CondeSoulrack/chessy-v1

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:0.5BQuant:BF16Ctx Length:32kPublished:Mar 4, 2026Architecture:Transformer Warm

CondeSoulrack/chessy-v1 is a 0.5 billion parameter model fine-tuned from Qwen2.5-0.5B by Carlos Conde. This model specializes in chess move prediction, taking a board position in FEN notation and outputting the best move in UCI format. It was trained on 3 million unique Stockfish-evaluated positions, making it highly optimized for chess-specific applications.

Loading preview...

Overview

CondeSoulrack/chessy-v1 is a specialized language model developed by Carlos Conde, fine-tuned from the Qwen2.5-0.5B architecture. Its primary function is to predict optimal chess moves given a board state.

Key Capabilities

  • Chess Move Prediction: The model excels at predicting the best chess move in UCI format when provided with a board position in FEN notation.
  • Specialized Training: It was fine-tuned using a dataset of 3 million unique Stockfish-evaluated positions sourced from Lichess/chess-position-evaluations, ensuring high accuracy for chess-related tasks.
  • Input/Output Format: The model processes input in the format FEN: <fen> MOVE: <uci_move>.

Good For

  • Chess Engines and Bots: Integrating intelligent move prediction into chess applications.
  • Chess Analysis Tools: Providing suggested moves or evaluating positions.
  • Educational Chess Software: Assisting players with move choices and understanding optimal play.