Skip to content
@metatensor

metatensor

Welcome to metatensor

metatensor logo

We develop software to enable modularity and colaboration in atomistic machine learning; through data, models and architecture exchange.

We have three main projects:

  • metatensor is a generic self-describing sparse data format, that can be used by multiple pieces of code to exchange data without knowing about each other;
  • metatomic is a interface to execute arbitrary atomistic machine learning models in arbitrary simulation engines;
  • metatrain is a command line tool to train state of the art atomistic machine learning models on custom datasets, and expand the capabilities of existing model architectures.

We also have a couple of projects built on top of the above three:

  • featomic is a tool to compute machine learning features for atomistic systems, which can then be used in larger models
  • lammps-metatomic is our fork of LAMMPS with added support for metatomic models

Pinned Loading

  1. metatensor metatensor Public

    Self-describing sparse tensor data format for atomistic machine learning and beyond

    Python 73 23

  2. metatomic metatomic Public

    Atomistic machine learning models you can use everywhere for everything

    Python 15 2

  3. metatrain metatrain Public

    Training and evaluating machine learning models for atomistic systems.

    Python 32 10

  4. featomic featomic Public

    Computing representations for atomistic machine learning

    Rust 71 15

Repositories

Showing 10 of 24 repositories

Top languages

Loading…

Most used topics

Loading…