deeponet

A library for scientific machine learning and physics-informed learning

Physics Informed Machine Learning Tutorials (Pytorch and Jax)

DeepONets, (Fourier) Neural Operators, Physics-Informed Neural Operators, and more in Julia

Source code of "Learning nonlinear operators in latent spaces for real-time predictions of complex dynamics in physical systems."

Source code of 'Deep transfer operator learning for partial differential equations under conditional shift'.

Code for training and inferring acoustic wave propagation in 3D

No need to train, he's a smooth operator

Official repo for separable operator networks -- extreme-scale operator learning for parametric PDEs.

NodeLab is a simple MATLAB-repository for node-generation and adaptive refinement for testing, and implementing various meshfree methods (including physics-informed neural networks, PINNs and DeepOnet) for solving PDEs in arbitrary domains.

Implementations of three neural operators and application in Bayesian inference problems

Various methods for solving PDEs using artificial intelligence

Nonlinear model reduction for operator learning

Source code of "On the influence of over-parameterization in manifold based surrogates and deep neural operators".

Benchmarking Surrogates for coupled ODE systems.

We implement a Multifidelity-DeepONet that leverages both high-fidelity CFD simulations and real-time, low-fidelity sensor data. We also proved that Multifidelity-DeepONet has better performance compare to all the others baseline methods in our experiments.

NVIDIA PhysicsNeMo 中文教程书 · 8 章 + 4 附录 · PINN 与神经算子三种实现（裸 PyTorch / SDK / GPU production）

I-FENN framework integrating DeepONets (PyTorch) with FEM (deal.II) for accelerated, HPC-enabled, coupled multiphysics simulations (thermoelasticity & poroelasticity). Code for Amin et al., CMAME 2026.

This repository implements an Extended Physics-Informed DeepONet (XPIDON) model for simulating autoclave composite curing processes. It features a nonlinear decoder, subdomain-specific input normalization, and an adaptive and automated domain decomposition strategy.

Scientific machine learning for JAX/Flax NNX: neural operators (FNO family, DeepONet, PINO, UNO), physics-informed networks (PINN, FBPINN, XPINN), E(3)-equivariant atomistic potentials (SchNet, PaiNN, NequIP), differentiable Kohn-Sham DFT, SINDy equation discovery, uncertainty quantification (conformal, GPs, SBI), PDEBench benchmarking.

A comparative analysis of DeepONet and FNO architectures, benchmarking their performance on Function-to-Function (Heat Equation) vs. Parameter-to-Function (Elastic Bar) PDE problems to motivate hybrid operator designs.