software
We highly value and support free and open-source software (FOSS) development. Besides using as much FOSS as possible, everybody in the team is encouraged to contribute to open-source projects and to share reproducible code for our projects. In this page, you can find some utilities developed by members of the team, which can be useful for your research. If you have any questions about the code, please feel free to contact us. The code for each group publication can be found on the publications page.
Dawgz
Scheduling library for Slurm clusters in Python.
Author: François Rozet
Dawgz allows to easily schedule job workflows on Slurm clusters, entirely from Python, replacing the need for bash scripts.
Azula
Diffusion models library in PyTorch.
Author: François Rozet
Azula is a library for training and sampling from diffusion models, with a focus on unifying the different formalisms and notations of the generative diffusion models literature into a single, convenient and hackable interface.
Zuko
Normalizing flows library in PyTorch.
Author: François Rozet
Zuko is a library for training and sampling from normalizing flows in PyTorch, providing more powerful abstractions for distributions than native PyTorch.
Inox
PyTorch-style neural networks library in JAX.
Author: François Rozet
Inox is a library for building and training neural networks in JAX, with a PyTorch-like interface.