wehub-resource-sync
113 lines
4.2 KiB
C++
113 lines
4.2 KiB
C++
//
|
|
// NPUConvolutionDepthwiseInt8.cpp
|
|
// MNN
|
|
//
|
|
// Created by MNN on b'2020/10/15'.
|
|
// Copyright © 2018, Alibaba Group Holding Limited
|
|
//
|
|
|
|
#include "NPUConvolutionDepthwiseInt8.hpp"
|
|
#include "NPUBackend.hpp"
|
|
#include "../3rdParty/include/graph/types.h"
|
|
#include <core/TensorUtils.hpp>
|
|
|
|
using namespace std;
|
|
|
|
namespace MNN {
|
|
|
|
NPUConvolutionDepthwiseInt8::NPUConvolutionDepthwiseInt8(Backend *b, const Op *op, const std::vector<Tensor *> &inputs,
|
|
const std::vector<Tensor *> &outputs)
|
|
: MNN::NPUCommonExecution(b,op) {}
|
|
|
|
ErrorCode NPUConvolutionDepthwiseInt8::onResize(const std::vector<Tensor *> &inputs, const std::vector<Tensor *> &outputs) {
|
|
mNpuBackend->setNetworkInput(inputs, mOp);
|
|
|
|
auto opName = mOp->name()->str();
|
|
|
|
auto conv2D = mOp->main_as_Convolution2D();
|
|
auto conv2DCommon = conv2D->common();
|
|
auto quantizedParams = conv2D->symmetricQuan();
|
|
|
|
auto kernelX = conv2DCommon->kernelX();
|
|
auto kernelY = conv2DCommon->kernelY();
|
|
auto outputCount = conv2DCommon->outputCount();
|
|
int weightSize = quantizedParams->weight()->size();
|
|
int inputCount = weightSize / (kernelX * kernelY * outputCount);
|
|
|
|
shared_ptr<hiai::op::QuantizedConvolutionDepthwise> conv(new hiai::op::QuantizedConvolutionDepthwise(opName));
|
|
|
|
auto int32ToInt8Scale = quantizedParams->scale()->data();
|
|
|
|
vector<float> filter_scale(int32ToInt8Scale, int32ToInt8Scale + quantizedParams->scale()->size());
|
|
|
|
auto xOp = mNpuBackend->getInputOps(mOp);
|
|
|
|
// om input weight const op
|
|
mConst_w = hiai::op::Const(opName + "_w_const");
|
|
{
|
|
ge::TensorDesc fdesc(ge::Shape({outputCount, inputCount, kernelY, kernelX}), ge::FORMAT_NCHW,
|
|
ge::DT_INT8); // in o h w ?
|
|
ge::TensorPtr filter = std::make_shared<ge::Tensor>();
|
|
filter->SetTensorDesc(fdesc);
|
|
filter->SetData((uint8_t *)quantizedParams->weight()->data(), weightSize);
|
|
|
|
mConst_w.set_attr_value(filter);
|
|
}
|
|
// om input bias const op
|
|
mConst_b = hiai::op::Const(opName + "_b_const");
|
|
{
|
|
ge::TensorDesc fdesc(ge::Shape({1, outputCount, 1, 1}), ge::FORMAT_NCHW, ge::DT_INT32);
|
|
ge::TensorPtr bias = std::make_shared<ge::Tensor>();
|
|
bias->SetTensorDesc(fdesc);
|
|
bias->SetData((uint8_t *)quantizedParams->bias()->data(), quantizedParams->bias()->size()*
|
|
sizeof(int32_t));
|
|
mConst_b.set_attr_value(bias);
|
|
}
|
|
|
|
auto padMode = "SPECIFIC"; // NOTSET
|
|
vector<int64_t> pad = {conv2DCommon->padY(), conv2DCommon->padY(), conv2DCommon->padX(), conv2DCommon->padX()};
|
|
if (PadMode_VALID == conv2DCommon->padMode() || PadMode_CAFFE == conv2DCommon->padMode() ) {
|
|
padMode = "VALID";
|
|
} else if (PadMode_SAME == conv2DCommon->padMode()) {
|
|
padMode = "SAME";
|
|
pad = {0,0,0,0};
|
|
}
|
|
|
|
(*conv)
|
|
.set_input_x(*xOp.get())
|
|
.set_input_filter(mConst_w)
|
|
.set_input_bias(mConst_b)
|
|
.set_attr_strides(ge::AttrValue::LIST_INT({conv2DCommon->strideY(), conv2DCommon->strideX()}))
|
|
.set_attr_dilations(ge::AttrValue::LIST_INT({conv2DCommon->dilateY(), conv2DCommon->dilateX()}))
|
|
.set_attr_pads(ge::AttrValue::LIST_INT(pad)) // 上下左右
|
|
.set_attr_pad_mode(padMode)
|
|
.set_attr_filter_quant_type(1)
|
|
.set_attr_x_quant_type(1)
|
|
.set_attr_x_quant_offset(127)
|
|
// .set_attr_x_quant_offset(0)
|
|
.set_attr_x_quant_scale(1.0)
|
|
.set_attr_filter_quant_scales(filter_scale);
|
|
|
|
shared_ptr<hiai::op::Activation> relu_conv(new hiai::op::Activation(opName + "_Relu"));
|
|
mRelu_conv = relu_conv;
|
|
|
|
auto relu = conv2DCommon->relu();
|
|
auto relu6 = conv2DCommon->relu6();
|
|
if (relu || relu6) {
|
|
(*relu_conv)
|
|
.set_input_x(*conv.get())
|
|
.set_attr_mode(relu?1:14);
|
|
}
|
|
|
|
if (relu || relu6) {
|
|
mNpuBackend->setOutputOps(mOp, {conv, mRelu_conv}, outputs);
|
|
}else{
|
|
mNpuBackend->setOutputOps(mOp, {conv}, outputs);
|
|
}
|
|
return NO_ERROR;
|
|
}
|
|
|
|
NPUCreatorRegister<TypedCreator<NPUConvolutionDepthwiseInt8>> __depthwise_conv_int8_op(OpType_DepthwiseConvInt8);
|
|
|
|
} // namespace MNN
|