# commaai/research

dataset and code for 2016 paper "Learning a Driving Simulator"

Repository: https://github.com/commaai/research
Canonical: https://ross.abutalabs.com/products/research
Language: Python
License: BSD-3-Clause
License Family: permissive
Last push: 2022-08-16T09:22:12+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3682, "days_push": 1478, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4123, forks 1157 (observed 2026-08-28T04:08:36.007920+00:00)

## What it is
The comma.ai driving dataset with accompanying code for the 2016 paper 'Learning a Driving Simulator'. It contains ~7.25 hours of highway driving video at 20 Hz plus synchronized vehicle telemetry (speed, steering angle, GPS, gyroscope) in HDF5 format, with example experiments for steering angle prediction and generative driving image models.

## Use cases
- train a steering angle predictor from driving video
- train a generative model that simulates driving images
- research datasets for autonomous driving
- learn deep learning with real car sensor data
- download highway driving footage with telemetry
- reproduce the Learning a Driving Simulator paper

## When to choose
- you need real-world driving video paired with steering/speed/GPS logs for ML research
- you want to reproduce or extend the 2016 driving simulator paper experiments
- you're studying computer vision or generative models on automotive data

## When to avoid
- you need a production driving simulator or openpilot software - this is research data only
- you need commercial-use rights - the dataset is CC BY-NC-SA 3.0
- you need modern tooling - the examples use TensorFlow 0.9 and Keras 1.0.6
- you need urban or off-highway driving data - footage is largely highway

## Facets
- artifact type: dataset
- maturity: maintenance
- function: machine-learning, computer-vision, data-science
- domain: machine-learning, autonomous-vehicles, computer-vision, data-science
- platform: python, cross-platform
- tags: driving-dataset, autonomous-driving, steering-prediction, generative-models, research-paper, hdf5, comma-ai

## Member repositories
- commaai/research (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:36.007920+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:23:05.466637+00:00, confidence not recorded.
  - readme: https://github.com/commaai/research (fetched 2026-08-28T04:08:36.007920+00:00, sha 3b0d79dc8512)
- Data as of 2026-08-30T08:39:29.467469+00:00.