DeepDrive in Universe

Setup

git clone https://github.com/deepdrive/deepdrive-universe

Note: You will need GPU acceleration to do inference at the 8Hz that the model runs. Slower (or faster) inference may work, but is not the standard way the model is run.

Baseline model - Tensorflow version

Thanks to Rafal jozefowicz for contributing this model

cd deepdrive-tf
wget https://www.dropbox.com/s/fsummbpqlfildnq/model.ckpt-20048?dl=1

Baseline model - Caffe version

Install the latest version of Caffe and download the model via

cd deepdrive
wget https://www.dropbox.com/s/z92c4otvyofgl3f/caffe_deep_drive_train_iter_35352.caffemodel?dl=1

Baseline models were trained with the standard hood camera in GTAV.

To enable the hood camera, hit v until you see something like this

If you see the steering wheel, change the camera settings like so:

In addition, enable borderless mode to avoid sending the window chrome through the network

Run the GTAV Universe environment

Follow the env setup instructions here

Run the model

python main.py -d [DeepDriver|DeepDriverTF] -r vnc://<your-env-ip>:5900+15900

Directly connected machines

To connect your Windows and Ubuntu machines directly via ethernet, follow these instructions for adding the interface to Ubuntu.

Use the Netmask provided by ipconfig in Windows for your ethernet interface.