computer-go-dataset
                                
                                 computer-go-dataset copied to clipboard
                                
                                    computer-go-dataset copied to clipboard
                            
                            
                            
                        datasets for computer go
Contents
- TYGEM
- Tom
- Foxwq
- Aya
- Professional
- AI
- AlphaGo
- ELF OpenGo
- FineArt
- PhoenixGo
- Zen
- CGI
- DolBaram
- Dancer
- Leela
- CNC
- The 1st World AI Go Open 2017
- Berry Genomics Cup 2018 World AI Weiqi Competition
- 2018 Tencent World AI WEIQI Competition
- The 2nd World AI Go Open 2018
- Berry Genomics Cup 2019 World AI Weiqi Competition
- 2019 China Securities Cup World AI WEIQI Open
- The 11th Computer Go UEC Cup
- 2020 World GO AI Championship
 
- CGOS
- LeelaZero
- KGS
- Minigo
- NNGS
- ELF OpenGo
- KataGo
1 TYGEM dataset
2005.11.02 - 2016.12.31
TYGEM "9D vs 9D" dataset (1,516,031 games).
1.1 Format(index)
- id
- TYGEM id
 
- date
- YYYY.MM.DD HH:MM
 
- white
- char[16]
 
- white english
- char[11]
 
- white nation
- 0 Korea
- 1 Japan
- 2 China
- 3 USA
- 4 Taiwan
- 5 China
 
- black
- char[16]
 
- black english
- char[11]
 
- black nation
- 0 Korea
- 1 Japan
- 2 China
- 3 USA
- 4 Taiwan
- 5 China
 
- result
- B+Resign
- W+Resign
- B+X.5
- W+X.5
- B+Time
- W+Time
- B+Offline
- W+Offline
 
- round
- 60 - 450
 
- byoyomi/times minutes
default:
- komi
- 6.5
 
- white rank
- 9d
 
- black rank
- 9d
 
- type
- ranked
 
- CA
- UTF-8
 
- GM
- 1
 
- FF
- 4
 
Due to the 25 MB file size limit, I have to split the ".index" file. You can merge it by yourself.
1.2 Format(kifu)
id\t;B[coord];W[coord];B[coord];W[coord]......
1.3 Convert to SGF
- 
- Thanks for Hiroshi Yamashita writing the converter tool.
 
- 
Converter.py - usage: python Converter.py kifu.index english_user_id kifu_folder save_folder
- example: python Converter.py kifu.index Lurk(P) Kifu Save
 
2 TOM dataset
2003.09.25 - 2011.12.28
TOM "9D vs 9D" dataset (50,956 games).
2.1 Format(index)
- id
- date
- white
- black
- komi
- result
- round
- byoyomi/times minutes
default:
- white rank
- 9d
 
- black rank
- 9d
 
- type
- ranked
 
- CA
- UTF-8
 
- GM
- 1
 
- FF
- 4
 
2.2 Format(kifu)
id\t;B[coord];W[coord];B[coord];W[coord]......
2.3 Convert to SGF
- Converter_Tom.py
- usage: python Converter_Tom.py Kifu.index user_id kifu_folder save_folder
- example: python Converter_Tom.py Kifu.index 930115 kifu save
 
3 Foxwq dataset
3.1 Github
3.2 9d vs 9d
2013.07.09 - 2019.10.17
166,184 games
4 Aya's selfplay games for training value network
19x19, 13x13, 9x9
5 Professional
1940.01.01 - 2017.01.09
73,522 games
5.1 Format(txt)
SGF\nSGF\nSGF\n...
5.2 SGF tags
- 
GM - 1
 
- 
FF - 4
 
- 
SZ - 19
 
- 
AP 
- 
CA - UTF-8
 
- 
GN 
- 
PW 
- 
WR (option) 
- 
PB 
- 
BR (option) 
- 
KM (option) - 0 <= komi <= 8
 
- 
RE - B+
- W+
- B+X
- W+X
- B+R
- W+R
- B+T
- W+T
 
- 
B/W/B/W ... 
6 AI dataset
- 
AlphaGo- AlphaGo Zero
- 83 games
 
- AlphaGo Ke
- AlphaGo vs Ke Jie
- 3 games
- 2017.05.23 - 2017.05.27
- 3W / 0L / 100%
 
- AlphaGo vs China Team
- 1 games
- 2017.05.26
- 1W / 0L / 100%
 
- AlphaGo + Lian Xiao vs AlphaGo + Gu Li
- 1 games
- 2017.05.26
 
- AlphaGo selfplay
- 55 games
 
 
- AlphaGo vs Ke Jie
- AlphaGo Master
- Master
- 60 games
- 2016.12.29 - 2017.01.04
- 60W / 0L / 100%
 
 
- Master
- AlphaGo Lee
- AlphaGo vs Lee Sedol
- 5 games
- 2016.03.09 - 2016.03.15
- 4W / 1L / 80%
 
- AlphaGo selfplay
- 3 games
 
 
- AlphaGo vs Lee Sedol
- AlphaGo Fan
- AlphaGo vs Fan Hui
- 5 games
- 2015.10.05 - 2015.10.09
- 5W / 0L / 100%
 
 
- AlphaGo vs Fan Hui
 
- AlphaGo Zero
- 
ELF OpenGo- https://github.com/pytorch/ELF/releases
 
- 
FineArt- fuheyuqi
- Fox Go Server
- 171 games
- 2017.02.20 - 2017.11.23
- 166W / 5L / 97.07%
 
- FineArt
- Fox Go Server
- 626 games
- 2016.11.01 - 2017.05.12
- 488W / 138L / 77.95%
 
- FineArt UEC
- 11 games
- 11W / 0L / 100%
 
- FineArt vs Gu Li / Tuo Jia Xi / Tan Xiao
- 1 games
- 1W / 0L / 100%
 
- Li Long
- Fox Go Server
- 903 games
- 2016.10.21 - 2017.09.12
- 795W / 108L / 88.03%
 
- Xing Tian
- Fox Go Server
- 45 games
- 2016.12.27 - 2017.01.02
- 38W / 7L / 84.44%
 
- Tian Xia Wu Gou
- Fox Go Server
- 204 games
- 2016.09.28 - 2016.10.31
- 159W / 45L / 77.94%
 
- Sweeper
- Fox Go Server
- 55 games
- 2016.09.02 - 2016.10.05
- 38W / 17L / 69.09%
 
- Tiger
- Fox Go Server
- 63 games
- 2016.07.28 - 2016.10.28
- 42W / 21L / 66.66%
 
 
- fuheyuqi
- 
PhoenixGo- BensonDarr
- Fox Go Server
- 666 games
- 2017.11.28 - 2018.04.24
- 641W / 24L / 96.39%
 
- Shi Kong Qi Dian
- Fox Go Server
- 100 games
- 2018.03.15 - 2018.03.20
- 100W / 0L / 100.00%
 
- Jin Mao Ce Shi
- Fox Go Server
- 2744 games
- 2018.03.20 - 2018.10.10
 
 
- BensonDarr
- 
Zen- DeepZenGo
- Fox Go Server
- 130 games
- 2016.11.19 - 2018.04.07
- 106W / 24L / 81.53%
 
- DeepZen(P)
- TYGEM
- 2806 games
- 2016.12.29 - 2017.07.13
- 2271W / 535L / 80.93%
 
- DeepZen(B)
- TYGEM
- 303 games
- 2017.07.13 - 2017.07.27
- 208W / 95L / 68.65%
 
- deepzengo
- WBaduk
- 455 games
- 2017.06.21 - 2017.08.28
- 437W / 18L / 96.04%
 
- Zen19L vs Godmoves
- KGS
- 3 games
- 2016.11.29 - 2016.12.01
- 0W / 3L / 0%
 
 
- DeepZenGo
- 
CGI- CGI(P)
- TYGEM
- 167 games
- 2017.07.03 - 2017.07.13
- 120W / 47L / 71.85%
 
- CGI(B)
- TYGEM
- 16 games
- 2017.07.13 - 2017.08.19
- 10W / 6L / 62.50%
 
- CGI
- Fox Go Server
- 193 games
- 2017.09.22 - 2017.11.20
- 161W / 32L / 83.41%
 
 
- CGI(P)
- 
DolBaram- DolBaram
- TYGEM
- 26 games
- 2017.05.08 - 2017.08.03
- 24W / 2L / 92.31%
 
- dolBaram9
- Fox Go Server
- 62 games
- 2017.11.02 - 2018.02.16
- 50W / 12L / 80.64%
 
 
- DolBaram
- 
Dancer- Dancer(P)
- TYGEM
- 171 games
- 2017.04.02 - 2017.08.27
- 135W / 36L / 78.95%
 
- Dancer(P)
- Fox Go Server
- 220 games
- 2017.12.11 - 2018.04.04
- 193W / 27L / 87.72%
 
 
- Dancer(P)
- 
Leela- Leela(P)
- TYGEM
- 1851 games
- 2017.03.29 - 2017.07.13
- 787W / 1064L / 42.52%
 
- Leela(B)
- TYGEM
- 481 games
- 2017.07.13 - 2017.08.11
- 194W / 287L / 40.33%
 
 
- Leela(P)
- 
CNC- cnculture(P)
- TYGEM
- 1237 games
- 2017.04.22 - 2017.07.13
- 519W / 718L / 41.96%
 
- cnculture(B)
- TYGEM
- 662 games
- 2017.07.13 - 2017.08.11
- 304W / 358L / 45.92%
 
 
- cnculture(P)
- 
The 1st World AI Go Open 2017 - AI: Abacus,AQ,CGI,DeepZenGo,Dolbaram,FineArt,Golois,Leela,MuGo,OracleWQ,Rayn,TianRang
- 37 games + 3 special games
- 2017.08.16 - 2017.08.18
- 1st: DeepZenGo
- 2nd: CGI
- 3rd: FineArt
 
- AI: 
- 
Berry Genomics Cup 2018 World AI Weiqi Competition - AI: BADUKi,Dolbaram,FineArt,Golois,HEROZ Kishin,Leela Zero,Phoenix Go,TS Go
- 26 games
- 2018.04.27 - 2018.04.28
- 1st: Phoenix Go
- 2nd: FineArt
- 3rd: Leela Zero
 
- AI: 
- 
2018 Tencent World AI WEIQI Competition - 2018.06.24, Preliminary Contest: AQ,Aya,BADUKi,Dolbaram,ELFOpenGo,FineArt,Golaxy,LeelaZero,Northern Lights,Octopus,Raynz
- 2018.07.09, Repecharge: FineArt,Golaxy,ELFOpenGo,AQ,LeelaZero,Octopus,BADUKi,Dolbaram
- 2018.07.29, Semi-final: FineArt,Golaxy,ELFOpenGo,AQ
- 2018.07.30, Final: FineArt,Golaxy
- 1st: FineArt
- 2nd: Golaxy
 
- 2018.06.24, Preliminary Contest: 
- 
The 2nd World AI Go Open 2018 - AI: AQ,BADUKi,Dolbaram,ELFOpenGo,Golaxy,Raynz,Octopus,Og-go
- 31 games
- 2018.08.13 - 2018.08.15
- 1st: Golaxy
- 2nd: AQ
- 3rd: Octopus
 
- AI: 
- 
Berry Genomics Cup 2019 World AI Weiqi Competition - AI: BADUKi,Dolbaram,Go Genius,Golaxy,Golois,Leela Zero,Raynz,YXT
- 28 games
- 2019.04.26 - 2019.04.28
- 1st: Golaxy
- 2nd: BADUKi
- 3rd: Leela Zero
 
- AI: 
- 
2019 China Securities Cup World AI WEIQI Open - AI: CGI,ClearStone,Dolbaram,FineArt,Globis-AQZ,Go Genius,Golaxy,HanDol,Handtalk,LeelaZero,MayaGo,Octopus,YaoGo,YiXiaoTian
- 51 games
- 2019.08.22 - 2019.08.25
- 1st: FineArt
- 2nd: Golaxy
- 3rd: HanDol
- 4th: LeelaZero
 
- AI: 
- 
The 11th Computer Go UEC Cup - AI: Akira,BaduGI,BSK,EsArgo,GLOBIS-AQZ,Go Genius,Golaxy,Katsunari,Kifuwarabe,Kugutsu,Maru,masacts,mayoigo,Natsukaze,nlp,QuinoaIgo,Ray,Rn
- 93 games
- 2019.12.14 - 2019.12.15
- 1st: Golaxy
- 2nd: GLOBIS-AQZ
- 3rd: BaduGI
 
- AI: 
- 
2020 World GO AI Championship - 1st: Golaxy
- 4th: KataGo 40b384f with training
 
- 1st: 
7 CGOS dataset
8 Leela Zero dataset
9 KGS dataset
10 Minigo dataset
- CloudyGo
- Github
- Youtube
- cloud.google.com
- 
Under v?-9x9 directory you can find - data/ Tensorflow tf.Record files corresponding to raw training data
- models/ The saved model files for each generation
- sgf/ Zip archives of all games in .sgf form, grouped by generation
 
- 
Under v?-19x19 directory you can find - models/ The saved model files for each generation
- sgf/ Zip archives of all games in .sgf form, grouped by generation
 
 
- 
11 NNGS dataset
- 435,495 SGF files
- 1995.07 - 2005.05
12 ELF OpenGo dataset
- Home Page
- v2-training-run
- 19,690,954 SGF files
 
- Analysis
13 KataGo dataset
SGF
Licensing
The repository is licensed under GPL v3. License is available here.