ZeroToNet is a neural network implementation built entirely from scratch, with no external dependencies. This repository contains the code for constructing and training neural networks.
ZeroToNet provides a comprehensive framework for creating neural networks from scratch. It includes modules as Dense architecture till now.
- Implementation of core neural network components (layers, activation functions, loss functions)
- Support for customizable network architectures and hyperparameters
- Modular design for easy extension and experimentation
To get started with ZeroToNet, simply clone this repository to your local machine: git clone https://github.com/khaaaleed-5/ZeroToNet.git
To use ZeroToNet, import the necessary modules into your Python script:
from ANN import NeuralNetwork, Dense, ActivationFunction, LossFunction
# Define your network architecture
network = NeuralNetwork()
network.add_layer(Dense(input_size=784, output_size=128,activation='relu'))
network.add_layer(Dense(input_size=128, output_size=10,activation='relu'))
network.add_layer(Dense(input_size=10, output_size=1,activation='sigmoid'))