Skip to content

Latest commit

 

History

History
38 lines (32 loc) · 1.5 KB

README.md

File metadata and controls

38 lines (32 loc) · 1.5 KB
Raw

🧠 TensorFlow MNIST Digit Classifier 🧠

A simple and efficient neural network model to classify handwritten digits using TensorFlow and the MNIST dataset (soon I will implement the image classification part, slow motion is better than no motion).

📌 Features

  • Simple Neural Network: Built using TensorFlow's Keras API for ease of use.
  • Data Normalization: Pre-processes the MNIST dataset for optimal training.
  • Dropout Layer: Helps prevent overfitting to improve model generalization.
  • Softmax Activation: Converts the model's output logits to probabilities.

🚀 Getting Started

  1. Clone the repo:
git clone [https://github.com/reecebaileyy/Image-Classification.git]
cd Image-Classification
  1. Install TensorFlow:
pip install tensorflow
  1. Run the script:
python image_classification.py

📈 Model Architecture

  • Input Layer: Flattens the 28x28 pixel images.
  • Hidden Layer: 128 neurons with ReLU activation.
  • Dropout Layer: 20% dropout rate.
  • Output Layer: 10 neurons for each digit (0-9).

📊 Model Performance

After training for 5 epochs, you can evaluate the model's performance on the test dataset. The accuracy metric will give you an idea of how well the model is performing.

🔜 Coming Soon

Custom Image Prediction: Soon, you'll be able to input your own handwritten digit images, and the script will predict the digit for you!

🤝 Contributing

Feel free to fork the repository, make changes, and submit pull requests. All contributions are welcome!