graphormer-base-pcqm4mv1

Model Card for pcqm4mv1_graphormer_base

The Graphormer is a graph classification model.

Model Details

Model Description

The Graphormer is a graph Transformer model, pretrained on PCQM4M-LSC, and which got 1st place on the KDD CUP 2021 (quantum prediction track).

• Developed by: Microsoft
• Model type: Graphormer
• License: MIT

Model Sources

<!-- Provide the basic links for the model. -->

• Repository: Github
• Paper: Paper
• Documentation: Link

Uses

Direct Use

This model should be used for graph classification tasks or graph representation tasks; the most likely associated task is molecule modeling. It can either be used as such, or finetuned on downstream tasks.

Bias, Risks, and Limitations

The Graphormer model is ressource intensive for large graphs, and might lead to OOM errors.

How to Get Started with the Model

See the Graph Classification with Transformers tutorial.

Citation [optional]

<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->

BibTeX:

@article{DBLP:journals/corr/abs-2106-05234,
  author    = {Chengxuan Ying and
               Tianle Cai and
               Shengjie Luo and
               Shuxin Zheng and
               Guolin Ke and
               Di He and
               Yanming Shen and
               Tie{-}Yan Liu},
  title     = {Do Transformers Really Perform Bad for Graph Representation?},
  journal   = {CoRR},
  volume    = {abs/2106.05234},
  year      = {2021},
  url       = {https://arxiv.org/abs/2106.05234},
  eprinttype = {arXiv},
  eprint    = {2106.05234},
  timestamp = {Tue, 15 Jun 2021 16:35:15 +0200},
  biburl    = {https://dblp.org/rec/journals/corr/abs-2106-05234.bib},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}