Views
No views yet
Fire_best.pt1from ultralytics import YOLO
2import torch
3torch.backends.cudnn.benchmark = True
4
5model = YOLO("path/to/yolo11n.pt")
6train_results = model.train(
7 data="path/to/FASDD_CV_Fire/data.yaml",
8 epochs=100,
9 imgsz=640,
10 batch=16,
11 optimizer="AdamW",
12 lr0=1e-4,
13 lrf=0.01,
14 dropout=0.15,
15 weight_decay=5e-4,
16 device=0,
17 val=False,
18 save=True,
19 plots=False
20)| P | R | mAP@0.5 | mAP@0.5:0.95 |
|---|---|---|---|
| 0.799 | 0.697 | 0.797 | 0.520 |
Fire_best.pt): AGPL-3.0 (Ultralytics-trained).1@inproceedings{profsam2025,
2 author = {Emmanuel U. Ugwu and Xinming Zhang},
3 title = {Promptable Fire Segmentation: Unleashing SAM2's Potential for Real-Time Mobile Deployment with Strategic Bounding Box Guidance},
4 booktitle = {ICIGP '26},
5 year = {2026},
6 address = {Wuhan, China},
7 month = jan,
8 note = {to appear}
9}1@software{profsam2025,
2 author = {Ugwu, Emmanuel U. and Zhang, Xinming},
3 title = {Promptable Fire Segmentation: Unleashing SAM2’s Potential for Real-Time Mobile Deployment with Strategic Bounding Box Guidance},
4 year = {2025},
5 doi = {10.5281/zenodo.17340313},
6 url = {https://doi.org/10.5281/zenodo.17340313}
7}