iam-tsr/yolov8n-helmet-detection

Helmet Detection with YOLOv8n

A lightweight, real-time object detector for “helmet” vs “no_helmet” built with Ultralytics YOLOv8n. The project includes data download Roboflow, fine-tuning, and evaluation.

Usage

from ultralytics import YOLO
from huggingface_hub import hf_hub_download

repo_id = "iam-tsr/yolov8n-helmet-detection"
filename = "best.pt"

# Load downloaded model
model = YOLO(filename)

# Run inference
result_many = model.predict(
    source="images/image.png",
    save=False,
    show=True,
)

Evaluation

Metric	Value
Precision (P)	0.855
Recall (R)	0.808
mAP@0.5	0.881
mAP@0.5:0.95	0.538