NNet
NNet copied to clipboard
mczerny
Simple implementation of neural networks in python
NNet 0.1
This is a simple implementation of neural networks in Python. Use at your own risk, as this is very messy, has very little internal testing, and created in an ad-hoc manner to suit my needs as they came up.
Example
The model expects NxP numpy arrays as input (X) where N is the number of observations and P is the number of features. The target values (Y) can either be a numpy array where each column is a class for classification or a number for regression. ``` from NNet import NeuralNetnnet = NeuralNet(100, learn_rate=1e-2, penalty=1e-8) nnet.fit(X_train, Y_train, SGD=True, batch=100) nnet.score(X_test, Y_test)
More information on the methods can be found in the Wiki.
mczerny