wehub-resource-sync
122 lines
3.8 KiB
C++
122 lines
3.8 KiB
C++
//
|
|
// WhereTest.cpp
|
|
// MNNTests
|
|
//
|
|
// Created by MNN on 2021/11/11.
|
|
// Copyright © 2018, Alibaba Group Holding Limited
|
|
//
|
|
|
|
#include <MNN/expr/Expr.hpp>
|
|
#include <MNN/expr/ExprCreator.hpp>
|
|
#include "MNNTestSuite.h"
|
|
#include "TestUtils.h"
|
|
|
|
using namespace MNN::Express;
|
|
static void print(VARP xc, std::string name) {
|
|
MNN_PRINT("%s begin:\n", name.c_str());
|
|
xc->readMap<void>();
|
|
xc->getTensor()->print();
|
|
MNN_PRINT("%s end:\n", name.c_str());
|
|
}
|
|
static bool _testFromIssue() {
|
|
{
|
|
// 初始化
|
|
VARP var = MNN::Express::_Input({10}, Dimensionformat::NCHW, halide_type_of<float>());
|
|
float *pData = var->writeMap<float>();
|
|
std::vector<float> data = {
|
|
0.5f, 0.5f, 1.0f, 1.0f, 1.5f, 1.5f, 1.0f, 1.0f, 0.2f, 0.3f,
|
|
};
|
|
memcpy(pData, data.data(), var->getInfo()->size*sizeof(float));
|
|
|
|
auto xc = _Greater(var, _Scalar<float>(1.0));
|
|
print(xc, "xc");
|
|
|
|
auto w = _Where(xc);
|
|
print(w, "where");
|
|
if (w->getInfo()->size != 2) {
|
|
return false;
|
|
}
|
|
if (w->readMap<int>()[0] != 4 || w->readMap<int>()[1] != 5) {
|
|
return false;
|
|
}
|
|
}
|
|
|
|
{
|
|
VARP var = _Input({10}, Dimensionformat::NCHW, halide_type_of<float>());
|
|
auto pData = var->writeMap<float>();
|
|
std::vector<float> data = {
|
|
0.5f, 0.5f, 1.0f, 1.0f, 1.5f, 1.5f, 1.0f, 1.0f, 0.2f, 0.3f,
|
|
};
|
|
memcpy(pData, data.data(), var->getInfo()->size * sizeof(float));
|
|
auto index = _Sort(var, 0, true, true);
|
|
print(index, "Sort index");
|
|
auto value = _Sort(var, 0, false, true);
|
|
print(value, "Sort value");
|
|
}
|
|
return true;
|
|
}
|
|
|
|
class WhereTest : public MNNTestCase {
|
|
public:
|
|
virtual ~WhereTest() = default;
|
|
virtual bool run(int precision) {
|
|
auto res = _testFromIssue();
|
|
if (!res) {
|
|
FUNC_PRINT(1);
|
|
return false;
|
|
}
|
|
return commonCase() &&
|
|
zeroCase();
|
|
|
|
}
|
|
bool commonCase() {
|
|
auto input = _Input({2, 3}, NCHW);
|
|
input->setName("input_tensor");
|
|
// set input data
|
|
const float inpudata[] = { 1.0, 0.0, 2.0,
|
|
3.0, 0.0, 4.0 };
|
|
auto inputPtr = input->writeMap<float>();
|
|
memcpy(inputPtr, inpudata, 6 * sizeof(float));
|
|
input->unMap();
|
|
auto output = _Where(input);
|
|
const std::vector<int> expectedOutput = {0, 0, 0, 2, 1, 0, 1, 2};
|
|
const std::vector<int> expectedShape = {4, 2};
|
|
|
|
auto realShape = output->getInfo()->dim;
|
|
if (!checkVector<int>(realShape.data(), expectedShape.data(), 2, 0)) {
|
|
MNN_ERROR("WhereTest shape mismatch!\n");
|
|
return false;
|
|
}
|
|
|
|
auto gotOutput = output->readMap<int>();
|
|
if (!checkVector<int>(gotOutput, expectedOutput.data(), 8, 0)) {
|
|
MNN_ERROR("WhereTest test failed!\n");
|
|
return false;
|
|
}
|
|
return true;
|
|
}
|
|
|
|
bool zeroCase() {
|
|
|
|
auto input = _Input({2, 4}, NCHW);
|
|
input->setName("input_tensor");
|
|
// set input data
|
|
const float inpudata_zero[] = { 0.0, 0.0, 0.0, 0.0,
|
|
0.0, 0.0, 0.0, 0.0 };
|
|
auto inputPtr = input->writeMap<float>();
|
|
memcpy(inputPtr, inpudata_zero, 8 * sizeof(float));
|
|
input->unMap();
|
|
auto output = _Where(input);
|
|
const std::vector<int> expectedOutput = {};
|
|
const std::vector<int> expectedShape = {0, 2};
|
|
|
|
auto realShape = output->getInfo()->dim;
|
|
if (!checkVector<int>(realShape.data(), expectedShape.data(), 2, 0)) {
|
|
MNN_ERROR("WhereTest zero shape mismatch!\n");
|
|
return false;
|
|
}
|
|
return true;
|
|
}
|
|
};
|
|
MNNTestSuiteRegister(WhereTest, "op/where");
|