Welcome to Raspberry Pi 5 repository

Model training_conversion

After completing the tutorial content of dataset annotation, we can use the motherboard to start training the model.

xxxxxxxxxx
This tutorial only introduces the model training and conversion of CLI. You can refer to the official website to modify the Python case
Due to performance reasons, the model training time of the Raspberry Pi 5 motherboard is relatively long, so you need to wait patiently.

1. Model training

Use CLI commands to train the model directly: Copy the yolo11n.pt file to the directory where the configuration file is located, and then open the terminal in the directory where the configuration file is located:

xxxxxxxxxx
cd /home/pi/ultralytics/ultralytics/data/yahboom_data/orange_data

xxxxxxxxxx
yolo detect train data=orange.yaml model=yolo11n.pt epochs=100 imgsz=640

data: Dataset configuration file

model: Pre-trained model file

epochs: Number of training rounds

imgsz: Input specified image size

2. Model conversion

The final model will be generated in the runs folder: generally choose the best.pt file for use

xxxxxxxxxx
/home/pi/ultralytics/ultralytics/data/yahboom_data/orange_data/runs/detect/train/weights

Convert the pt model to an onnx or ncnn model:

xxxxxxxxxx
cd /home/pi/ultralytics/ultralytics/data/yahboom_data/orange_data/runs/detect/train/weights

xxxxxxxxxx
yolo export model=best.pt format=onnx
yolo export model=best.pt format=ncnn

References

https://docs.ultralytics.com/modes/train/