A neural network manually implemented in Python2.7 to classify handwritten numerical digits from MNIST data
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
raw_data
.gitignore
README.md
neural_funcs.py
nn.py

README.md

neural-net

A neural network implemented in Python to classify handwritten numerical digits from MNIST data

To run the network, first clone this repo and ensure that numpy and pandas are installed on your machine. Then simply start an iPython session, import the nn module, and call fit() on an instantiated network:

import nn
clf = nn.NeuralNetwork()
clf.fit()

This will automatically use the MNIST dataset and report the results of evaluation on a test using default parameters.