PINA: a PyTorch Framework for Deep Differential Equation Learning for Research and Production Environments
Keywords: PINN, Neural Operator, Differential Equations, Software
TL;DR: A PyTorch open-interface software for learning differential equation via deep learning methods
Abstract: The last years have manifested the artificial intelligence revolution in several fields. Within the scientific computing community, there has been a remarkable effort to exploit the advancements in machine learning to address the limitations of conventional methods for solving differential equations. Notably, the physics-informed neural network (PINN) and neural operator (NO) approaches have emerged as central players due to their promising and innovative approaches to compute differential equations' solutions. In this contribution, we are going to present a versatile software designed for tackling differential equation learning using PINN and NO methodologies. The package is called PINA, an open-source Python library built upon the robust foundations of PyTorch and Lightning. It empowers end-users to formulate their problem and craft their models to effortlessly compute the solution. The modular structure of PINA permits it to adapt PINN and NO schemas for user specifics, thus offering the freedom to select the most suitable learning techniques for their particular problem domain. Furthermore, by leveraging the capabilities of the Lightning package, PINA adapts to various hardware setups, including GPUs and TPUs. This adaptability positions PINA as an ideal candidate for the transition of these methodologies into production and industrial pipelines, where computational efficiency and scalability are of paramount importance.
Supplementary Material: zip
Primary Area: infrastructure, software libraries, hardware, etc.
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 6990
Loading