Our MGT model is configured in configs/is2re/10k/MGT/gmt.yml and implemented in ocpmodels/models/GMTNet. One of them, MGT.py, holds the main code of the MGT model. The OC20-Ni dataset is selected from the OC20 dataset in the open-source OCP project and stored in this link(https://doi.org/10.5281/zenodo.12200723).
ocp is the Open Catalyst Project's
library of state-of-the-art machine learning algorithms for catalysis.
It provides training and evaluation code for tasks and models that take arbitrary chemical structures as input to predict energies / forces / positions, and can be used as a base scaffold for research projects. For an overview of tasks, data, and metrics, please read our papers:
Projects developed on ocp:
- CGCNN [
arXiv] [code] - SchNet [
arXiv] [code] - DimeNet [
arXiv] [code] - ForceNet [
arXiv] [code] - DimeNet++ [
arXiv] [code] - SpinConv [
arXiv] [code] - GemNet-dT [
arXiv] [code] - PaiNN [
arXiv] [code] - Graph Parallelism [
arXiv] [code] - GemNet-OC [
arXiv] [code] - SCN [
arXiv] [code] - eSCN [
arXiv] [code] - EquiformerV2 [
arXiv] [code]
See installation instructions.
Dataset download links and instructions are in DATASET.md.
A detailed description of how to train and evaluate models, run ML-based relaxations, and generate EvalAI submission files can be found in TRAIN.md.
Our evaluation server is hosted on EvalAI. Numbers (in papers, etc.) should be reported from the evaluation server.
Interactive tutorial notebooks can be found here to get familiar with various components of the codebase.
We provide several pretrained model weights for download here.
For all non-codebase related questions and to keep up-to-date with the latest OCP announcements, please join the discussion board.
All code-related questions and issues should be posted directly on our issues page.
- This codebase was initially forked from CGCNN by Tian Xie, but has undergone significant changes since.
- A lot of engineering ideas have been borrowed from github.com/facebookresearch/mmf.
- The DimeNet++ implementation is based on the author's Tensorflow implementation and the DimeNet implementation in Pytorch Geometric.
ocp is released under the MIT license.
If you use this codebase in your work, please consider citing:
@article{ocp_dataset,
author = {Chanussot*, Lowik and Das*, Abhishek and Goyal*, Siddharth and Lavril*, Thibaut and Shuaibi*, Muhammed and Riviere, Morgane and Tran, Kevin and Heras-Domingo, Javier and Ho, Caleb and Hu, Weihua and Palizhati, Aini and Sriram, Anuroop and Wood, Brandon and Yoon, Junwoong and Parikh, Devi and Zitnick, C. Lawrence and Ulissi, Zachary},
title = {Open Catalyst 2020 (OC20) Dataset and Community Challenges},
journal = {ACS Catalysis},
year = {2021},
doi = {10.1021/acscatal.0c04525},
}