Object Detection
ultralytics
YOLOv26
English
Russian
yolo
yolov8
yolov11
gaming
computer-vision
real-time
Instructions to use DragonCheat-AI/Dragon-4-Lite with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- ultralytics
How to use DragonCheat-AI/Dragon-4-Lite with ultralytics:
from ultralytics import YOLOvv8 model = YOLOvv8.from_pretrained("DragonCheat-AI/Dragon-4-Lite") source = 'http://images.cocodataset.org/val2017/000000039769.jpg' model.predict(source=source, save=True) - YOLOv26
How to use DragonCheat-AI/Dragon-4-Lite with YOLOv26:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
Dragon 4 Lite
Dragon 4 Lite β lightweight real-time object detection model for Standoff 2 (and similar games with Standoff 2 characters).
π― Model Details
| Property | Value |
|---|---|
| Architecture | YOLOv26n |
| Task | Object Detection |
| Classes | ct (Counter-Terrorist), tr (Terrorist) |
| Input Size | 520Γ520 |
| File Size | ~5-6 MB |
| Framework | Ultralytics |
| License | MIT |
π Performance (DragonTest-V2, 175 images)
| Test | Score |
|---|---|
| Lite-30 | 87.43% |
| Lite-40 | 88.00% |
| Lite-50 | 86.29% |
| Lite-70 | 82.29% |
| Hard-40 | 84.57% |
| Hard-50 | 82.86% |
| Hard-60 | 82.29% |
| UltraHard | 2.29% |
Average (excluding UltraHard): 84.82%
β‘ Key Features
- Real-time inference β works on mid-range GPUs (GTX 1650+, RTX 3050+)
- Low VRAM usage β ~2 GB
- Robust β does not confuse
ctandtr(a common problem in older models) - Generalizes well β works on new maps/skins not seen during training
- Optimized for Standoff 2 β trained on real game screenshots
- Downloads last month
- 7
Model tree for DragonCheat-AI/Dragon-4-Lite
Unable to build the model tree, the base model loops to the model itself. Learn more.