maziarraissi/PINNs

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

6111 stars · 1608 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

The original reference implementation of Physics-Informed Neural Networks (PINNs), which train neural networks to solve and discover nonlinear partial differential equations while respecting physical laws. It provides continuous-time and discrete-time models for data-driven PDE solutions and parameter discovery, written in Python with TensorFlow 1.x.

Use cases

• solve partial differential equations with neural networks
• discover PDE coefficients from data
• build physics-informed surrogate models
• learn PINNs from the original paper's code
• run forward and inverse PDE problems with deep learning
• reproduce Raissi et al. PINN benchmark results

When to choose

• you want the canonical reference implementation accompanying the original PINN papers
• you are studying or extending PINN research and need the exact algorithms from the 2017-2019 publications
• you need data-driven solution or discovery of nonlinear PDEs in a TensorFlow 1.x setting

When to avoid

• you need an actively maintained library - the repo is explicitly no longer maintained
• you use modern PyTorch, JAX, or TensorFlow 2 - the authors recommend pinns-torch, pinns-jax, or pinns-tf2 instead
• you need production-grade, well-tested scientific software rather than research code

Facets

library · maturity maintenance

machine-learning deep-learning simulation math machine-learning deep-learning simulation mathematics artificial-intelligence python physics-informed-neural-networks pde-solver scientific-computing surrogate-modeling tensorflow research-code

Member repositories

For agents

markdown · JSON · MCP: product_card(name="maziarraissi/PINNs")