README.md

metadata

license: mit
datasets:
  - Xmaster6y/stockfish-debug
name: Xmaster6y/gpt2-stockfish-debug
results:
  - task: train
    metrics:
      - name: train-loss
        type: loss
        value: 0.151
        verified: false
      - name: eval-loss
        type: loss
        value: 0.138
        verified: false

Model Card for gpt2-stockfish-debug

Details

Simple completion tuning using an equivalent of:

{"prompt":"FEN: {fen}\nMOVE:", "completion": " {move}"}