Public Snippets
Python implementation of a radial basis function network. The basis functions are (unnormalized) gaussians, the output layer is linear and the weights are learned by a simple pseudo-inverse.
Run this script in the folder where the .png images are stored. mencoder will create a avi video with the given framerate (fps).
This implementation of a standard feed forward network (FNN) is short and efficient, using numpy's array multiplications for fast forward and backward passes. The source code comes with a little example, where the network learns the XOR problem.
