Welcome to the NumPy Neural Network Library! Have you ever desired a collection of machine learning algorithms implemented exclusively in NumPy, even if it's not the most efficient but still quite readable? This library is crafted to furnish you with a complete toolkit for constructing and training neural networks using NumPy, a robust numerical computing library for Python.

• Modular Design: The library is structured in a modular way, allowing users to easily construct, customize, and train various types of neural network architectures.
• Multiple Architectures: It includes implementations of popular neural network architectures such as Multi-Layer Perceptrons (MLP), Convolutional Neural Networks (CNN), Transformers, Variational Autoencoders (VAE), and Diffusion models etc.
• Flexible and Extensible: Users can easily extend the library by implementing custom layers, loss functions, and optimizers.
• Efficient NumPy Implementation: All computations are performed using NumPy arrays, ensuring efficient execution on CPU/GPU.

To install the library, simply clone this repository and install the required dependencies:

git clone https://github.com/CaiFeipeng/numpy-nn.git
cd numpy-nn
pip install -r requirements.txt