wehub-resource-sync
53 lines
1.9 KiB
C++
53 lines
1.9 KiB
C++
//
|
|
// DequantizeTflite.cpp
|
|
// MNNConverter
|
|
//
|
|
// Created by MNN on 2020/05/27.
|
|
// Copyright © 2018, Alibaba Group Holding Limited
|
|
//
|
|
|
|
#include "liteOpConverter.hpp"
|
|
#include "TfliteUtils.hpp"
|
|
|
|
|
|
DECLARE_OP_COVERTER(DequantizeTflite);
|
|
|
|
MNN::OpType DequantizeTflite::opType(int quantizedModel){
|
|
return MNN::OpType_Dequantize;
|
|
}
|
|
|
|
MNN::OpParameter DequantizeTflite::type(int quantizedModel){
|
|
return MNN::OpParameter_Dequantize;
|
|
}
|
|
|
|
void DequantizeTflite::run(MNN::OpT *dstOp, const std::unique_ptr<tflite::OperatorT> &tfliteOp, const std::vector<std::unique_ptr<tflite::TensorT> > &tfliteTensors, const std::vector<std::unique_ptr<tflite::BufferT> > &tfliteModelBuffer, const std::vector<std::unique_ptr<tflite::OperatorCodeT> > &tfliteOpSet, int quantizedModel){
|
|
DCHECK(1 == tfliteOp->inputs.size()) << "Dequantize should have one input now";
|
|
auto inputIndex = tfliteOp->inputs[0];
|
|
const auto& inputTensor = tfliteTensors[inputIndex];
|
|
const auto& outputTensor = tfliteTensors[tfliteOp->outputs[0]];
|
|
if (inputTensor->type != tflite::TensorType_UINT8) {
|
|
if (outputTensor->type != tflite::TensorType_UINT8) {
|
|
dstOp->type = MNN::OpType_Identity;
|
|
dstOp->main.type = MNN::OpParameter_NONE;
|
|
return;
|
|
}
|
|
}
|
|
|
|
auto dequantizeParam = new MNN::DequantizeT;
|
|
|
|
dequantizeParam->modelFormat = MNN::ModeFormat_TFLITE;
|
|
|
|
dequantizeParam->type = TfliteDequantDataTypeToMNN(inputTensor->type);
|
|
|
|
auto quantizedParam = new MNN::QuantizedParamT;
|
|
|
|
quantizedParam->zeroPoint = static_cast<int32_t>(inputTensor->quantization->zero_point[0]);
|
|
quantizedParam->scale = inputTensor->quantization->scale[0];
|
|
dequantizeParam->inputQuantizedParam = std::unique_ptr<MNN::QuantizedParamT>(quantizedParam);
|
|
|
|
dstOp->main.value = dequantizeParam;
|
|
}
|
|
|
|
using namespace tflite;
|
|
REGISTER_CONVERTER(DequantizeTflite, BuiltinOperator_DEQUANTIZE);
|