---
license: mit
library_name: ultralytics
pipeline_tag: object-detection
tags:
  - birdlense
  - yolov11
  - object-detection
  - birds
  - rodents
  - feeder-camera
---

# BirdLense Detector (3-class)

YOLO detector weights for BirdLense Hub (CV/ML roadmap #368).

## Classes (`model.names`)

0. Bird  
1. Rodent  
2. Background

## Intended use

- Binary detector stage in BirdLense two-stage pipeline
- Feeder-camera wildlife monitoring
- Local/on-device inference

## Integration in BirdLense

Default detector path in repository:

`app/processor/models/detection/weights/best.pt`

Relevant config keys:

- `processor.models.binary`
- `processor.detector_weight_contract` (`off | warn | enforce`)
- optional OpenVINO path:
  - `processor.models.binary_openvino`
  - `BIRDLENSE_BINARY_OPENVINO_PATH`

## Quick start (Ultralytics)

```python
from ultralytics import YOLO

model = YOLO("best.pt")
print(model.names)  # expected: {0:'Bird', 1:'Rodent', 2:'Background'}
results = model.predict("sample.jpg", conf=0.25)