Data I/O

The first step in any FiftyOne workflow is loading your data into a FiftyOne dataset. FiftyOne natively supports dozens of common dataset formats, custom formats, and ready-to-use datasets from the Dataset Zoo and the Hugging Face integration.

New to FiftyOne? The Getting Started guides walk through loading data for common use cases, and FiftyOne Enterprise users can follow the Enterprise getting started guide to connect to their deployment.

In FiftyOne Enterprise, you can also query your data lake directly with Data Lens 🚀 and work with cloud-backed media 🚀 stored in S3, GCS, MinIO, or Azure.

And when you’re ready to move data out of FiftyOne — to a training pipeline, an annotation tool, or another team — you can export datasets in dozens of standard or custom formats.