mirror of https://github.com/alibaba/MNN.git
37 lines
1013 B
C++
37 lines
1013 B
C++
//
|
|
// MultiInputConvExecution.hpp
|
|
// MNN
|
|
//
|
|
// Created by MNN on 2023/03/20.
|
|
// Copyright © 2018, Alibaba Group Holding Limited
|
|
//
|
|
#ifndef MultiInputConvExecution_hpp
|
|
#define MultiInputConvExecution_hpp
|
|
|
|
#include "backend/cuda/core/CUDABackend.hpp"
|
|
#include "core/Execution.hpp"
|
|
#include "CutlassGemmParam.hpp"
|
|
#include "MNNCUDADefine.hpp"
|
|
#include "MNNCUDAFunction.cuh"
|
|
#include "cutlass/CutlassConvCommonExecution.hpp"
|
|
|
|
namespace MNN {
|
|
namespace CUDA {
|
|
|
|
class MultiInputConvExecution : public CutlassConvCommonExecution {
|
|
public:
|
|
MultiInputConvExecution(const MNN::Op* op, Backend* backend);
|
|
virtual ~MultiInputConvExecution();
|
|
virtual ErrorCode onResize(const std::vector<Tensor*> &inputs, const std::vector<Tensor*> &outputs) override;
|
|
virtual ErrorCode onExecute(const std::vector<Tensor*> &inputs, const std::vector<Tensor*> &outputs) override;
|
|
|
|
private:
|
|
bool mNeedWeightFill;
|
|
bool mNeedBiasFill;
|
|
};
|
|
|
|
|
|
} // namespace CUDA
|
|
} // namespace MNN
|
|
|
|
#endif /* MultiInputConvExecution */ |