christopher commited on
Commit
a004998
1 Parent(s): 40a153d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +38 -1
README.md CHANGED
@@ -36,4 +36,41 @@ license: cc0-1.0
36
  pretty_name: Lichess puzzles
37
  tags:
38
  - chess
39
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
36
  pretty_name: Lichess puzzles
37
  tags:
38
  - chess
39
+ ---
40
+ # Dataset Card for Lichess Puzzles
41
+
42
+ <!-- Provide a quick summary of the dataset. -->
43
+
44
+ 4,062,423 chess puzzles, rated and tagged. See them in action on [Lichess](https://lichess.org/training/themes).
45
+
46
+ ## Dataset Details
47
+
48
+ ```python
49
+ Dataset({
50
+ features: ['PuzzleId', 'FEN', 'Moves', 'Rating', 'RatingDeviation', 'Popularity', 'NbPlays', 'Themes', 'GameUrl', 'OpeningTags'],
51
+ num_rows: 4062423
52
+ })
53
+ ```
54
+
55
+ ```json
56
+ {'PuzzleId': '00008',
57
+ 'FEN': 'r6k/pp2r2p/4Rp1Q/3p4/8/1N1P2R1/PqP2bPP/7K b - - 0 24',
58
+ 'Moves': 'f2g3 e6e7 b2b1 b3c1 b1c1 h6c1',
59
+ 'Rating': 1931,
60
+ 'RatingDeviation': 76,
61
+ 'Popularity': 94,
62
+ 'NbPlays': 6319,
63
+ 'Themes': 'crushing hangingPiece long middlegame',
64
+ 'GameUrl': 'https://lichess.org/787zsVup/black#47',
65
+ 'OpeningTags': None}
66
+ ```
67
+
68
+ ### Dataset Creation
69
+
70
+ Generating these chess puzzles took more than 50 years of CPU time. We went through 300,000,000 analyzed games from the Lichess database, and re-analyzed interesting positions with Stockfish 12/13/14/15 NNUE at 40 meganodes. The resulting puzzles were then [automatically tagged](https://github.com/ornicar/lichess-puzzler/tree/master/tagger). To determine the rating, each attempt to solve is considered as a Glicko2 rated game between the player and the puzzle. Finally, player votes refine the tags and define popularity.
71
+
72
+ ## Dataset Structure
73
+
74
+ <!-- This section provides a description of the dataset fields, and additional information about the dataset structure such as criteria used to create the splits, relationships between data points, etc. -->
75
+
76
+ ### Dataset