BlueSunflower/gpt2-medium-chess

Model description

GPT-2 medium finetuned on 8 million chess games (short algebraic notation)

Data: Chess DB + sample from lichess + sample from CCRL

Example context: "1-0 2700 1350 1.e4 e5 2.Nf3 Nc6" (white score-black score whiteelo blackelo moves)

Model results

• —ELO (measured against Stockfish) ~ 1340
• —% legal moves 98.5%
• —checkmates in one move (from BigBench benchmark) - 46.5%

license: agpl-3.0 ---