中文Sign up

mondk/my-chess-engine-test

🤗 Hugging Face sourcemit35 MBother✓ 1 checksumupdated today
Have the files? Submit a model torrent

If you have the complete model files and permission to share them, replace the example folder path and run this command. It checks your files, creates a torrent, and submits the magnet and checksums to Pirate Face. Keep your torrent client seeding so others can download from peers. Pirate Face does not receive the model files. Get your community key from your account. Read the script first if you like.

curl -fsSL https://pirateface.co/package.sh | bash -s -- --repo mondk/my-chess-engine-test ./model-folder
Needs a seeder →

GUIDE FOR USING CUSTOM-TRAINED UCI ENGINE

  1. Install dependencies on the machine running the engine: pip install torch python-chess

  2. Extract this zip file, ensuring the following 2 files are in the SAME directory:

    • chess_model.pt
    • uci_engine.py

  3. Add the engine to any GUI software supporting external UCI engines, such as:

    • Arena Chess GUI
    • BanksiaGUI
    • CuteChess
    • Lucas Chess
    • En Croissant

    In the "Add Engine" / "Install Engine" section, point to the execution command:

python /path/to/uci_engine.py

(or python3, depending on your operating system)

  1. Playing as a Lichess Bot (official method, allows custom engines):
    • Register a dedicated Lichess account and upgrade it to a BOT account (see guide at: https://lichess.org/api#tag/Bot)
    • Use the "lichess-bot" repo (by lichess), pointing the engine path to uci_engine.py
    • The engine will automatically play against other bots on Lichess—matching the goal of "having it play against other bots like Stockfish does".

IMPORTANT FAIR PLAY NOTE:

  • Chess.com and Lichess DO NOT allow external engine usage in RATED GAMES against HUMAN PLAYERS. Violations will result in account bans.
  • Allowed use cases: offline practice/analysis (Play vs Computer / local GUI), or using an official BOT account to play against other bots.

ENGINE PERFORMANCE:

  • This is a small CNN model trained on a small dataset (random vs Stockfish
  • evaluated at 0.1s/move). Its playing strength is roughly equivalent to ~1000-1200 ELO,
  • and it is not as strong as full Stockfish. To make it stronger: generate more data,
  • increase Stockfish think time during data generation, increase epoch counts/network channels,
  • or integrate deeper search algorithms.