When installing and learning a new tool, the hardest part is usually figuring out the first steps that actually resemble your real workflow. FiftyOne’s getting started guides are designed to show you the way. These guides walk you through complete workflows based on real computer vision use cases, including step-by-step instructions and companion notebooks wired together so you can quickly load datasets and explore FiftyOne’s core features.

Each getting started guide begins with what dataset or problem it focuses on, an estimated difficulty level, expected time to completion, and what you’ll have at the end. The main content is broken into easy to follow step-by-step sections.

Prerequisites and system requirements are listed up front and are also included in the accompanying notebook. You can treat each notebook as a self-contained lab that you can open on your own machine or in an environment like Google Colab. Most notebooks will start with installing FiftyOne and its dependencies.

Our FiftyOne guides include a range of workflows and industries so you can onboard in the context that makes sense for you. Here are a few examples.

This guide walks you through the fundamentals of object detection with FiftyOne. You’ll load a detection dataset, add predictions, use FiftyOne Brain to uncover mistakes and edge cases, and run model evaluations.

The self-driving guide showcases multi-sensor and multi-modal workflows that mirror modern AV stacks via the nuScenes dataset. You’ll work with complex scenes, annotations, and visual embeddings that are grounded in the real-world challenges of autonomy.

This guide helps you load and perform operations on 3D data such as point clouds and 3D meshes. You will also work with common label types like 3D annotations and polylines.

This medical imaging guide helps you load and evaluate MRI and CT datasets. You’ll work with formats like DICOM and NIfTI, organize grouped datasets, and apply and evaluate segmentation masks to medical data.

FiftyOne offers pre-formatted datasets via the FiftyOne Dataset Zoo and a variety of pre-trained models via the FiftyOne Model Zoo. This FiftyOne guide will walk you through loading datasets from the dataset zoo, applying models from the model zoo, and working toward integrating your own datasets and models.

Working through several of these guides will give you a mental model of where FiftyOne fits in your workflow. From there, you can look at the more focused FiftyOne tutorials for advanced workflows like fine-tuning, anomaly detection, or visual search. The docs also include copy-paste recipes for common operations like de-duplication, dataset merging, or format conversion.

If you’re building in a team or production setting, be sure to check out FiftyOne Enterprise. The patterns you learn will map directly to Enterprise capabilities like data like integration, fine-grained dataset versioning, and advanced labeling and data quality analysis.

So, start with the guide that matches your domain, and then apply the same patterns to your own data. When you’re ready to scale to bigger teams and bigger datasets, talk to us about FiftyOne Enterprise and turn your workflows into production-ready visual AI applications.