This project focuses on digit recognition using TensorFlow and offers two different approaches for achieving high accuracy:
-
Using Dense Layers: Model is trained using dense layers, achieving a training accuracy of 98% and a test accuracy of 97%.
-
Using CNN Layers: Model is trained using Convolutional Neural Network (CNN) layers, achieving an impressive training accuracy of 99.5% and a test accuracy of 98.5%.
It aims to classify handwritten digits from the MNIST dataset.
The project is organized into two main folders:
Using CNN
: Contains code related to digit recognition using Convolutional Neural Networks.Using Dense
: Contains code related to digit recognition using Dense Neural Networks.
Follow the steps below to get started with this project:
-
Clone this repository by running
git clone https://github.com/yourusername/digit-recognition.git
-
Navigate to the appropriate folder (
CNN
orDense
) to explore the specific code.
- Python 3
- Keras
- TensorFlow
- NumPy
- Matplotlib