Commit
·
55ba0fc
1
Parent(s):
e06ac00
Update README.md
Browse files
README.md
CHANGED
@@ -1,5 +1,5 @@
|
|
1 |
|
2 |
-
Model description
|
3 |
|
4 |
|
5 |
GPT-2 medium finetuned on 8 million chess games (short algebraic notation)
|
@@ -9,12 +9,12 @@ Data: Chess DB + sample from lichess + sample from CCRL
|
|
9 |
Example context: "1-0 2700 1350 1.e4 e5 2.Nf3 Nc6" (white score-black score white_elo black_elo moves)
|
10 |
|
11 |
|
12 |
-
Model results
|
13 |
|
14 |
|
15 |
-
ELO (measured against Stockfish) ~ 1340
|
16 |
-
% legal moves 98.5%
|
17 |
-
checkmates in one move (from BigBench benchmark) - 46.5%
|
18 |
|
19 |
|
20 |
|
|
|
1 |
|
2 |
+
# Model description
|
3 |
|
4 |
|
5 |
GPT-2 medium finetuned on 8 million chess games (short algebraic notation)
|
|
|
9 |
Example context: "1-0 2700 1350 1.e4 e5 2.Nf3 Nc6" (white score-black score white_elo black_elo moves)
|
10 |
|
11 |
|
12 |
+
# Model results
|
13 |
|
14 |
|
15 |
+
- ELO (measured against Stockfish) ~ 1340
|
16 |
+
- % legal moves 98.5%
|
17 |
+
- checkmates in one move (from BigBench benchmark) - 46.5%
|
18 |
|
19 |
|
20 |
|