Class	Precision	Recall	mAP50	mAP50-95
All	0.923	0.745	0.852	0.585
Coverall	0.944	0.711	0.840	0.623
Face_Shield	1.000	0.989	0.995	0.712
Gloves	0.826	0.672	0.798	0.497
Goggles	0.866	0.531	0.750	0.445
Mask	0.977	0.819	0.878	0.648

Trained for 46 epochs (early stopped, best at epoch 31), imgsz=512, batch=32, on a Tesla T4.

Usage

python

from ultralytics import YOLO

model = YOLO("best.pt")
results = model.predict("image.jpg")

Training details

Base model: yolo26s.pt
Augmentations: HSV, flip, mosaic + custom Albumentations pipeline (Blur, MedianBlur, ToGray, CLAHE, RandomBrightnessContrast, RandomGamma)
Framework: Ultralytics 8.4.87