# Datasets

To get a good understanding of how a synthesizer performs, we recommend running it on a wide variety of datasets that represent real world scenarios. The SDV ecosystem library offers a variety of [public demo datasets](https://docs.sdv.dev/sdgym/customization/datasets/public-sdv-datasets) that are available for anyone to use. By default, SDGym is set to use them.

{% hint style="success" %}
**Understand your data.** Use SDGym to summarize the datasets by key characteristics such as the size of the data, number of tables, numbers of each kinds of columns, and more. For more information, see the [Explore Datasets](https://docs.sdv.dev/sdgym/customization/datasets/explore-datasets) guide.
{% endhint %}

In addition to the public demo datasets, you can also provide your own datasets that are stored in an Amazon S3 bucket. For more information see the [Dataset Format](https://docs.sdv.dev/sdgym/customization/datasets/dataset-format) guide.