Quantized YOLOX cannot be use

[nockn]

Hi.

I downloaded a YOLOX from the Model Zoo.(for PyTorch, Vitis AI 3.0)
I trained it on a custom dataset, and performed quantization and compilation.

However, running it on the board causes PetaLinux to crash after a few predictions.

Do you know what caused this?

Thank you.

#include <stdio.h>

#include <cstdio>
#include <cstdlib>
#include <fstream>
#include <iomanip>
#include <iostream>
#include <opencv2/opencv.hpp>
#include <sstream>
#include <string>
#include <typeinfo>
#include <vitis/ai/yolovx.hpp>

using namespace std;
using namespace cv;

#define YOLO_MODEL "YOLOX_nano_pt"
#define CAMERA_ID    1


int main()
{
	// init YOLO
	std::unique_ptr<vitis::ai::YOLOvX, std::default_delete<vitis::ai::YOLOvX>>
		yolo = vitis::ai::YOLOvX::create(YOLO_MODEL, true);

	cv::VideoCapture cap(CAMERA_ID, cv::CAP_V4L2);

	while (1) {
		cv::Mat frame;
		cap >> frame;

		vitis::ai::YOLOvXResult results = yolo->run(frame);
		
		for (auto& result : results.bboxes) {
		 	int label = result.label;
		 	auto& box = result.box;

		 	cout << "RESULT: " << label << "\t" << std::fixed
		 		 << std::setprecision(2) << box[0] << "\t" << box[1] << "\t"
		 		 << box[2] << "\t" << box[3] << "\t" << std::setprecision(6)
		 		 << result.score << "\n";
		}
	}

	return 0;
}