YOLO11s — PPE Compliance Detection (ONNX)
A YOLO11s model fine-tuned to detect Personal Protective Equipment (PPE) compliance on construction sites.
Model Details
| Architecture | YOLO11s (Ultralytics) |
| Format | ONNX (CPU-optimised) |
| Input size | 640x640 |
| Training resolution | 1280x1280 |
| Epochs | 100 |
| Hardware | NVIDIA RTX 4060 |
Performance
| Metric | Value |
|---|---|
| Test mAP50 | 93.2% |
| Test mAP50-95 | 63.9% |
| Minority class AP50 | >91% (vest, no-vest) |
Classes
| ID | Class | Type |
|---|---|---|
| 0 | hardhat | Compliant |
| 1 | no-hardhat | Violation |
| 2 | vest | Compliant |
| 3 | no-vest | Violation |
| 4 | person | Neutral |
Training Data
Merged from three public datasets (~10K images):
- Construction Site Safety (Roboflow)
- SHWD - Safety Helmet Wearing Dataset (GitHub)
- Pictor-PPE (GitHub)
Usage
from ultralytics import YOLO
model = YOLO("best.onnx")
results = model("image.jpg", imgsz=640, conf=0.25)
results[0].plot()
Live Demo
Try it on HuggingFace Spaces.