Downloads · 30 days
2.3K
28% of all-time downloads
fire-smoke-yolov8n
rabahdev/fire-smoke-yolov8n
fire-smoke-yolov8n is a object detection model from rabahdev. Use it when you need objects located in an image. It is set up for ultralytics. The card lists the license as agpl-3.0.
YOLOv8n fine-tuned on the D-Fire dataset for detecting smoke and fire in images.
All-time downloads
8.3K
Public
Repo size
12.8 MB
Likes
6
Public
Repo makeup
Click a slice to open those files.
.pt12.5 MB · 97%
At a glance
- Task
- Object Detection
- Library
- ultralytics
- License
- agpl-3.0
- Access
- Public
- Created
- Jun 8, 2026
- Updated
- Jun 8, 2026
- SHA
- 13017fe8
Model card
From the Hugging Face model README
YOLOv8n — Fire & Smoke Detector (D-Fire fine-tune)
YOLOv8n fine-tuned on the D-Fire dataset for detecting smoke and fire in images.
Classes
0— smoke1— fire
Files
best.pt— fine-tuned weights (use this)last.pt— final-epoch weights
Results (test split, 4,306 images)
| Metric | All | Smoke | Fire |
|---|---|---|---|
| mAP50 | 0.754 | — | — |
| mAP50-95 | 0.430 | 0.499 | 0.362 |
| Precision | 0.766 | — | — |
| Recall | 0.688 | — | — |
Training config
- Base:
yolov8n.pt(COCO pretrained) - Epochs: 50, image size: 640, batch: 16
- Optimizer: MuSGD (auto), lr0=0.01
- Device: Apple MPS
Usage
from huggingface_hub import hf_hub_download
from ultralytics import YOLO
ckpt = hf_hub_download(repo_id="rabahdev/fire-smoke-yolov8n", filename="best.pt")
model = YOLO(ckpt)
results = model("image.jpg")
results[0].show()
Dataset
- D-Fire (smoke + fire detection), YOLO format
- Train: 14,122 — Val: 3,099 — Test: 4,306
License
AGPL-3.0 (inherited from Ultralytics YOLOv8).
- Task
- Object Detection
- Library
- ultralytics
- License
- agpl-3.0
- Created
- Jun 8, 2026
- Updated
- Jun 8, 2026