Deepdrive is a simulator that allows anyone with a PC to push the state-of-the-art in self-driving.
Deepdrive 2.0
The easiest way to experiment with self-driving AI
Features
Baseline agent
Use our
pretrained example agent,
training code,
and
dataset
to get started building your own AI and try to beat our high score!
Dataset
Download the 100GB, 8.2 hour, dataset
we used to train the baseline agent, complete with depth,
DAgger labels,
and capture augmentation used to achieve scores on-par with the path-follower.
Modifiable source
Our AI,
Unreal plugin,
simulation assets and code,
and Unreal Engine are all on GitHub, so you can
modify the agent and the environment to automate labeling and iterate on ideas in minutes
instead of weeks or months.
Eight cameras with depth
Enable up to eight cameras with images and depth
transferred at blazing speeds via shared memory. i.e.
Eight 512x512 48 bit RGB with 16 bit depth channel cameras, uncompressed on
a GTX 980 over PCIe 3.0 x 16 are available at 20Hz as NumPy arrays.
Linux / Windows support
First class support for Linux and Windows, one command
python install.py
, and you'll be piloting Deepdrive in minutes.
Visual scripting
Quickly change anything in the simulation with
Unreal
Blueprints,
using the full power of Unreal's C++ API.
C++ extendability
Everything is C++ under the hood,
so you can integrate existing systems and highly optimize anything you want.
Python API
Interface through the Gym API
using a reward function based on speed, safety, legality, and comfort.
Example forward agent
import deepdrive def main(): env = deepdrive.start() forward = deepdrive.action(steering=0, throttle=1, brake=0) done = False while not done: observation, reward, done, info = env.step(forward) if __name__ == '__main__': main()
GitHub
Deepdrive - Agents, training, and API to simulator - start here!
Deepdrive sim - Unreal project, Unreal Plugin, assets, and Python extension