SeqNN 0.0.3

A sequential neural network python extension

SeqNN

A simple sequential neural network python extension

Table of Contents

• About the Project
  • Built With
• Getting Started
  • Installation
• Usage
• Contributing
• License
• Contact

About The Project

SeqNN is a simple single threaded sequential neural network python extension written in c++ and compiled with SWIG.

Features

• Modular layers
  • 2 Dimensional Convolutional
  • 2 Dimensional Max/Min Pool
  • Fully Connected Dense Layer
• Regularization
  • Weight decay
  • Soft weight sharing for the 2D Convolutional layer
  • Early Stopping

Built With

• SWIG

Getting Started

Installation

1. Clone the repo

git clone https://github.com/BrettCleary/DigitCNN

OR

2. Install with PyPi

pip install SeqNN

Usage

Example usage with a subset of the MNIST dataset is given in DigitCNN.py. Here is a sample output:

Error Rate (%) after training  2  number of epochs is  22.8
Error Rate (%) after training  4  number of epochs is  14.4
Error Rate (%) after training  6  number of epochs is  12.8
Error Rate (%) after training  8  number of epochs is  12.0
Error Rate (%) after training  10  number of epochs is  10.8
Error Rate (%) after training  12  number of epochs is  10.8
Error Rate (%) after training  14  number of epochs is  10.0
Error Rate (%) after training  16  number of epochs is  10.8
Error Rate (%) after training  18  number of epochs is  10.4
Error Rate (%) after training  20  number of epochs is  9.6

The error rate for test dataset is  6.4

Contributing

Any contributions you make are greatly appreciated.

1. Fork the Project
2. Create your Feature Branch (git checkout -b feature/AmazingFeature)
3. Commit your Changes (git commit -m 'Add some AmazingFeature')
4. Push to the Branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

License

Distributed under the MIT License. See LICENSE for more information.

Contact

Project Link: https://github.com/BrettCleary/SeqNN