Train a YOLOv8 model on license plate images (class 0 = plate bounding box).
After training, copy the .pt file here:
models/plate_detector.pt
Then set PLATE_MODEL_PATH=models/plate_detector.pt in your .env file.
The system will automatically use your custom model for plate detection. If the
file is not found, it falls back to the built-in OpenCV Haar cascade.
Main… See the full description on the dataset page: https://huggingface.co/datasets/Asilbek14/Crosswalk_Violation.
