metadata
license: mit
language:
- en
metrics:
- "[email protected]\t0.697"
- "[email protected]:0.95\t0.438"
- "precision \t0.764"
- "recall\t0.646"
base_model:
- Ultralytics/YOLOv8
tags:
- Pytorch
- detection
- object-detection
- yolo
- ultralytics
- yolov8x
pipeline_tag: object-detection
📌 YOLOv8x Accident Evaluator
YOLOv8x-Accident-Evaluator is a custom object detection model trained to detect road accident scenarios and severity levels using the Roboflow Accident Evaluator dataset.
🚀 Model Details
- Architecture: YOLOv8x (You Only Look Once, version 8, extra-large)
- Trained by: Uppada Enos & Ultralytics Team
- Contact: Uppada Enos
- License: MIT
🏷️ Detection Classes
- detected-injury
- fire
- high severity
- low severity
- medium severity
- smoke
🧪 Performance Metrics
| Metric | Score |
|---|---|
| [email protected] | 0.697 |
| [email protected]:0.95 | 0.438 |
| Precision | 0.764 |
| Recall | 0.646 |
Per-class [email protected]:
- detected-injury: 0.741
- fire: 0.611
- high: 0.869
- low: 0.834
- medium: 0.795
- smoke: 0.331
📂 Usage
from ultralytics import YOLO
model = YOLO("Enos-123/accident-evaluator-yolov8x")
results = model("example.jpg", imgsz=640, conf=0.25)
results.show()
results.save(save_dir="inference_output")