metadata
license: apache-2.0
pipeline_tag: object-detection
library_name: ultralytics
model_type: yolov9
datasets:
- benediktkol/DDOS
metrics:
- accuracy
YOLO298B is a custom‑trained Ultralytics YOLO model (best.pt) built by Team 6 (SJSU).
It detects aerial classes in aerial imagery collected by autonomous drones.
| Attribute | Value |
|---|---|
| Architecture | YOLO‑v9‑S |
| Input size | 640 × 640 px |
| Classes | n |
| Checkpoint | 5.5 MB |
Intended uses & limitations
| Use‑case | ✅ Recommended | 🚫 Not recommended |
|---|---|---|
| Real‑time obstacle detection on UAVs | ✔️ | |
| Academic research / benchmarking | ✔️ | |
| Safety‑critical deployment w/o human | ❌ |
Training data
Dataset: benediktkol/DDOS – contains drone‑view images with obstacles (<brief description>).
Split: 80 % train · 10 % val · 10 % test
Quick start
from ultralytics import YOLO
model = YOLO("vsham001/Yolo298B")
results = model("https://ultralytics.com/images/bus.jpg")
results[0].show()