Setting up your development environment

You have decided to contribute to optimum-neuron: at this stage you should have an up-to-date copy of the optimum-neuron repository installed locally (either a clone of the original repository if you have write access or a clone from your own fork if you are an external contributor).

Before contributing and submit your first pull-request, you need to prepare your development environment by installing a few development tools.

Prepare a python virtual environment

NOTE: ❗If you are using the Hugging Face Deep Learning AMI, you can reuse the virtual environment that is automatically activated when logging to the machine and skip this step.

$ python3 -m venv .venv
$ source .venv/bin/activate

Note: optimum-neuron requires at least python 3.10

Install development tools

First, you need to install the tools that are used to check that your contribution complies with the few optimum-neuron styling and coding rules.

$ pip install .[quality]
$ pre-commit install

Then, and only if you plan to modify the optimum-neuron code itself, you need to install the test environment.

$ pip install .[tests]

Creating a development branch

You cannot contribute your changes directly to the optimum-neuron main branch, so you need first to create a development branch containing your changes:

$ git checkout main
$ git pull
$ git checkout -b <my-awesome-contribution>

Committing your changes

All contributions are reviewed by the optimum-neuron maintainers: in order to speed up the review, you are strongly encouraged to submit your changes in small, atomic changes.

optimum-neuron has a few styling and coding policies that are enforced by the quality tools you previously installed.

You can however apply the styling tools manually before committing using explicit commands:

For any contribution:

$ pre-commit run end-of-file-fixer
$ pre-commit run trailing-whitespace

For python code:

```shell
$ pre-commit run ruff-check
$ pre-commit run ruff-format

Submitting your pull-request

We have prepared a pull-request template with a few instructions that you should read carefully before submitting.