[IPython-User] Fwd: WIP: theano ipython tutorials

James Bergstra james.bergstra@gmail....
Sat Dec 22 10:01:57 CST 2012


I finally uploaded some theano tutorials in ipython that I'd been
sitting on for a while (see below if you're interested!) but this
brought up two quick questions about IPython:

1. what's the best way to run all the notebooks from top to bottom to
make sure that there is no stale code?

2. what's the best way to publish the notebooks as images or a web
page or something so that they can have a more web-friendly image?
Right now, you have to download them and fire them up with IPython to
get the *html* that finally lets you see them.

- James

On Sat, Dec 22, 2012 at 10:39 AM, James Bergstra
<james.bergstra@gmail.com> wrote:
> I think the IPython notebook is a brilliant way to teach theano and
> machine learning in Python.  I've been working off and on toward a set
> of tutorials similar to theano tutorial + deep learning tutorials, and
> it's at a point where I'd like to share it, and maybe get some
> feedback & contributions.
>
> The code is here:
> https://github.com/jaberg/IPythonTheanoTutorials
>
> Currently the notebooks are:
>
> Intro to Python.ipynb
> Intro to Scikit Data (skdata).ipynb
> Links to Related Work.ipynb
> Model - Autoencoders and Variations with PyAutodiff.ipynb
> Model - Convnet with PyAutodiff.ipynb
> Model - LIF Neurons with Theano.ipynb
> Model - Linear SVM with PyAutodiff.ipynb
> Model - Logistic Regression with Theano.ipynb
> Model - Multilayer Perceptron with PyAutodiff.ipynb
> Notation for Machine Learning.ipynb
> Preprocessing - Image Whitening.ipynb
> Theano Tutorial (Part 1 - Introduction).ipynb
> Theano Tutorial (Part 2 - Simple Computation).ipynb
> Theano Tutorial (Part 3 - Functions and Shared Variables).ipynb
> Theano Tutorial (Part 4 - Random Variables).ipynb
> util.py
>
> Note that pure python support files (e.g. util.py) can be included
> alongside the notebook files, so that the notebooks can be more
> focussed on just the important parts of the code.
>
> There are a few known issues in this set of tutorials of various
> kinds, but I just decided it was silly to be sitting on it and hiding
> all the good parts as well, so here it is! Here's a home-made
> christmas present, theano ppl :)
>
> - James


More information about the IPython-User mailing list