# OpenDriveLab/UniAD

[CVPR 2023 Best Paper Award] Planning-oriented Autonomous Driving

Repository: https://github.com/OpenDriveLab/UniAD
Canonical: https://ross.abutalabs.com/products/uniad
Language: Python
License: Apache-2.0
License Family: permissive
Topics: autonomous-driving, end-to-end-autonomous-driving, motion-prediction, multi-object-tracking, perception-prediction-planning, bev-segmentation, motion-planning, occupancy-prediction, autonomous-driving-framework
Last push: 2025-10-29T03:13:35+00:00

## Health v2 (maintenance only)
Score: 44/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 49, release rhythm 8, longevity 97
- inputs: {"age_days": 1365, "days_push": 308, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4737, forks 549 (observed 2026-08-28T04:08:57.764561+00:00)

## What it is
UniAD is a unified end-to-end autonomous driving framework that hierarchically casts perception, prediction, and planning tasks under a planning-oriented philosophy. It won the CVPR 2023 Best Paper Award and was released as UniAD 2.0 in October 2025 with upgrades to mmdet3d 1.x and PyTorch 2.x.

## Use cases
- run end-to-end autonomous driving research on nuScenes
- benchmark motion prediction for autonomous vehicles
- evaluate open-loop planning with collision metrics
- train multi-task perception-prediction-planning models
- do BEV segmentation and occupancy prediction from camera data
- reproduce CVPR 2023 best paper results

## When to choose
- you need a unified multi-task autonomous driving pipeline rather than standalone modules
- you want a research baseline for end-to-end planning on nuScenes
- you need SOTA motion prediction and occupancy prediction benchmarks

## When to avoid
- you need a production-ready autonomous driving stack for real vehicles
- you lack the required GPUs or nuScenes dataset access
- you only need a single task like object detection without the full pipeline

## Facets
- artifact type: framework
- maturity: active
- function: machine-learning, deep-learning, computer-vision, simulation
- domain: autonomous-vehicles, deep-learning, computer-vision, artificial-intelligence
- platform: python
- tags: autonomous-driving, end-to-end, motion-prediction, multi-object-tracking, planning, bev-segmentation, cvpr-2023, nuscenes, nuplan, gpu, linux

## Member repositories
- OpenDriveLab/UniAD (main) score 44

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:57.764561+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 2026-08-29T18:18:58.880268+00:00, confidence not recorded.
  - readme: https://github.com/OpenDriveLab/UniAD (fetched 2026-08-28T04:08:57.764561+00:00, sha 85b116c0ef0d)
- Data as of 2026-08-30T08:39:29.467469+00:00.