Simple Neural Net implementation, Taken from the MNIST Nueral Net example implementation.
UI is implemented in ReactJS
and server side API/Nueral Net is implemented in GO
Online DEMO can be accessed HERE
- Install Go and Node.js
- Clone Repo in Go
SRC
directory - Open console in cloned repo directory and perform these commands
$ npm install
$ npm run build // Builds React Bundle
$ go build // Build API server and web server
$ ./khs-neural-net // Executes Go Binary
- Open browser to http://localhost:8000
Network has been pre trained, but it can be retrained with the following commands
- Unzip mnist_dataset
- From a command shell execute the following commands
$ go build
$ ./khs-neural-net -mnist train
References used in this implementation
-
Modified Go implementation from https://github.com/sausheong/gonn/commits/master
-
YOLO and MNIST reference app https://github.com/SkalskiP/ILearnMachineLearning.js