DI-drive
Updated on 2022.2.25 DI-drive-v0.3.1 (beta)
DI-drive - Decision Intelligence Platform for Autonomous Driving simulation.
DI-drive is application platform under OpenDILab
Introduction
DI-drive is an open-source application platform under OpenDILab. DI-drive applies different simulator/datasets/cases in Decision Intelligence Training & Testing for Autonomous Driving Policy. It aims to
- run Imitation Learning, Reinforcement Learning, GAIL etc. in a single platform and simple unified entry
- apply Decision Intelligence in any parts of driving simulation
- suit most of the driving simulators input & output
- run designed driving cases and scenarios
and most importantly, to put these all together!
DI-drive uses DI-engine, a Reinforcement Learning
platform to build most of the running modules and demos. DI-drive currently supports Carla,
an open-source Autonomous Drining simulator to operate driving simualtion, and MetaDrive,
a diverse driving scenarios for Generalizable Reinforcement Learning. Users can specify any of them to run in global config under core
.
Installation
DI-drive needs to have the following modules installed:
- Pytorch
- DI-engine
MetaDrive can be easily installed via pip
.
If Carla server is used for simulation, users need to install 'Carla Python API' in addition.
Please refer to the documentation for details about installation and user guide of DI-drive.
We provide IL and RL tutorials, and full guidance for quick run existing policy for beginners.
Please refer to FAQ for frequently asked questions.
Model Zoo
Imitation Learning
- Conditional Imitation Learning
- Learning by Cheating
- from Continuous Intention to Continuous Trajectory
Reinforcement Learning
- BeV Speed RL
- Implicit Affordance
- Latent DRL
- MetaDrive Macro RL
DI-drive Casezoo
DI-drive Casezoo is a scenario set for training and testing of Autonomous Driving policy in simulator. Casezoo combines data collected by real vehicles and Shanghai Lingang road license test Scenarios. Casezoo supports both evaluating and training, whick makes the simulation closer to real driving.
Please see casezoo instruction for details about Casezoo.
Contributing
We appreciate all contributions to improve DI-drive, both algorithms and system designs.
License
DI-engine released under the Apache 2.0 license.
Citation
@misc{didrive,
title={{DI-drive: OpenDILab} Decision Intelligence platform for Autonomous Driving simulation},
author={DI-drive Contributors},
publisher = {GitHub},
howpublished = {\url{https://github.com/opendilab/DI-drive}},
year={2021},
}