TRM: Tiny Recursive Model

This is a project aimed to create a lightweight TRM that can solve sudoku puzzles using logical reasoning techniques. The model is designed to be efficient and easy to use, making it suitable for educational purposes and small-scale applications.

To learn more about how this project is setup and the model works, see the documentation.

Features

• Lightweight: The model is designed to be small and efficient, making it easy to run on various devices.
• Logical Reasoning: Utilizes logical reasoning techniques to solve sudoku puzzles.
• Easy to Use: Simple interface for users to input puzzles and receive solutions.
• Educational: A great tool for learning about logical reasoning and problem-solving techniques.

Installation

To install the Tiny Recursive Model, clone the repository and install the required dependencies:

git clone https://github.com/s4nj1th/tiny-recursive-model.git
cd tiny-recursive-model
pip install -r requirements.txt

Usage

0. Prepare your environment by ensuring you have Python 3.7+ and the required libraries installed.

   • Install dependencies using:

     pip install -r requirements.txt

   • Ensure you have a compatible GPU if you plan to train the model.
   • Store the dataset of sudoku puzzles in the data/ directory (as sudoku.csv).

1. Run the tiny-recursive-model.ipynb notebook.

   • This trains and tests the TRM on sudoku puzzles.
   • The models are saved at checkpoints/.

2. Run the main.py script to input a sudoku puzzle and get the solution.

   • Input puzzles can be provided in a text file format.

   • Example command:

     python main.py --input puzzle.txt --output solution.txt

     Note: The input puzzle should be in a 9x9 grid format, with empty cells represented by zeros.

License

This project is licensed under the MIT License. See the LICENSE file for details.