Jet is a cross-platform C++ and Python-based library for simulating quantum circuits using tensor network contractions. Accelerate quantum circuit simulations with task-based parallelism across multiple devices easily, with built-in support for concurrent tensor contractions on CPUs and GPUs.
Run anywhere. Jet runs on a variety of systems, from single-board machines to massively-parallel supercomputers, make the most of your CPU and GPU resources!
Fast. Jet accelerates tensor contractions using a novel task-based parallelism approach, executing simulations even faster by distributing contractions across multiple devices.
Machine learning. Jet provides tools for training machine learning models based on tensor network architectures such as MPS, PEPS, MERA, and TTN.
Qudits. Model quantum systems with an arbitrary number of basis states, allowing simulations from qubit to photonic algorithms, and everything in between.
Once you have Jet up and running, explore our tutorials and demonstrations to learn about tensors and how they are modelled in Jet, and how to take advantage of task-based parallelization to speed up your tensor contractions.