---
license: agpl-3.0
tags:
  - object-detection
  - yolo
  - yolo11
  - onnx
  - ultralytics
  - computer-vision
pipeline_tag: object-detection
---

# LunchSpot YOLO11n (ONNX)

This repository hosts an **unmodified ONNX export of Ultralytics' YOLO11n**
(nano) checkpoint, pretrained on COCO. It is **not a custom-trained or
fine-tuned model** — the weights and detection classes are exactly the
official Ultralytics release. This repo exists purely as a deployment
artifact for the [LunchSpot](https://github.com/<your-org>/lunchspot)
project, so the app can pull the model file at build/deploy time instead
of committing a ~10 MB binary to the application repository.

If you need the model for a different purpose, prefer the official
Ultralytics release (GitHub / PyPI `ultralytics` package) over this mirror.

## Credits & license

- Original model: [Ultralytics YOLO11](https://github.com/ultralytics/ultralytics)
- License: **AGPL-3.0**, inherited unchanged from Ultralytics' pretrained
  weights. Any use of this file is subject to that license (or an
  Ultralytics Enterprise license if you have one).
- No weights were modified. Only the export format changed (PyTorch → ONNX).

## What this model does

Standard YOLO11n object detector over the 80 COCO classes, exported to
ONNX for CPU inference via ONNX Runtime. LunchSpot uses it as-is and, at
the application layer, filters detections down to two COCO classes:

| COCO class | Used for                         |
|------------|-----------------------------------|
| `person`   | occupancy / people counting       |