---
license: mit
tags:
  - mlx
  - yolo
  - yolo11
  - object-detection
  - apple-silicon
  - omniparser
library_name: mlx
---

# yolo11-mlx: YOLO11 on Apple MLX

Pre-converted weights for running YOLO11 detection models on Apple Silicon via MLX Metal GPU.

## OmniParser (UI Element Detection)

Microsoft's [OmniParser v2](https://huggingface.co/microsoft/OmniParser-v2.0) — YOLO11m fine-tuned on 67K web screenshots for detecting interactive UI elements (buttons, links, inputs, icons).

| File | Size | Description |
|------|------|-------------|
| omniparser_mlx.npz | 80.6 MB | MLX weights (float32) |
| omniparser_mlx.json | metadata | nc=1, class="icon" |

## Usage





## Performance

~110ms inference on M4 (vs ~400ms ultralytics CPU).

## Code

[github.com/walter-grace/yolo11-mlx](https://github.com/walter-grace/yolo11-mlx)