/Convolutional_Neural_Network-for-Handwritten-Digit-Predection

Neural Network on MNIST Dataset: Implementation of a single-layer Neural Network using TensorFlow's Keras library. Train on MNIST, check predictions interactively with Drawingboard.py.

Primary LanguageJupyter Notebook

Convolutional Neural Network Implementation with MNIST Dataset

This repository demonstrates the implementation of a Neural Network with a single layer using TensorFlow's Keras library. The model is trained on the MNIST dataset, consisting of 70,000 images. The dataset is split into two parts: the training set (x_train, y_train) and the test set (x_test, y_test).

Files and Structure

  • Model.ipynb:

    • Jupyter notebook containing the implementation of the Neural Network.
    • Uses TensorFlow's Keras library for building and training the model.
    • Converts 2D image arrays to 1D for compatibility with the Neural Network.
    • Includes a run method for easy execution.

  • Drawingboard.py:

    • Python script to check predictions after running the model.
    • Allows users to interactively draw digits on a drawing board and receive predictions from the trained model.

How to Use

  1. Clone this repository:

    git clone https://github.com/your-username/neural-network-mnist.git
cd neural-network-mnist

  2. Open and run the Model.ipynb notebook to train the Neural Network and make predictions.

  3. After running the notebook, execute the Drawingboard.py script to interactively draw digits and check the model's predictions.

    python Drawingboard.py

Note

  • The Model.ipynb notebook includes a run method that guides you through the training process. Follow the instructions in the notebook for seamless execution.

  • The license and dependencies sections have been removed from this README for simplicity.

Contribution

Contributions and suggestions are welcome! Feel free to open issues or submit pull requests to enhance the functionality or documentation of this project.

Happy coding! 😊🖌️🤖