Running Custom YoloV5 TensorRT engine on Jetson Nano | Rocket Systems

Running Custom YoloV5 TensorRT engine on Jetson Nano
Custom YoloV5 TensorRT engine on Jetson Nano
YoloV5 TensorRT engine on Jetson Nano
In this video we will see how you can convert and build custom yolov5 model into TensorRT engine on Jetson Nano.
Repo link: github.com/mailrocketsystems/...
Jetson Yolov5 tensort: • Jetson Yolo TensorRT
Jetson computer vision full course: • Jetson Computer Vision...
Watch full Jetson playlist: • Jetson Computer Vision
Nvidia Jetson Inference: • Nvidia Jetson Inferencing
Contact: [e-mail]
Running Custom YoloV5 TensorRT engine on Jetson Nano,Custom YoloV5 TensorRT engine on Jetson Nano,YoloV5 TensorRT engine on Jetson Nano,yolov5 custom dataset,tensorrt jetson nano,running inference with tensorrt,linux en jetson nano,custom object detection transfer learning,como configurar jetson nano,yolov5,qué es jetson nano,object detection jetson nano,yolov4 custom object detection,nvidia jetson nano developer kit,tensorrt yolov5,custom car detection

[@joaquinarino3390]

you are the best, it was really helpful!!!🚀🚀🚀

[@manuelvelasco7581]

fue realmente util tu video.....

[@mapriyaditriputra992]

Good videos sir, its working for me, but can i ask something, how to block all print in terminal when detecting?

[@adnanahmad4168]

And one more thing can you please upload a video on deploying custom Yolov8 model on Jetson nano. Please!

[@enollamisaki1254]

hello there, which version of yolov5 are you use?

[@legacyofmocarony2944]

Im trying to do with my custom model, it says "no module named ultralytics" when im about to generate the wts file. Is it safe if im installing the ultralytics?

[@trungquyenha9951]

hey sir, i am working with yolov5n, the weight is only 1/3 compare to the s, how can i config it. I tried to convert into engine and got the weights error.

[@adnanahmad4168]

You have tested it for video. What if I need to use Camera and do detection using camera?

[@RocketSystems]

You can use camera. The code uses opencv, so you can replace the video filename with your camera source

[@user-rx8jq2jk5e]

Every time I make ，the following error, my opencv is 4.5.2 and it has been installed as a function import ，could you heip me
cu:2:10: fatal error: opencv2/opencv.hpp: No such file or directory
#include
^~~~~~~~~~~~~~~~~~~~
compilation terminated.

[@haihuoctv4651]

Hello @Rocket Systems , Thank you for this tutorial , this help me a lot
i have 1 question that : When i try to generate .engine file (Serialize the model):
./yolov5 det -s best.wts best.engine s
facing with this error :
1. [E] [TRT] Parameter check failed at: ../builder/Network.cpp::addScale::737, condition: shift.count > 0 ? (shift.values != nullptr) : (shift.values == nullptr)
2. yolov5_det : /home/....../model.cpp :95: nvinfer1::IScaleLayer* addBatchNorm2d(nvinfer1::INetworkDefinition*, std::map&, nvinfer1::ITensor&, std::__cxx11::string, float): Assertion `scale_1' failed.
it have been trained with custom data set , have configured the config.h file about the number of class
In my training , i have changed the network architecture of yolov5s (in yolov5s.yaml)
Have you face with this issue ?

[@RocketSystems]

Files provided in yolov5 directory are for standard network architecture, if you have updated it then you will have to update the files accordingly.

[@RocketSystems]

Hi, I have not done testing with custom architecture so cannot say much on that. Will try sometime later

[@haihuoctv4651]

@@RocketSystems OK , got it
Thanks very much for your tutorials!

[@haihuoctv4651]

@@RocketSystems i have 1 more question that which version of Yolov5 that you combine to your workspace (JetsonYolov5)? i mean version 7 or 6.2 or...?

[@thomasemmanuelgabrielmacat2962]

Sir, have you resolved the problem? I'm facing the same issue...

[@namle9770]

# Build engine
./yolov5_det -s best52.wts best52.engine s
I have a question: if my model (best52) size is "m". So please help me edit above script. Many thanks!