Last active
January 4, 2016 00:39
-
-
Save Slater-Victoroff/8543588 to your computer and use it in GitHub Desktop.
Simplest Possible Geometric Restricted Boltzmann Machine. Doesn't include training, just random generation and prediction for now.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import numpy as np | |
def transfer_function(x, y): | |
return np.power(np.prod(x, axis=1)[:, None] * np.prod(y, axis=0), 1./x.shape[1]) | |
def gnn(c): | |
return normalize([np.random.random(c[i] * c[i + 1]).reshape((c[i], c[i + 1])) for i in range(len(c) - 1)]) | |
def predict(weights, input_vector, reverse=False): | |
current_net = [input_vector] + weights | |
if reverse: current_net = [input_vector] + [layer.T for layer in weights[::-1]] | |
return reduce(transfer_function, current_net) | |
def normalize(weights): | |
constants = [np.power(np.prod(w, axis=1), -1./w.shape[1]).reshape(w.shape[0],1) for w in weights] | |
return [np.multiply(w, constant) for w, constant in zip(weights, constants)] |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment