This model detects fire, smoke, and related fire indicators in images and videos. Built on YOLOv26-S architecture and trained on 8,939 annotated images.

Classes

fire - Active flames
smoke - Smoke plumes
other - Related fire indicators

Performance

Metric	Score
mAP@50	94.9%
mAP@50-95	68.0%
Precision	89.6%
Recall	88.8%

Training Details

Parameter	Value
Base Model	YOLOv26-S
Epochs	100
Batch Size	16
Image Size	640x640
Optimizer	AdamW
Learning Rate	0.01

Usage

Installation

pip install ultralytics

Inference

python

from ultralytics import YOLO

# Load model
model = YOLO("path/to/best.pt")

# Run inference
results = model.predict("image.jpg", conf=0.25)

# Process results
for result in results:
    boxes = result.boxes
    for box in boxes:
        cls = int(box.cls[0])
        conf = float(box.conf[0])
        label = model.names[cls]
        print(f"Detected: {label} ({conf:.2f})")

Video Inference

python

from ultralytics import YOLO

model = YOLO("path/to/best.pt")
results = model.predict("video.mp4", save=True, conf=0.25)

Webcam (Real-time)

python

from ultralytics import YOLO

model = YOLO("path/to/best.pt")
results = model.predict(source=0, show=True, conf=0.25)

Detection Samples

Training Curves

Use Cases

Building fire safety monitoring
Wildfire early detection systems
Industrial safety surveillance
Smart home fire detection
Drone-based fire monitoring

Limitations

May have reduced accuracy in low-light conditions
Smoke detection can be affected by fog/steam
Best performance on images similar to training data

Dataset

Trained on fire detection dataset from Roboflow Universe:

8,939 images
License: CC BY 4.0

Citation

bibtex

@misc{yolov26-fire-detection,
  author = {Salah AL-Haismawi},
  title = {YOLOv26 Fire Detection},
  year = {2026},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/SalahALHaismawi/yolov26-fire-detection}}
}

License

MIT License

Acknowledgments

Ultralytics for YOLO implementation
Roboflow for the dataset