Poutyne: all-purpose library for neural networks developed by our students Frédérik Paradis and David Beauchemin

Two IID student members, Frédérik Paradis and David Beauchemin, Ph.D. students in the Faculty of Sciences and Engineering at Université Laval, recently launched version 1.0 of their Poutyne library, developed with the objective of facilitating the development of neural network solutions. 


What Ph.D. students Frédérik Paradis and David Beauchemin have developed is essentially a one-size-fits-all library that facilitates the development of neural networks.

This library offers tools that allow you to train neural networks through the PyTorch library, perform backups, or even log training statistics. With a user-friendly and flexible interface, Poutyne adapts to the needs of the user – both for traditional classification networks and for research. Recently released, version 1.0 of Poutyne supports better documentation and examples, better indication of training progress with a progress bar and colours, built-in multi-GPUS support, and more.

Let’s keep in touch!

Would you like to be informed about IID news and activities? Subscribe now to our monthly newsletter.