detect_face_plate_sign

This model has be trained for the Panoramax project in order to detect:

• people face to blur them
• licence plates to blur them
• road signs to classify them with other models

The last model has been trained on yolo11l with imgsz of 2048 and 300 epochs, the older one on yolo8s.

Here is the last run validation :

Validating runs/detect/train5/weights/best.pt...
Ultralytics 8.3.29 🚀 Python-3.12.3 torch-2.5.1+cu124 CUDA:0 (NVIDIA GeForce RTX 4090, 24207MiB)
YOLO11l summary (fused): 464 layers, 25,281,625 parameters, 0 gradients, 86.6 GFLOPs
                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 83/83 [00:08<00:00,  9.78it/s]
                   all        329       1209      0.812      0.768      0.815      0.412
                  sign        231        507      0.879      0.836      0.898      0.561
                 plate        202        410      0.833      0.849      0.889      0.438
                  face        118        292      0.724      0.619      0.657      0.237
Speed: 1.2ms preprocess, 18.4ms inference, 0.0ms loss, 1.6ms postprocess per image