API reference

Note

This is the documentation for the atomistic capabilities of metatensor-torch version 0.5.4. For other versions, look in the following pages:

Version 0.5

0.5.4

https://docs.metatensor.org//metatensor-torch-v0.5.4/atomistic/reference/index.html

0.5.3

https://docs.metatensor.org//metatensor-torch-v0.5.3/atomistic/reference/index.html

0.5.2

https://docs.metatensor.org//metatensor-torch-v0.5.2/atomistic/reference/index.html

0.5.1

https://docs.metatensor.org//metatensor-torch-v0.5.1/atomistic/reference/index.html

0.5.0

https://docs.metatensor.org//metatensor-torch-v0.5.0/atomistic/reference/index.html

Version 0.4

0.4.0

https://docs.metatensor.org//metatensor-torch-v0.4.0/atomistic/reference/index.html

Version 0.3

0.3.0

https://docs.metatensor.org//metatensor-torch-v0.3.0/atomistic/reference/index.html

Older versions

0.2.1

https://docs.metatensor.org//metatensor-torch-v0.2.1/atomistic/reference/index.html

0.2.0

https://docs.metatensor.org//metatensor-torch-v0.2.0/reference/torch/index.html

The atomistic machine learning models capabilities of metatensor are part of the TorchScript API, and are installed when installing the TorchScript API. We provide code for using atomistic systems as input of a machine learning model with metatensor.torch.atomistic.System, and exporting trained models with metatensor.torch.atomistic.MetatensorAtomisticModel.

• Systems
• Models
• Atomic Simulation Environment (ASE) integration

C++ API reference

The C++ API can be accessed by including <metatensor/torch/atomistic.hpp>.

• C++ API reference