Installation#

This section explains how to install the required software stack natively on your system. However, you might prefer using CompressAI-Vision via docker instead.

1. Python dependencies#

Here we assume you are running Ubuntu 20.04 LTS (or newer)

First, create and activate the virtualenv with:

python3 -m venv venv
source ./venv/bin/activate
pip install -U pip

You might want to define the python version explicitly, i.e. with python3.8 -m venv venv. Python3.8+ is required.

While in the activated virtualenv, run the installation:

bash scripts/install.sh

The script will install the following software stack:

After running the script within the virtualenv, you might need to deactivate and activate it for the effects to take place.

2. If your pipeline includes traditional codecs#

Install binaries from sources, clone the selected repo from the following links and select the desired tag

Please refer to the respective Readmes to build the binaries, that will be launched from compressai-vision code.