> For the complete documentation index, see [llms.txt](https://guide.petoi.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://guide.petoi.com/extensible-modules/petoi-ai-vision-module/custom-petoi-vision-model/steps-on-how-to-get-started/step-1-choose-or-make-your-own-dataset.md).

# Step 1: Choose or Make your own dataset

Before we can teach our AI to "see," we have to give it something to look at! In machine learning, a dataset is the collection of images we use to train our model. Think of it as the textbook your AI will study.

You have two paths to get started: you can either grab a textbook that is already written, or you can write your own from scratch.

{% hint style="info" %}
**Choose the fast track** when you want to learn the workflow. **Create custom data** when your robot must recognize something unique.
{% endhint %}

***

## Path A: Choose a Pre-Labeled Dataset (The Fast Track) 🚀

If you want to jump straight into training and see results quickly, this is the path for you. You will use open-source images that other developers have already collected and labeled.

* Best for: Beginners, quick testing, and general objects (like cars, animals, or people).
* Where to find them:
  * Roboflow Universe: A massive library of over 750,000 pre-labeled computer vision datasets ready to download.
  * Kaggle: A popular data science community with thousands of free, high-quality image datasets.
* Your Mission: Browse these platforms, find a dataset that interests you, and download it in the correct format for your model.

{% hint style="success" %}
**Fast-track goal:** Download a dataset that matches the objects you want to detect.
{% endhint %}

## Path B: Create a Custom Training Dataset (The Real-World Experience) 📸

If you want your AI to recognize something highly specific—like a specific part on a circuit board, or your own pet—you have to build the dataset yourself. This is how professional AI pipelines operate in the real world!

* Best for: Highly specific projects, custom hardware, and unique environments.
* How it works:
  1. Gather Images: Take a wide variety of photos of your target object.
  2. Upload: Bring those images into an annotation tool like Label Studio.
  3. Draw Bounding Boxes: Manually draw boxes around your target object in every single image so the AI knows exactly what to look for.
  4. Export: Save your newly labeled dataset!

> 💡 **Pro-Tip for Custom Data:** Quality matters more than quantity! Make sure you take photos of your object from different angles, against different backgrounds, and in different lighting conditions so your AI doesn't get confused later.

{% hint style="warning" %}
Keep every label consistent. A `ball` must always use the same class name.
{% endhint %}

***

### Before you continue

Confirm that your dataset is ready:

* [ ] Every image contains a target object.
* [ ] Each target object has a correct bounding box.
* [ ] Each class uses one consistent label.

***

Next Up: Once you have your dataset ready, it is time to move to the cloud. Click next to proceed to [Step 2: Train Model](/extensible-modules/petoi-ai-vision-module/custom-petoi-vision-model/steps-on-how-to-get-started.md#step-2-train-model).


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## 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, and the optional `goal` query parameter:

```
GET https://guide.petoi.com/extensible-modules/petoi-ai-vision-module/custom-petoi-vision-model/steps-on-how-to-get-started/step-1-choose-or-make-your-own-dataset.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

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.