Two public datasets supported by highly detailed maps to test, experiment, and teach self-driving vehicles how to understand the world around them.

Cities are complicated.

Untangling that complexity is difficult.

Solving a challenging problem requires experts with access to high-quality data.

The first self-driving datasets with...

...highly detailed maps

Publicly available datasets for self-driving research rarely include rich map data, even though detailed maps are critical to the development of self-driving systems.

We decided it’s time to change that.

Argoverse is the first large-scale self-driving data collection to include HD maps with geometric and semantic metadata — such as lane centerlines, lane direction, and driveable area. All of the detail we provide makes it possible to develop more accurate perception algorithms, which in turn will enable self-driving vehicles to safely navigate complex city streets.

...hundreds of thousands of interesting scenarios for motion forecasting

Motion forecasting — the process of predicting the movements of other vehicles, pedestrians, and road users based on past and present data — is one of the most challenging parts of developing safe self-driving vehicles. That’s because there are many variables, many potential outcomes, and many moving objects on the roads around us.

Building more accurate motion forecasting models requires access to countless scenarios that illustrate the complexity of our roads and the things that happen on them. To support this, Argoverse contains a motion forecasting dataset with more than 300,000 curated scenarios, including unprotected left turns and lane changes, and provides a benchmark to promote testing, teaching, and learning.

Inside Argoverse


Learn about new competitions and additions to the Argoverse datasets by subscribing to our mailing list.