/DeepStream-Yolov8-Jetson-Nano

Learn to run Yolov8 on Jetson Nano using DeepStream

Primary LanguageC++

Learn to run YOLOv8 on Jetson Nano Using DeepStream

To run deepstream applications, Make sure DeepStream SDK is properly installed.

Check Versions:

deepstream-app --version
Check this tutorial to learn how to install deepstream : https://www.youtube.com/watch?v=Ufd86duobvc
Check tensorrt version. tensorrt –version
Check cuda version. nvcc --veersion

Clone this repo: git clone https://github.com/AarohiSingla/DeepStream-Yolov8-Jetson-Nano.git

Paste yolov8s.pt, yolov8s.onnx, yolov4.weights, yolov4.cfg in this repo.

compile the lib:

export CUDA_VER=10.2

For jetson nano (DeepStream 6.0.1 / 6.0 / 5.1 = 10.2)

Make the lib:

make -C nvdsinfer_custom_impl_Yolo clean && make -C nvdsinfer_custom_impl_Yolo

(MAke the changes in MakeFile before executing it)

Edit the config_infer_primary_yoloV8.txt file according to your model (example for YOLOv8s with 80 classes)

[property] onnx-file=yolov8s.onnx .. num-detected-classes=80

Edit the deepstream_app_config file:

... [primary-gie] ... config-file=config_infer_primary_yoloV8.txt

Testing the model:

deepstream-app -c deepstream_app_config.txt