> For the complete documentation index, see [llms.txt](https://aegion-dynamic.gitbook.io/learning-resources/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://aegion-dynamic.gitbook.io/learning-resources/machine-learning/env-recommendations.md).

# Environment Setup

## Recommended Setup

* Ubuntu (Current LTS Version)
* [pyenv](https://github.com/pyenv/pyenv) - for python environment management
* [poetry](https://python-poetry.org/) - For project and dependency management
* [docker](https://docs.docker.com/engine/install/ubuntu/)
* CUDA Drivers
* VSCode
* Increase the swap size on you machine to at least 20GB ([tutorial](https://www.digitalocean.com/community/tutorials/how-to-add-swap-space-on-ubuntu-20-04))

## Don't

At Aegion, we don't use `anaconda`. While `anaconda` is a very popular package management system, in order to ensure that all our projects have the the most detailed dependency trees, we recommend the learner to familiarize themselves with `poetry` to ensure that all the dependencies are captured accurately with precise versioning and environment constraints. This is critical as we move prototypes to production.

Check out the following tutorial to get familiar with `poetry`:

{% embed url="<https://realpython.com/dependency-management-python-poetry/>" %}