MNN/source/backend/tensorrt/execution/TRTCast.cpp

52 lines
1.6 KiB
C++
Executable File

//
// TRTCast.cpp
// MNN
//
// Created by MNN on 2019/09/11.
// Copyright © 2018, Alibaba Group Holding Limited
//
#include "TRTCast.hpp"
#include <core/TensorUtils.hpp>
#include "TRTBackend.hpp"
#include "schema/current/MNNPlugin_generated.h"
using namespace std;
namespace MNN {
TRTCast::TRTCast(Backend *b, const Op *op, const std::vector<Tensor *> &inputs,
const std::vector<Tensor *> &outputs)
: MNN::TRTCommonExecution(b, op) {
}
std::vector<ITensor *> TRTCast::onEncode(const std::vector<ITensor *> &xOp) {
auto plu = createPluginWithOutput(mOutputs);
auto castPara = mOp->main_as_CastParam();
DataType srcT = castPara->srcT();
DataType dstT = castPara->dstT();
plu->main.type = MNNTRTPlugin::Parameter_OneHotInfo;
plu->main.value = new MNNTRTPlugin::OneHotInfoT;
auto onehotp = plu->main.AsOneHotInfo();
onehotp->outerSize = mInputs[0]->elementSize();
if((srcT == DataType_DT_INT32 || srcT == DataType_DT_INT64) && dstT == DataType_DT_FLOAT){
auto interpPlugin = (nvinfer1::IPluginExt *)MNNTRTCreatePlugion(mOp, plu.get());
nvinfer1::IPluginLayer *plugin = mTrtBackend->getNetwork()->addPluginExt(&xOp[0], 1, *((nvinfer1::IPluginExt *)interpPlugin));
if (plugin == nullptr) {
printf("Interp plugin == nullptr !!!\n");
}
mTrtBackend->pushReleaseLayer(interpPlugin);
return {plugin->getOutput(0)};
}else{
MNN_ASSERT(false);
return {};
}
}
TRTCreatorRegister<TypedCreator<TRTCast>> __cast_op(OpType_Cast);
} // namespace MNN