{

if (argc != 3) {

std::cerr << "Usage: [apps] [path/to/onnx/model] [path/to/image]" << std::endl;

return EXIT_FAILURE;

}

const std::string ONNX_MODEL_PATH = argv[1];

const std::string IMAGE_PATH = argv[2];

Ort::ImageClassificationOrtSessionHandler osh(Ort::IMAGENET_NUM_CLASSES, ONNX_MODEL_PATH, 0);

osh.initClassNames(Ort::IMAGENET_CLASSES);

cv::Mat img = cv::imread(IMAGE_PATH);

if (img.empty()) {

std::cerr << "Failed to read input image" << std::endl;

return EXIT_FAILURE;

}

cv::resize(img, img, cv::Size(IMG_WIDTH, IMG_HEIGHT));

float* dst = new float[IMG_WIDTH * IMG_HEIGHT * IMG_CHANNEL];

osh.preprocess(dst, img.data, IMG_WIDTH, IMG_HEIGHT, IMG_CHANNEL, Ort::IMAGENET_MEAN, Ort::IMAGENET_STD);

std::chrono::high_resolution_clock::time_point begin = std::chrono::high_resolution_clock::now();

for (int i = 0; i < TEST_TIMES; ++i) {

auto inferenceOutput = osh({reinterpret_cast<float*>(dst)});

const int TOP_K = 5;

// osh.topK({inferenceOutput[0].first}, TOP_K);

std::cout << osh.topKToString({inferenceOutput[0].first}, TOP_K) << std::endl;

}

std::chrono::high_resolution_clock::time_point end = std::chrono::high_resolution_clock::now();

auto elapsedTime = std::chrono::duration_cast<std::chrono::milliseconds>(end - begin);

std::cout << elapsedTime.count() / 1000. << "[sec]" << std::endl;

delete[] dst;

return 0;