⚛️ MLIP: SOTA Machine-Learning Interatomic Potentials in JAX 🚀

👀 Overview

mlip is a Python library for training and deploying Machine Learning Interatomic Potentials (MLIP) written in JAX. It provides the following functionality:

• Multiple model architectures (for now: MACE, NequIP and ViSNet)
• Dataset loading and preprocessing
• Training and fine-tuning MLIP models
• Batched inference with trained MLIP models
• MD simulations with MLIP models using multiple simulation backends (for now: JAX-MD and ASE)
• Energy minimizations with MLIP models using the same simulation backends as for MD.

The purpose of the library is to provide users with a toolbox to deal with MLIP models in true end-to-end fashion. Hereby we follow the key design principles of (1) easy-of-use also for non-expert users that mainly care about applying pre-trained models to relevant biological or material science applications, (2) extensibility and flexibility for users more experienced with MLIP and JAX, and (3) a focus on high inference speeds that enable running long MD simulations on large systems which we believe is necessary in order to bring MLIP to large-scale industrial application. See our inference speed benchmark below. With our library, we observe a 10x speedup on 138 atoms and up to 4x speed up on 1205 atoms over equivalent implementations relying on Torch and ASE.

See the Installation section for details on how to install MLIP-JAX and the example Google Colab notebooks linked below for a quick way to get started. For detailed instructions, visit our extensive code documentation.

This repository currently supports implementations of:

• MACE
• NequIP
• ViSNet

As the backend for equivariant operations, the current version of the code relies on the e3nn library.

📦 Installation

mlip can be installed via pip like this:

pip install mlip

However, this command only installs the regular CPU version of JAX. We recommend that the library is run on GPU. This requires also installing the necessary versions of jaxlib which can also be installed via pip. See the installation guide of JAX for more information. At time of release, the following install command is supported:

pip install -U "jax[cuda12]"

Note that using the TPU version of jaxlib is, in principle, also supported by this library. However, it has not been thoroughly tested and should therefore be considered an experimental feature.

Also, some tasks in mlip will require JAX-MD as a dependency. As the newest version of JAX-MD is not available on PyPI yet, this dependency will not be shipped with mlip automatically and instead must be installed directly from the GitHub repository, like this:

pip install git+https://github.com/jax-md/jax-md.git

Furthermore, note that among our library dependencies we have pinned the versions for jaxlib, matscipy, and orbax-checkpoint to one specific version only to prioritize reliability, however, we plan to allow for a more flexible definition of our dependencies in upcoming releases.

⚡ Examples

In addition to the in-depth tutorials provided as part of our documentation here, we also provide example Jupyter notebooks that can be used as simple templates to build your own MLIP pipelines:

• Inference and simulation
• Model training
• Addition of new models

To run the tutorials, just install Jupyter notebooks via pip and launch it from a directory that contains the notebooks:

pip install notebook && jupyter notebook

The installation of mlip itself is included within the notebooks. We recommend to run these notebooks with GPU acceleration enabled.

Alternatively, we provide a Dockerfile in this repository that you can use to run the tutorial notebooks. This can be achieved by executing the following lines from any directory that contains the downloaded Dockerfile:

docker build . -t mlip_tutorials
docker run -p 8888:8888 --gpus all mlip_tutorials

Note that this will only work on machines with NVIDIA GPUs. Once running, you can access the Jupyter notebook server by clicking on the URL displayed in the console of the form "http://127.0.0.1:8888/tree?token=abcdef...".

🤗 Pre-trained models (via HuggingFace)

We have prepared pre-trained models trained on a subset of the SPICE2 dataset for each of the models included in this repo. They can be accessed directly on InstaDeep's MLIP collection, along with our curated dataset or directly through the huggingface-hub Python API:

from huggingface_hub import hf_hub_download

hf_hub_download(repo_id="InstaDeepAI/mace-organics", filename="mace_organics_01.zip", local_dir="")
hf_hub_download(repo_id="InstaDeepAI/visnet-organics", filename="visnet_organics_01.zip", local_dir="")
hf_hub_download(repo_id="InstaDeepAI/nequip-organics", filename="nequip_organics_01.zip", local_dir="")
hf_hub_download(repo_id="InstaDeepAI/SPICE2-curated", filename="SPICE2_curated.zip", local_dir="")

Note that the pre-trained models are released on a different license than this library, please refer to the model cards of the relevant HuggingFace repos.

🚀 Inference time benchmarks

In order to showcase the runtime efficiency, we conducted benchmarks across all three models on two different systems: Chignolin (1UAO, 138 atoms) and Alpha-bungarotoxin (1ABT, 1205 atoms), both run for 1 ns of MD simulation on a H100 NVIDIA GPU. All model implementations are our own, including the Torch + ASE benchmarks, and should not be considered representative of the performance of the code developed by the original authors of the methods. Further details can be found in our white paper (see below).

MACE (2,139,152 parameters):

Systems	JAX + JAX-MD	JAX + ASE	Torch + ASE
1UAO	6.3 ms/step	11.6 ms/step	44.2/step
1ABT	66.8 ms/step	99.5 ms/step	157.2/step

ViSNet (1,137,922 parameters):

Systems	JAX + JAX-MD	JAX + ASE	Torch + ASE
1UAO	2.9 ms/step	6.2 ms/step	33.8 ms/step
1ABT	25.4 ms/step	46.4 ms/step	101.6 ms/step

NequIP (1,327,792 parameters):

Systems	JAX + JAX-MD	JAX + ASE	Torch + ASE
1UAO	3.8 ms/step	8.5 ms/step	38.7 ms/step
1ABT	67.0 ms/step	105.7 ms/step	117.0 ms/step

🙏 Acknowledgments

We would like to acknowledge beta testers for this library: Isabel Wilkinson, Nick Venanzi, Hassan Sirelkhatim, Leon Wehrhan, Sebastien Boyer, Massimo Bortone, Scott Cameron, Louis Robinson, Tom Barrett, and Alex Laterre.

📚 Citing our work

We kindly request to cite our white paper when using this library:

C. Brunken, O. Peltre, H. Chomet, L. Walewski, M. McAuliffe, V. Heyraud, S. Attias, M. Maarand, Y. Khanfir, E. Toledo, F. Falcioni, M. Bluntzer, S. Acosta-Gutiérrez and J. Tilly, Machine Learning Interatomic Potentials: library for efficient training, model development and simulation of molecular systems, available on the arXiv.