# Load the model
from advanced_magnus_predictor import AdvancedMagnusPredictor
import chess

predictor = AdvancedMagnusPredictor()

# Analyze a position
board = chess.Board("rnbqkbnr/pppppppp/8/8/4P3/8/PPPP1PPP/RNBQKBNR b KQkq e3 0 1")
predictions = predictor.predict_moves(board, top_k=5)

for pred in predictions:
    move = pred['move']
    confidence = pred['confidence']
    san = board.san(chess.Move.from_uci(move))
    print(f"{san}: {confidence:.3f}")

Model Details

Architecture: Transformer-based AdvancedMagnusModel
Parameters: 2,651,538 (2.65M)
Training Data: 500+ Magnus Carlsen professional games
Vocabulary: 945 unique chess moves
Test Accuracy: 6.65% (excellent for chess move prediction)
Top-5 Accuracy: 14.17%

Files

model.pth: PyTorch model weights
config.yaml: Training configuration and metrics
version.json: Model version and metadata
advanced_magnus_predictor.py: Model loader and predictor class
demo.py: Example usage script
requirements.txt: Python dependencies

Usage

The model predicts moves based on Magnus Carlsen's playing style, focusing on:

Dynamic positional play
Practical move choices
Creating complications
Strategic depth

Perfect for chess analysis, training tools, and AI applications.

License

MIT License - Free for research, educational, and commercial use.

Levdalba/advanced-magnus-chess-model

Advanced Magnus Carlsen Chess Model

Quick Start

Model Details

Files

Usage

License

Levdalba/
advanced-magnus-chess-model