How to Run Multiple Streams with DeepStream SDK on Jetson Nano using Ultralytics YOLOv8 | Episode 82

Поделиться
HTML-код
  • Опубликовано: 8 ноя 2024

Комментарии • 17

  • @muhammadrizwanmunawar
    @muhammadrizwanmunawar Месяц назад

    I like this video, I did spent most of my time on deepstream work in last years. Great to see its usage with YOLOv8 😊

    • @Ultralytics
      @Ultralytics  Месяц назад

      Glad you enjoyed the video! DeepStream and YOLOv8 make a powerful combo for real-time applications. If you're interested in more details, check out our guide here: docs.ultralytics.com/guides/deepstream-nvidia-jetson/ 🚀

  • @ArjunS-k1f
    @ArjunS-k1f Месяц назад

    I have a Jetson Xavier AGX. Will it be challenging to run DeepStream 7, since it's optimized for the Jetson Orin series?

    • @Ultralytics
      @Ultralytics  Месяц назад

      Running DeepStream 7 on a Jetson Xavier AGX should be feasible, but you might encounter some performance differences compared to the Jetson Orin series. Ensure your JetPack version is compatible and follow the setup guide here: docs.ultralytics.com/guides/deepstream-nvidia-jetson/. If you face any issues, consider checking for updates or optimizations specific to your hardware. 😊

  • @TheodoreBC
    @TheodoreBC Месяц назад

    Bro, if the Jetson Nano bucks like a wild bronco under multiple streams, any tips for roping it in without overheating? Could rugged outback conditions impact performance in any mysterious ways?

    • @Ultralytics
      @Ultralytics  Месяц назад

      Hey there! To keep your Jetson Nano cool while handling multiple streams, make sure to enable MAX Power Mode and Jetson Clocks for optimal performance. Also, consider using a heatsink or fan to manage heat. Rugged conditions might affect performance, so ensure proper ventilation and avoid direct sunlight. For more tips, check out our guide: docs.ultralytics.com/guides/nvidia-jetson/ 🌟

  • @m033372
    @m033372 2 месяца назад

    Awesome video! Curious how the Jetson Nano manages real-time performance when running multiple YOLOv8 streams. Are there any significant limitations we should be aware of in practical applications, or can we expect smooth operations out of the box? Also, ever tried pitting this setup against a jet engine for the ultimate stream showdown?

    • @Ultralytics
      @Ultralytics  2 месяца назад

      Thanks for watching! 😊 The Jetson Nano can handle multiple YOLOv8 streams, but performance depends on factors like model size and input resolution. You might experience some limitations in processing speed, especially with higher resolutions or more streams. For optimal performance, consider using FP16 or INT8 precision with TensorRT. For more details, check out our guide: docs.ultralytics.com/guides/deepstream-nvidia-jetson/
      And as for the jet engine showdown, that sounds like an epic experiment! 🚀

  • @krisnayoga-g9r
    @krisnayoga-g9r 2 месяца назад

    please discuss how to insert attention mechanisms into the yolov8 layer

    • @Ultralytics
      @Ultralytics  2 месяца назад

      Integrating attention mechanisms into YOLOv8 involves modifying the model architecture to include attention layers, which can enhance feature extraction. For detailed guidance, you might want to explore the Ultralytics documentation or community discussions for insights on custom model modifications. If you have specific implementation questions, feel free to ask! 😊
      Check out the docs here: docs.ultralytics.com/models/yolov8/

  • @Sashreek-dz7jb
    @Sashreek-dz7jb 18 дней назад

    I tried the steps in this video to build the TensorRT engine for my DeepStream application, but after the build is complete I'm encountering a segmentation fault during engine creation. The exact error is: 'Thread 1 "deepstream-app" received signal SIGSEGV, Segmentation fault.' Does anyone know how to resolve this? I've tried rebuilding the engine and checking for memory issues, but it's still crashing. Any help would be appreciated!

    • @Ultralytics
      @Ultralytics  18 дней назад

      Segmentation faults can be tricky! Ensure you're using the latest versions of DeepStream and TensorRT. Double-check your model conversion steps, especially when exporting to ONNX. Sometimes, mismatched input dimensions or unsupported layers can cause issues. You might also want to verify your Jetson Nano's memory usage during the process. For more detailed guidance, check out our DeepStream guide docs.ultralytics.com/guides/deepstream-nvidia-jetson/. Good luck! 🚀

  • @asdfds6752
    @asdfds6752 Месяц назад

    Trying to run this on a ubuntu laptop but get several errors when running the app. Is there anything that should be different on a laptop or other ubuntu machines?

    • @asdfds6752
      @asdfds6752 Месяц назад

      My bad, it was an error in the "deepstream_app_config.txt". Was pointing at the wrong yoloV8.txt file. everything works fine

    • @Ultralytics
      @Ultralytics  Месяц назад

      Glad you figured it out! If you need more help, feel free to ask. For common issues, check out our guide: docs.ultralytics.com/guides/yolo-common-issues/ 😊

  • @oilbender
    @oilbender 2 месяца назад

    No need for openCV cuda?

    • @Ultralytics
      @Ultralytics  2 месяца назад +1

      For running YOLOv8 on Jetson Nano with DeepStream, OpenCV CUDA isn't necessary. DeepStream handles GPU acceleration efficiently. If you have more questions, feel free to ask! 😊
      For more details, check out our guide: docs.ultralytics.com/guides/deepstream-nvidia-jetson/