Metrics of the Model that Detects License Plates of Vehicles in Turkey
I trained a license plate detection model on the Roboflow platform for 100 epochs with the YOLOv11l model on a dataset of 3000 photos, which I divided into 2695 training, 300 validation and 5 testing. The training validation metrics of the model are given in the graph below.
Metrics of the Model that Detects Traffic Signs Recognition
In the same way, again using the Roboflow platform, I divided the dataset into 2400 training, 450 validation, and 150 testing on 3000 photos, which are frequently encountered in urban areas: traffic signs. I trained 24 different traffic signals with the YOLOv11l model for 100 epochs, and the training and validation metrics are given in the graph below.
