ConvNetJS a javascript toolkit for neural networks

ConvnetJs is a javascript library to train deeplearning models  on your browser. This library do not require additional software support nor compilers and GPUs.
   
         This library totally deals with solving neural networks in javascript.
It deals with these aspects says stanford site:

• Common Neural Network modules(fully connected layers, non-linearities)
• Classification (SVM/Softmax) and Regression (L2) cost functions
• Ability to specify and train Convolutional Networks that process images
• An experimental Reinforcement Learning module, based on Deep Q Learning.

   It is developed  by a Phd member of stanford named kaparthy.


There are two ways to use it

1. On a  browser
2. On a server using npm.
     To install in npm type

npm install convnetjs

     If you want to get started click here