https://github.com/anshumansinha3301/neural-network
MNIST Using Neural Network project from scratch (Using Numpy and Matplotlib)
https://github.com/anshumansinha3301/neural-network
anshumansinha3301 bitfusiondynamics matplotlib-python mnist-handwriting-recognition neural-networks numpy open-source python
Last synced: about 2 months ago
JSON representation
MNIST Using Neural Network project from scratch (Using Numpy and Matplotlib)
- Host: GitHub
- URL: https://github.com/anshumansinha3301/neural-network
- Owner: anshumansinha3301
- License: mit
- Created: 2025-01-03T13:23:55.000Z (about 2 months ago)
- Default Branch: main
- Last Pushed: 2025-01-03T13:24:46.000Z (about 2 months ago)
- Last Synced: 2025-01-03T14:35:46.730Z (about 2 months ago)
- Topics: anshumansinha3301, bitfusiondynamics, matplotlib-python, mnist-handwriting-recognition, neural-networks, numpy, open-source, python
- Homepage:
- Size: 116 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Neural-Network
This project is a Python-based implementation of a neural network built from scratch to classify handwritten digits,
likely using the MNIST dataset. It begins by loading the dataset and preprocessing it, which includes normalizing
pixel values by scaling them between 0 and 1. The dataset is split into training and development sets to evaluate the model's performance.
The neural network architecture consists of two layers. The first is a hidden layer that uses the
ReLU (Rectified Linear Unit) activation function to introduce non-linearity. The second is an output layer that
applies the softmax activation function to generate probabilities for multi-class classification.
The project includes forward propagation, where the inputs are passed through the layers, and the output is computed.
It also handles one-hot encoding for the target labels and uses basic optimization techniques to update model parameters.
The model's performance is evaluated on both the training and development datasets, showcasing an end-to-end approach
to building a neural network from scratch.