Skip to content

Lightweight neural network library built from the ground up use torch GPU tensors to facilitate linear algebra.

Notifications You must be signed in to change notification settings

Carson-Fricke/pynet

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 

Repository files navigation

pynet

Pynet is a small neural network library built upon pytorch tensors intended to demonstrate backpropogation with various activation functions and gradient descent optimizers. It was written in the spirit of "built from scratch," and only uses pytorch tensors to speed up performance and allow for GPU parallelization.

Getting Started

Simply download the code and import the pynet module to get started.

Alternatively, download the code and run the inbuilt test file which trains a Dense neural network on MNIST digit classification.

Implementation

All of the basic mathematic tensor operations rely on the pytorch library, and could be easilly subbed out for numpy or for a handwritten implementation.

Forwards and backwards proppogation is implemented for various Layer types. For example, Dense is a typical Dense layer, and DenseResidual carries previous layers activations on to the next layer.

A wide variety of activation functions and gradient descent opimizers are also implemented.

About

Lightweight neural network library built from the ground up use torch GPU tensors to facilitate linear algebra.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages