Skip to content

Latest commit

 

History

History
57 lines (50 loc) · 2.36 KB

INSTALL.md

File metadata and controls

57 lines (50 loc) · 2.36 KB
Raw

Installation

We use AWS EC2 instances, but you can also setup the environment and run all experiments on your own computer or cluster.

Pre-installation

  • Launch an AWS EC2 g4dn.xlarge instance with Deep Learning Base AMI (Ubuntu 18.04) and ssh to it.
  • Install Miniconda.
  • (Optional) Generate and add SSH key to your github account.

Install CARLA

mkdir carla
cd carla
wget https://carla-releases.s3.eu-west-3.amazonaws.com/Linux/CARLA_0.9.11.tar.gz
tar -xvzf CARLA_0.9.11.tar.gz
echo "export CARLA_ROOT=/home/ubuntu/carla" >> /home/ubuntu/.bashrc
cd Import && wget https://carla-releases.s3.eu-west-3.amazonaws.com/Linux/AdditionalMaps_0.9.11.tar.gz
cd .. && bash ImportAssets.sh
rm CARLA_0.9.11.tar.gz Import/AdditionalMaps_0.9.11.tar.gz
rm Import/AdditionalMaps_0.9.10.1.tar.gz Import/AdditionalMaps_0.9.11.tar.gz
cd && source .bashrc

Setup carla-roach

  • Clone this repository and create conda environment.
git clone git@github.com:zhejz/carla-roach.git
cd carla-roach
conda env create -f environment.yml --name carla
conda activate carla
easy_install ${CARLA_ROOT}/PythonAPI/carla/dist/carla-0.9.11-py3.7-linux-x86_64.egg
  • We use wandb for logging, please register a free account and login to it.
wandb login

Post-installation

  • Create an AMI for this instance.
  • Create launch template for that AMI.

For RL training

Our RL training crashed more often on CARLA 0.9.11 than on CARLA 0.9.10.1. So for RL training we create an environment with the CARLA 0.9.10.1 release. Get a new AWS instance, all installation steps remain the same, just the CARLA version is changed

wget https://carla-releases.s3.eu-west-3.amazonaws.com/Linux/CARLA_0.9.10.1.tar.gz
wget https://carla-releases.s3.eu-west-3.amazonaws.com/Linux/AdditionalMaps_0.9.10.1.tar.gz

and the filename of the CARLA library is different

easy_install ${CARLA_ROOT}/PythonAPI/carla/dist/carla-0.9.10-py3.7-linux-x86_64.egg

You may want to create a specific AMI and launch template for this environment. On your local machine, you may consider to install two CARLA and have two environment variable ${CARLA_ROOT_911} and ${CARLA_ROOT_910}.