README.pcqm4mv2-2d.md

metadata

language: []
pretty_name: PCQM4Mv2 2D
config_name: pcqm4mv2-2d
dataset_size: 293908175
size_categories: n>1M
license: mit
license_link: https://github.com/snap-stanford/ogb/blob/master/LICENSE
tags:
  - graphs
  - molecules
  - chemistry
  - parquet
  - torch-geometric
  - ogb
task_categories:
  - graph-regression
homepage: https://ogb.stanford.edu/docs/lsc/pcqm4mv2/

pcqm4mv2-2d

OGB-LSC PCQM4Mv2 with 2D graph features derived from SMILES for HOMO-LUMO gap regression.

License: MIT License (OGB)

Splits

Split	Rows	File	Size (MB)
train	3,378,573	train.parquet	261.02
validation	73,545	valid.parquet	6.81
test-dev	147,034	test-dev.parquet	13.02
test-challenge	147,429	test-challenge.parquet	13.05

Features

• mol_id: int64 unique identifier per molecule
• x: list[int64[9]], shape (num_nodes, 9), atom feature vector
• edge_index: int64[2, num_edges], COO adjacency
• edge_attr: list[int64[3]], shape (num_edges, 3), bond features
• num_nodes: int64 number of atoms
• smiles: string canonical SMILES
• target: float32 HOMO-LUMO gap

Citation

@article{hu2021ogblsc,
  title={OGB-LSC: A Large-Scale Challenge for Machine Learning on Graphs},
  author={Hu, Weihua and Fey, Matthias and Ren, Hongyu and Nakata, Maho and Dong, Yuxiao and Leskovec, Jure},
  journal={arXiv preprint arXiv:2103.09430},
  year={2021}
}