This repository contains DeepMind's entry to the PCQM4M-LSC (quantum chemistry) and MAG240M-LSC (academic graph) tracks of the OGB Large-Scale Challenge (OGB-LSC).
For full details regarding this entry, please see our technical report.
pcq/
: Scripts for training, evaluating on the PCQ dataset.mag/
: Scripts for training, evaluating on the MAG dataset.
To cite this work:
@article{deepmind2021ogb,
author = {Ravichandra Addanki and Peter Battaglia and David Budden and Andreea
Deac and Jonathan Godwin and Thomas Keck and Wai Lok Sibon Li and Alvaro
Sanchez-Gonzalez and Jacklynn Stott and Shantanu Thakoor and Petar
Veli\v{c}kovi\'{c}},
title = {Large-scale graph representation learning with very deep GNNs and
self-supervision},
year = {2021},
journal={arXiv preprint arXiv:2107.09422},
}