Skip to content

Reproducibility and Sharing Results

Reproducible research depends on being able to recreate an experiment's environment, code, and data at any later point. This page gives a short overview of the practices involved.

Before you begin

  • Launch jobs

    Submit and manage jobs on the cluster with Slurm.

 

Environment

Pin the exact versions of Python and its dependencies so an experiment can be recreated on any node. See the portability guide for managing dependencies and lockfiles with uv.

Version control

Track code changes with Git and push to a hosted remote such as GitHub or GitLab. Tag or reference the exact commit used to produce a given set of results, so the code behind any experiment or publication stays retrievable.

Sharing datasets

To share a dataset with the Mila community, request its addition to /network/datasets. See the Datasets page for ways to publicly share a Mila-hosted dataset — Academic Torrent, Google Drive, and registering a DOI for citation.

For datasets that do not need to live on Mila's storage, the Hugging Face Hub hosts and versions datasets alongside model cards and loading scripts, and GitHub (or GitLab) works well for small datasets kept alongside the code that produces or consumes them.

Research paper distribution

When publishing results, bundle the code, environment, and dataset references readers need to reproduce them — for example through a public repository release alongside the paper.

Acknowledging Mila

Researchers are requested to acknowledge Mila when publications or reports are published on research that used compute resources provided by the organization. Our ability to showcase the successful use of our resources builds an important case for long-term and sustainable funding of Mila and our operations, i.e. have more compute resources available to researchers!

The exact wording of the acknowledgment may vary, but please ensure that Mila is mentioned. The following can be used as a guideline:

"This research was enabled in part by compute resources provided by Mila (mila.quebec)."

If you find useful some of the software developed at Mila or the technical help you can get from our experts you can also mention it in the acknowledgment:

"This research was enabled in part by compute resources, software and technical help provided by Mila (mila.quebec)."

Next step

 

Comments