <utility>
#include <vector>
#include "framework/operator.h" #include "framework/operator.h"
#include "framework/program/program-optimize/fusion_op_register.h" #include "framework/program/program-optimize/fusion_op_register.h"
#include "operators/kernel/conv_add_add_prelu_kernel.h" #include "operators/kernel/conv_add_add_prelu_kernel.h"
......
...@@ -19,8 +19,8 @@ limitations under the License. */ ...@@ -19,8 +19,8 @@ limitations under the License. */
#include <vector> #include <vector>
#include "framework/operator.h" #include "framework/operator.h"
#include "framework/program/program-optimize/fusion_op_register.h" #include "framework/program/program-optimize/fusion_op_register.h"
#include "operators/op_param.h"
#include "operators/kernel/conv_add_kernel.h" #include "operators/kernel/conv_add_kernel.h"
#include "operators/op_param.h"
namespace paddle_mobile { namespace paddle_mobile {
namespace operators { namespace operators {
......
...@@ -39,8 +39,7 @@ class FusionConvAddPReluOpMatcher : public framework::FusionOpMatcher { ...@@ -39,8 +39,7 @@ class FusionConvAddPReluOpMatcher : public framework::FusionOpMatcher {
std::vector<std::shared_ptr<framework::Node>> *removed_nodes) { std::vector<std::shared_ptr<framework::Node>> *removed_nodes) {
node->Folder(node_.Depth(), Type(), node->Folder(node_.Depth(), Type(),
{{G_OP_TYPE_ELEMENTWISE_ADD, {{"Y", "Y"}}}, {{G_OP_TYPE_ELEMENTWISE_ADD, {{"Y", "Y"}}},
{G_OP_TYPE_PRELU, {{"Alpha", "Alpha"}}} {G_OP_TYPE_PRELU, {{"Alpha", "Alpha"}}}},
},
removed_nodes); removed_nodes);
} }
std::string Type() { return G_OP_TYPE_FUSION_CONV_ADD_PRELU; } std::string Type() { return G_OP_TYPE_FUSION_CONV_ADD_PRELU; }
......
...@@ -17,12 +17,12 @@ limitations under the License. */ ...@@ -17,12 +17,12 @@ limitations under the License. */
#pragma once #pragma once
#include <string> #include <string>
#include <vector>
#include <utility> #include <utility>
#include <vector>
#include "framework/operator.h" #include "framework/operator.h"
#include "framework/program/program-optimize/fusion_op_register.h" #include "framework/program/program-optimize/fusion_op_register.h"
#include "operators/op_param.h"
#include "operators/kernel/conv_bn_add_relu_kernel.h" #include "operators/kernel/conv_bn_add_relu_kernel.h"
#include "operators/op_param.h"
namespace paddle_mobile { namespace paddle_mobile {
namespace operators { namespace operators {
......
...@@ -20,8 +20,8 @@ limitations under the License. */ ...@@ -20,8 +20,8 @@ limitations under the License. */
#include <vector> #include <vector>
#include "framework/operator.h" #include "framework/operator.h"
#include "framework/program/program-optimize/fusion_op_register.h" #include "framework/program/program-optimize/fusion_op_register.h"
#include "operators/op_param.h"
#include "operators/kernel/dwconv_bn_relu_kernel.h" #include "operators/kernel/dwconv_bn_relu_kernel.h"
#include "operators/op_param.h"
namespace paddle_mobile { namespace paddle_mobile {
namespace operators { namespace operators {
......
...@@ -17,9 +17,9 @@ limitations under the License. */ ...@@ -17,9 +17,9 @@ limitations under the License. */
#pragma once #pragma once
#include <string> #include <string>
#include "operators/op_param.h"
#include "framework/operator.h" #include "framework/operator.h"
#include "operators/kernel/im2sequence_kernel.h" #include "operators/kernel/im2sequence_kernel.h"
#include "operators/op_param.h"
namespace paddle_mobile { namespace paddle_mobile {
namespace operators { namespace operators {
......
...@@ -49,11 +49,14 @@ void ElementwiseAddKernel<GPU_CL, float>::Compute( ...@@ -49,11 +49,14 @@ void ElementwiseAddKernel<GPU_CL, float>::Compute(
cl_mem input_image = input->GetCLImage(); cl_mem input_image = input->GetCLImage();
cl_mem bias_image = bias->GetCLImage(); cl_mem bias_image = bias->GetCLImage();
cl_mem output_image = output->GetCLImage(); cl_mem output_image = output->GetCLImage();
status = clSetKernelArg(kernel, 0, sizeof(cl_mem), reinterpret_cast<void *>(&input_image)); status = clSetKernelArg(kernel, 0, sizeof(cl_mem),
reinterpret_cast<void *>(&input_image));
CL_CHECK_ERRORS(status); CL_CHECK_ERRORS(status);
status = clSetKernelArg(kernel, 1, sizeof(cl_mem), reinterpret_cast<void *>(&bias_image)); status = clSetKernelArg(kernel, 1, sizeof(cl_mem),
reinterpret_cast<void *>(&bias_image));
CL_CHECK_ERRORS(status); CL_CHECK_ERRORS(status);
status = clSetKernelArg(kernel, 2, sizeof(cl_mem), reinterpret_cast<void *>(&output_image)); status = clSetKernelArg(kernel, 2, sizeof(cl_mem),
reinterpret_cast<void *>(&output_image));
CL_CHECK_ERRORS(status); CL_CHECK_ERRORS(status);
int width = input->ImageWidth(); int width = input->ImageWidth();
int height = input->ImageHeight(); int height = input->ImageHeight();
...@@ -67,13 +70,17 @@ void ElementwiseAddKernel<GPU_CL, float>::Compute( ...@@ -67,13 +70,17 @@ void ElementwiseAddKernel<GPU_CL, float>::Compute(
cl_mem bias_image = bias->GetCLImage(); cl_mem bias_image = bias->GetCLImage();
cl_mem output_image = output->GetCLImage(); cl_mem output_image = output->GetCLImage();
int tensor_w = input->dims()[3]; int tensor_w = input->dims()[3];
status = clSetKernelArg(kernel, 0, sizeof(cl_mem), reinterpret_cast<void *>(&input_image)); status = clSetKernelArg(kernel, 0, sizeof(cl_mem),
reinterpret_cast<void *>(&input_image));
CL_CHECK_ERRORS(status); CL_CHECK_ERRORS(status);
status = clSetKernelArg(kernel, 1, sizeof(cl_mem), reinterpret_cast<void *>(&bias_image)); status = clSetKernelArg(kernel, 1, sizeof(cl_mem),
reinterpret_cast<void *>(&bias_image));
CL_CHECK_ERRORS(status); CL_CHECK_ERRORS(status);
status = clSetKernelArg(kernel, 2, sizeof(cl_mem), reinterpret_cast<void *>(&output_image)); status = clSetKernelArg(kernel, 2, sizeof(cl_mem),
reinterpret_cast<void *>(&output_image));
CL_CHECK_ERRORS(status); CL_CHECK_ERRORS(status);
status = clSetKernelArg(kernel, 3, sizeof(cl_int), reinterpret_cast<void *>(&tensor_w)); status = clSetKernelArg(kernel, 3, sizeof(cl_int),
reinterpret_cast<void *>(&tensor_w));
CL_CHECK_ERRORS(status); CL_CHECK_ERRORS(status);
int width = input->ImageWidth(); int width = input->ImageWidth();
int height = input->ImageHeight(); int height = input->ImageHeight();
......
...@@ -14,8 +14,8 @@ limitations under the License. */ ...@@ -14,8 +14,8 @@ limitations under the License. */
#ifdef POOL_OP #ifdef POOL_OP
#include <vector>
#include "operators/pool_op.h" #include "operators/pool_op.h"
#include <vector>
#include "framework/op_proto_maker.h" #include "framework/op_proto_maker.h"
#include "framework/op_registry.h" #include "framework/op_registry.h"
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册