Skip to content

Installation

Installing ado

ado can be installed in one of three ways:

  1. From PyPi
  2. From GitHub
  3. By cloning the GitHub repository locally
Warning

Before proceeding ensure you are using a supported Python version: run python --version in your terminal and check that you are on either Python 3.10, 3.11, or 3.12.

It is also highly recommended to create a virtual environment for ado, to avoid dependency conflicts with other packages. You can do so with:

python -m venv ado-venv

And activate it with

source ado-venv/bin/activate

This method installs the ado-core package from PyPi

pip install ado-core
pip install git+https://github.com/IBM/ado.git
git clone https://github.com/IBM/ado.git
cd ado
pip install .

Installing plugins

ado uses a plugin system to provide additional actuators and operators. We maintain a set of actuators and operators in the ado main repo which you can see here. Some plugins may also have packages on PyPi. You can install these actuators as follows:

Info

Some plugins may have dependencies that may require credentials to access. Check the plugins's docs if you encounter issues installing a specific actuator.

The following plugin packages are available: ado-sfttrainer, ado-vllm-performance and ado-ray-tune

pip install $PLUGIN_NAME

For actuators

pip install "git+https://github.com/IBM/ado.git#subdirectory=plugins/actuators/$ACTUATOR_NAME"

For operators

pip install "git+https://github.com/IBM/ado.git#subdirectory=plugins/operators/$OPERATOR_NAME"

If you've cloned the ado repository locally in the previous step, you can run from the top-level of the cloned repository

pip install plugins/actuators/$ACTUATOR_NAME

or

pip install plugins/operators/$OPERATOR_NAME

What's next