If you want to learn about deep learning I highly recommend the excellent WildML series written by Denny Britz.
The notebooks in this project are variations on his 'Implementing a Neural Network from Scratch in Python' exploring expressing the same using Tensorflow, Keras, and skflow.
Note: I'm new to these frameworks, so there maybe better and more correct ways to express this than I came up with. Let me know.
Note: You will have to update requirements.txt and change the tensorflow dependency if you install on something else then Mac OS X, see here.
# Create a new virtual environment
virtualenv venv
source venv/bin/activate
# Install requirements
pip install -r requirements.txt
# Start the notebook server
KERAS_BACKEND=tensorflow jupyter notebook