BrainJS

BrainJS is a neural network tool that allows you to visually create artificial neural networks. What makes this tool much cooler than any other tool out there, is that it is completely written in HTML5 and JavaScript.
I X
  • New Node Use this button to create new nodes. Nodes in different layers will automatically be (fully) connected.
  • Select Node Use this button to select nodes. Selected nodes can be deleted or show extra information.
  • Reset network This button will reset the network and all parameters to the original settings.
  • Settings By clicking this button the parameters of the network (e.g. learning rate and number of iterations during training) can be set.
  • Train Network This button is used to train the network. Input-output relations are specified and then used to train the network.
  • Single Run To use a trained network, this button opens an input window where input values can be specified. The network then calculates the associated output values.
  • Save Network This button lets you save a network so you can use it again later on.
  • Load Network Use this button to load a previously created (and saved) network.

Parameter settings that will during training of the neural network.

Are you sure you want to delete this node?

This action cannot be undone.

Are you sure you want to reset the network? This will remove all nodes and reset all parameters.

This action cannot be undone.

Please enter the training data. Each sample is separated by a new line and all values should be within the interval [0, 1]. An example of what the data looks like is as follows:

input: 0.9, 0.4, 0; output: 1.0, 0.0
input: 1.0, 1.0, 0; output: 0.2, 0.9

Note: use the Cancel button if you like to change the network; training data is automatically saved.

The textfield below displays the network in a text-format. Save this text to a file on your computer and use this text later on again to load the network.

Please insert the network below. Paste the text here from the file that you stored when saving the network.