# From Ultralytics

<figure><img src="/files/fny7adCeTkmXg2LfNMGn" alt="Ultralytics logo" width="500"><figcaption></figcaption></figure>

Ultralytics develops and maintains popular open-source computer vision libraries including YOLOv8 and YOLO11. These are state-of-the-art object detection models well suited for edge deployment with the Nx AI Manager.

For a list of pre-built YOLO models that are already available in the Nx AI Cloud catalogue, see [Common Models](/nx-ai-manager-v6.1.2/ai-models-support/importing-models/common-models.md).

## Exporting a YOLO model for the Nx AI Manager

To use a custom or fine-tuned YOLO model with the Nx AI Manager, export it to ONNX format and then upload it to the Nx AI Cloud.

### Step 1: Install Ultralytics

```bash
pip install ultralytics
```

### Step 2: Export the model to ONNX

Replace `yolov8n.pt` with your own model checkpoint if you have trained a custom model:

```bash
yolo export model=yolov8n.pt format=onnx
```

This produces a `yolov8n.onnx` file in the current directory. The ONNX export is compatible with the Nx AI Manager's CPU, Intel OpenVino, and NVIDIA CUDA runtimes.

{% hint style="info" %}
If your model uses a non-standard input size, pass `imgsz=<size>` to the export command, for example `yolo export model=yolov8n.pt format=onnx imgsz=640`.
{% endhint %}

### Step 3: Upload to Nx AI Cloud

Upload the `.onnx` file to the Nx AI Cloud. The platform will automatically convert it for all supported AI accelerators.

{% content-ref url="/pages/Z7ndCO8n1txSBW7CSIq7" %}
[Upload your model](/nx-ai-manager-v6.1.2/nx-ai-cloud-ui/upload-your-model.md)
{% endcontent-ref %}


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://nx.docs.scailable.net/nx-ai-manager-v6.1.2/ai-models-support/importing-models/from-ultralytics.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.