Skip to content
Conditional Affordance Learning for Driving in Urban Environments using reduced data
Python Jupyter Notebook Shell
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

Conditional Affordance Learning

Initial code :

Reference [Paper]

Our model uses concatenated images to give us a wider receptive field and also performs considerably well on a reduced dataset by performing key frame extraction.

Find more about our work in our presentation


# install anaconda2 if you don't have it yet
source ~/.profile
# or use source ~/.bashrc - depending on where anaconda was added to PATH as the result of the installation
# now anaconda is assumed to be in ~/anaconda2

Now we will:

  1. create a conda environment named CAL and install all dependencies
  2. download the binaries for CARLA version 0.8.2 [CARLA releases]
  3. download the model weights
git clone
cd CAL

# create conda environment
conda env create -f requirements.yml
source activate CAL

# run download script

Run the Agent

In CARLA_0.8.2/ start the server with for example: (see the CARLA documentation for details)

./ Town01 -carla-server -windowed -benchmark -fps=20 -ResX=800 - ResY=600

Open a second terminal, cd into CAL/PythonClient/ and run:

python -c Town02 -v -n test

This runs the basic_experiment benchmark. '-n' is the naming flag (in this example the run is named "test"). If you want to run the CORL 2017 benchmark you need to run

python -c Town02 -v -n test --corl-2017

If you want to continue an experiment, you can add the 'continue-experiment' flag.


cd training/

# download and untar the dataset
tar -xzvf dataset.tar.gz

# create the training environment
conda env create -f requirements.yml
source activate training_CAL

Now, open training_CAL.ipynb. The notebook walks you through the steps to train a network on the dataset.

You can’t perform that action at this time.