What is Netket?

NetKet is an open-source project delivering cutting-edge methods for the study of many-body quantum systems with artificial neural networks and machine learning techniques.

23 august 2021: NetKet 3 ❤️ Jax

18 months in development, NetKet 3.0 indicates a major new step for the NetKet project. NetKet has been totally rewritten in Python and is now a Jax-based library. This guarantees outstanding performance while allowing researchers to exploit machine-learning frameworks to define advanced Neural-Networks.

GPUs and Google’s TPUs are now supported too!

Update now and try the new examples!

Neural Quantum States

NetKet provides state-of-the-art Neural-Network Quantum states, and advanced learning algorithms to find the ground-state of many-body Hamiltonians.

Easy to Learn

NetKet has a library of simple Neural Quantum states and sensible defaults, allowing you to get started with as few as 10 lines of code.

Highly Customizable

NetKet provides a modular infrastructure for the development and application of machine-learning techniques to many-body quantum systems. You can set up your custom many-body Hamiltonian, observables, lattices, and machines in minutes.


Netket is based on Jax, therefore you can use any Neural Network Architecture written in one of the several Jax Frameworks, such as haiku or flax.

Run in Parallel

The learning algorithms used in NetKet are intrinsically amenable to massive parallelism. NetKet is built using MPI primitives, and can scale up to thousands of CPU cores.

Collaborative Effort

NetKet wants to be a common platform for the development of new algorithms to study the most challenging open problems in many-body quantum physics. Building upon a set of well-tested primitives and on a solid infrastructure, researchers can get publication-grade results in less time.