MNIST

Classification of hand written digits using a simple neural network and a shallow convolutional neural network for result comparison using TensorFlow 2 and Keras.

Tools

1. Visual Studio Code
2. Python 3.8
3. TensorFlow 2.x (Keras is built-in)
4. Matplotlib
5. Jupyter Notebook

Results (10 epochs)

Model	Training Accuracy	Validation Accuracy	Training Loss	Validation Loss
Artificial Neural Network	98.52%	98.00%	0.0445	0.0706
Convolutional Neural Network	99.65%	98.78%	0.0104	0.0470

For better visualizations please look at the code