Skip to content

A clean, pure C++/CUDA implementation of Capsule Networks, no cuDNN, TF, Keras, or libraries.

Notifications You must be signed in to change notification settings

TennisGazelle/CUDA-CapsuleNetwork-Methods

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

53 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Neural Nets

By Daniel Lopez Computer Science Graduate Student at the University of Nevada, Reno.

Three implementations of Neural Nets, built on top of one another for CUDA implementation and further researach/development. Written in C++/CUDA for full control ove data management and moving (not implying TensorFlow or other projects are badly written, or don't allow low-level control of data)

Building/Running

The preferred IDE is CLion by JetBrains, but choose whatever floats your boat :)

Dependencies

This project is build with CMake and only uses the armadillo library as dependency. Instructions for Armadillo download installation may be found on their website.

Building

mkdir build
cd build
cmake ..

Running

cd build/
./NeuralNets

Nothing too special comes up yet, look at the file src/main.cpp to see what you're supposed to do right now.