Skip to content

Latest commit

 

History

History
61 lines (46 loc) · 2.76 KB

README.md

File metadata and controls

61 lines (46 loc) · 2.76 KB
Raw

The Professional Go Annotation Dataset

Teaser image

The ProfessionAl Go Annotation DatasEt (PAGE) is an extensively annotated Go (weiqi, baduk) dataset for large-scale data-driven analytics. PAGE contains 98,525 games played by 2,007 professional players from 1950 to 2021. Our proposed PAGE incorporates game-level metadata and in-game statistics. The game-level metadata pertinent to games, players, and tournaments is annotated by consolidating data from numerous reliable sources. The comprehensive in-game statistics are generated from KataGo.

The Professional Go Annotation Dataset
Yifan Gao, Danni Zhang, Haoyue Li
IEEE Transactions on Games, 2023

PGD: A Large-scale Professional Go Dataset for Data-driven Analytics
Yifan Gao
IEEE Conference on Games (CoG), 2022

News

2023-04-01

  • This project is still quickly updating. Check ToDo list to see what will be released next.

2023-05-11

  • Our paper "The Professional Go Annotation Dataset" describing this dataset has been accepted by IEEE Transactions on Games. 🎉

2023-07-11

  • Initial public release of the proposed PAGE.
  • Dataset is licensed under CC BY-NC-SA 4.0, which allows free use for non-commercial purposes only.
  • For questions and feedback, please open a Github issue on this repo.

ToDo

  • Add example scripts to conduct preliminary analysis on various features of PAGE.
  • Provide code for downstream tasks described in the journal paper.
  • Add scripts to generate in-game statistics from SGF game records.
  • Add games data after 2021 to expand the dataset coverage.
  • Extend KataGo analysis time per game to produce more robust in-game statistics.

Download

All data is hosted on OneDrive and Baidu NetDisk:

Baidu NetDisk：download link
code：grvo

OneDrive: download link
code：PAGE

Path Size Files Format Description
PAGE 59.6 GB 98,541 Main folder
ownership 28.5 GB 98,525 NumPy In-game statistics, with multiple NumPy matrixes.
metadata.csv 13.3 MB 1 CSV Game-level Metadata.
recommended_move.h5 5.49 GB 1 HDF5 In-game statistics.
ingame_statistics.h5 1.26 GB 1 HDF5 In-game statistics without ownership and recommended move.
zip 24.3 GB 13 ZIP Contents of ownership folder as a ZIP archive.

Baselines and examples

ToDo

Analysis your own data

ToDo

Related resources

ToDo

Acknowledgements

ToDo