Skip to content

Latest commit

 

History

History
63 lines (40 loc) · 1.64 KB

README.md

File metadata and controls

63 lines (40 loc) · 1.64 KB
Raw

Hopfield network

Hopfield network implementations using numpy

It uses numpy for matrix multiplication. Here you can't find usage of Deep Learning framework. So this code can be useful for you, if you want to learn Hopfield networks from scratch. Supports only CPU computing.

This code is highly inspired by this Hopfield-Network project. takyamamoto, thanks a lot!

Getting Started

To download project:

git clone https://github.com/Vadbeg/hopfield_network.git

Installing

To install all libraries you need, print in hopfield_network directory:

pip install -r requirements.txt

It will install all essential libraries

Config

To use project with start_training.py or build_plots.pt you need to setup config. Config is located in config.py Config class. Example:

class Config:
    """Config class"""

    num_iter = 40
    threshold = 50

    image_size = (64, 64)

    asynchronous = True

Usage

After libraries installation you are free to use this project. It is great to start from start_training.py script. You can change data folders in this script and start training. Script will show you plots with training result.

Also you can use build_plots.py script. It will show you different plots with insights about how Hopfield network works.

Built With

  • numpy - The math framework used.

Authors