//
//  ReLU6GradTest.cpp
//  MNNTests
//
//  Created by MNN on 2022/07/12.
//  Copyright © 2018, Alibaba Group Holding Limited
//

#include <MNN/expr/Expr.hpp>
#include <MNN/expr/ExprCreator.hpp>
#include "MNNTestSuite.h"
#include "TestUtils.h"
#include "../tools/train/source/grad/OpGrad.hpp"

using namespace MNN;
using namespace MNN::Express;

class ReLU6GradTest : public MNNTestCase {
public:
    char name[20] = "ReLU6";
    virtual ~ReLU6GradTest() = default;

    virtual bool run(int precision) {
        {
            const int len = 4;
            auto input = _Input({len}, NCHW);
            const float inpudata[] = {-1.0, -2.0, 3.0, 6.0};
            auto inputPtr          = input->writeMap<float>();
            memcpy(inputPtr, inpudata, len * sizeof(float));

            auto output = _Relu6(input);
            auto opExpr = output->expr().first;

            auto grad = OpGrad::get(opExpr->get()->type());
            float outputDiff[] = {0.1, -0.2, -0.3, 0.4};
            auto inputGrad = grad->onGrad(opExpr, {_Const(outputDiff, {len})});

            const std::vector<float> expectedOutput = {0.0f, 0.0f, -0.3f, 0.0f};
            auto gotOutput = inputGrad[0]->readMap<float>();

            for (int i = 0; i < len; ++i) {
                auto diff = ::fabsf(gotOutput[i] - expectedOutput[i]);
                if (diff > 0.000001) {
                    MNN_ERROR("%s grad test failed, expected: %f, but got: %f!\n", name, expectedOutput[i], gotOutput[i]);
                    return false;
                }
            }
        }
        {
            float minValue = -3.0f;
            float maxValue = 1.0f;
            const int len = 4;
            auto input = _Input({len}, NCHW);
            const float inpudata[] = {-1.0f, -2.0f, 3.0f, 6.0f};
            auto inputPtr          = input->writeMap<float>();
            memcpy(inputPtr, inpudata, len * sizeof(float));

            auto output = _Relu6(input, minValue, maxValue);
            auto opExpr = output->expr().first;

            auto grad = OpGrad::get(opExpr->get()->type());
            float outputDiff[] = {0.1f, -0.2f, -0.3f, 0.4f};
            auto inputGrad = grad->onGrad(opExpr, {_Const(outputDiff, {len})});

            const std::vector<float> expectedOutput = {0.1f, -0.2f, 0.0f, 0.0f};
            auto gotOutput = inputGrad[0]->readMap<float>();

            for (int i = 0; i < len; ++i) {
                auto diff = ::fabsf(gotOutput[i] - expectedOutput[i]);
                if (diff > 0.000001) {
                    MNN_ERROR("%f-%f, %s grad test failed, expected: %f, but got: %f!\n", minValue, maxValue, name, expectedOutput[i], gotOutput[i]);
                    return false;
                }
            }
        }
        return true;
    }
};

MNNTestSuiteRegister(ReLU6GradTest, "grad/relu6");