Demo on Convolutional Neural Networks using Keras/Tensorflow on the MNIST dataset for HackBU on 11/9/17
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
misc
notebooks
LICENSE
README.md
help.txt
model.yaml
model_weights.h5
requirements.txt
test-model.py
train-cnn.py

README.md

Convolutional Neural Networks Demo

This repository is for my introductory demo on Convolutional Neural Networks - with MatPlotlib for visualization.

Setting Up

I'll be assuming Anaconda (Python 3.6) is installed for this demo, but I provided a table with the pip equivalents just in case. I can't ensure everything will work as instructed if not using Anaconda due to everyone having their own setups on their local machines.

Anaconda3 pip
conda create --name cnn --file requirements.txt pip install -r requirements.txt
source activate cnn ---

Writing Code

I'll be using Jupyter Notebook to demonstrate this code step-by-step

  • jupyter notebook should come installed with Anaconda3.

To start Jupyter Notebook, either start the application on Windows, or in the terminal on mac/linux just run the following command:

jupyter notebook