File size: 1,396 Bytes
6df111c
 
 
 
 
 
 
a2a61ab
0b85532
27bb612
6df111c
ade7c64
ccf4c15
ade7c64
ccf4c15
ade7c64
ccf4c15
ade7c64
 
 
 
 
 
 
 
7414592
ade7c64
a2a61ab
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
---
license: apache-2.0
datasets:
- nimashoghi/mptrj
pipeline_tag: graph-ml
tags:
- chemistry
- diffusers
- anemoi
- medvae
---
# Position-based Equivariant Graph Neural Network (`pos-egnn`)
This repository contains PyTorch model for loading and performing inference using the `pos-egnn`, a foundation model for Chemistry and Materials.

**GitHub**: https://github.com/IBM/materials/tree/main/models/pos_egnn

**HuggingFace**: https://huggingface.co/ibm-research/materials.pos-egnn

<p align="center">
    <img src="posegnn.svg">
</p>

## Introduction
We present `pos-egnn`, a Position-based Equivariant Graph Neural Network foundation model for Chemistry and Materials. The model was pre-trained on 1.4M samples (i.e., 90%) from the Materials Project Trajectory (MPtrj) dataset to predict energies, forces and stress. `pos-egnn` can be used as a machine-learning potential, as a feature extractor, or can be fine-tuned for specific downstream tasks.

Besides the model weigths `pos-egnn.v1-6M.pt` (download from [HuggingFace](https://huggingface.co/ibm-research/materials.pos-egnn)), we also provide an `example.ipynb` notebook (download from [GitHub](https://github.com/ibm/materials)), which demonstrates how to perform inference, feature extraction and molecular dynamics simulation with the model.

For more information, please reach out to [email protected] and/or [email protected]