Metadata-Version: 2.4
Name: yolodetection
Version: 1.0.4
Summary: Yolo + sahi toolkit
Author-email: Maxime Phaneuf <maxime.phaneuf@chaac.tech>
License: MIT
Requires-Python: >=3.8
Description-Content-Type: text/markdown
Requires-Dist: pathlib
Requires-Dist: ultralytics
Requires-Dist: sahi

# yolodetection

`yolodetection` is a lightweight wrapper around Ultralytics YOLO + SAHI sliced inference.  
It’s designed so that **you only ever interact with one class: `YOLOInstance`**.

Main goals:
- Simple interface
- Handles large images using SAHI slicing
- Exports clean, JSON-ready detections (with centroids)
- Optional saving of prediction JSON + visualization masks to disk

---

## 🔧 Installation

```bash
pip install yolodetection
```

## Usage
```
from yolodetect import YoloInstance

instance = YoloInstance("models/best.pt")
instance.predict(/home/images/)
```
