SimpleNN is a lightweight and modular artificial neural network (ANN) framework built entirely using NumPy. It supports essential neural network components including dense layers, activation functions, loss functions, and backpropagation.

• Fully connected Dense layers
• Common activation functions: ReLU, Sigmoid, Tanh, Softmax
• Loss functions: MSE, Binary Cross-Entropy (BCE), Mean Absolute Error (MAE)
• Backpropagation and weight updates with learning rate
• Forward and backward pass implementation
• Model saving and loading with .npz format
• Custom training loop with logging