MNN/source/backend/cpu/CPUQuantizedLogistic.cpp

65 lines
2.5 KiB
C++
Raw Normal View History

2019-04-17 10:49:11 +08:00
//
// CPUQuantizedLogistic.cpp
// MNN
//
// Created by MNN on 2018/12/12.
// Copyright © 2018, Alibaba Group Holding Limited
//
2019-12-27 22:16:57 +08:00
#include "backend/cpu/CPUBackend.hpp"
#ifdef MNN_SUPPORT_DEPRECATED_OP
#include "backend/cpu/CPUQuantizedLogistic.hpp"
2019-12-27 22:16:57 +08:00
#include "backend/cpu/CPUFixedPoint.hpp"
#include "backend/cpu/CPUQuantizationUtils.hpp"
#include "core/Macro.h"
#include "backend/cpu/compute/OptimizedComputer.hpp"
2019-04-17 10:49:11 +08:00
namespace MNN {
CPUQuantizedLogistic::CPUQuantizedLogistic(Backend *backend, const Op *op) : Execution(backend) {
mLogisticParam = op->main_as_QuantizedLogistic();
}
ErrorCode CPUQuantizedLogistic::onResize(const std::vector<Tensor *> &inputs, const std::vector<Tensor *> &outputs) {
MNN_ASSERT(1 == inputs.size() && 1 == outputs.size());
MNN_ASSERT(0 == mLogisticParam->outputQuantizedParam()->zeroPoint() &&
1. / 256 == mLogisticParam->outputQuantizedParam()->scale());
static constexpr int kInputIntegerBits = 4;
const double inputRealMultiplier =
mLogisticParam->inputQuantizedParam()->scale() * static_cast<double>(1 << (31 - kInputIntegerBits));
QuantizeMultiplierGreaterThanOne(inputRealMultiplier, &mInputMultiplier, &mInputLeftShift);
mInputZeroPoint = mLogisticParam->inputQuantizedParam()->zeroPoint();
2019-04-17 10:49:11 +08:00
mInputRangeRadius = CalculateInputRadius(kInputIntegerBits, mInputLeftShift);
return NO_ERROR;
}
ErrorCode CPUQuantizedLogistic::onExecute(const std::vector<MNN::Tensor *> &inputs,
const std::vector<MNN::Tensor *> &outputs) {
auto input = inputs[0], output = outputs[0];
std::vector<int> inputDims, outputDims;
for (int i = 0; i < input->buffer().dimensions; i++) {
inputDims.push_back(input->buffer().dim[i].extent);
}
for (int i = 0; i < output->buffer().dimensions; i++) {
outputDims.push_back(output->buffer().dim[i].extent);
}
Optimized::Logistic(input->host<uint8_t>(), inputDims, mInputZeroPoint,
2019-04-17 10:49:11 +08:00
mInputRangeRadius, mInputMultiplier, mInputLeftShift, output->host<uint8_t>(), outputDims);
return NO_ERROR;
}
class CPUQuantizedLogisticCreator : public CPUBackend::Creator {
public:
virtual Execution *onCreate(const std::vector<Tensor *> &inputs, const std::vector<Tensor *> &outputs,
const MNN::Op *op, Backend *backend) const {
return new CPUQuantizedLogistic(backend, op);
}
};
} // namespace MNN
- build: - unify schema building in core and converter; - add more build script for android; - add linux build script for python; - ops impl: - add floor mod support in binary; - use eltwise impl in add/max/sub/mul binary for optimization; - remove fake double support in cast; - fix 5d support for concat; - add adjX and adjY support for batch matmul; - optimize conv2d back prop filter; - add pad mode support for conv3d; - fix bug in conv2d & conv depthwise with very small feature map; - optimize binary without broacast; - add data types support for gather; - add gather ND support; - use uint8 data type in gather v2; - add transpose support for matmul; - add matrix band part; - add dim != 4 support for padding, reshape & tensor convert; - add pad type support for pool3d; - make ops based on TensorFlow Lite quantization optional; - add all & any support for reduction; - use type in parameter as output type in reduction; - add int support for unary; - add variable weight support for conv2d; - fix conv2d depthwise weights initialization; - fix type support for transpose; - fix grad outputs count for reduce grad and reshape grad; - fix priorbox & detection output; - fix metal softmax error; - python: - add runSessionWithCallBackInfo interface; - add max nodes limit (1400) for visualization tool; - fix save error in python3; - align default dim; - convert: - add extra design for optimization; - add more post converting optimizers; - add caffe v1 weights blob support; - add cast, unary, conv transpose support for onnx model; - optimize batchnorm, conv with variable weights, prelu, reshape, slice, upsample for onnx model; - add cos/sin/atan/tan support for unary for tensorflow model; - add any/all support for reduction for tensorflow model; - add elu, conv3d, pool3d support for tensorflow model; - optimize argmax, batchnorm, concat, batch to space, conv with variable weights, prelu, slice for tensorflow model; - others: - fix size computer lock; - fix thread pool deadlock; - add express & parameters in express; - rewrite blitter chooser without static map; - add tests for expr;
2019-10-29 13:37:26 +08:00
#endif
namespace MNN {
REGISTER_CPU_OP_CREATOR_OLD(CPUQuantizedLogisticCreator, OpType_QuantizedLogistic);
};