This is very cool! These are small tutorial interactions plotting a function, running a regression, etc using various scientific frameworks in IPython (the kickass interactive interpreter format for python). What really cool about this is that you can browse the individual notebooks by tool (numpy for numeric processing scikit-learn or deep-learning for tensor-flow and others). When you click a .ipynb file in github it will render that notebook for you. This lays out a step by step view of the results from each command or group of commands. Definitely worth checking out.
I love that github reads the .ipynb files natively.
So I just began using Jupyter for my own note taking on an EC2 free tier server, really loving it. Granted, I don't really do any data science stuff, but the wiki markup + the ability to run python code in my notes is grand.
What I love about the evolution (having started with python myself) is, that it not only does that for python anymore, but for a plethora of other languages as well.
So it helps me for example develop ideas in python, as well as in R, what I am learning a little bit right now.
I have an impression that R is now lagging behind Python when it comes to deep learning. Definitely such tutorials are helping get more people on board of the data science train with python.
Python is a good language and so is R. I don't understand the sentiment that the language is "lagging" behind Python.
It is frustrating to read these opinions on the final death of R, but as a R user the language has real grown by leaps and bounds. Sure Python is a nice language that has packages for Data Science. I even started in Python but R has won me over for data science due to it being language specific. As you look at the growth of R in recent years and it over taking over Data Science I still run into these ideas that R is antiquated and "lagging behind" when it usually is Python and Pandas catching up in actual practice.
The parent specifically said for Deep Learning though.
As someone who switched from R to Python 18 months ago, I haven't kept up to date in R world. What is the equivalent of Theano (and the packages built on top of it) in R?
There are a few R packages for deep learning, although I wouldn't say that they compare with Theano yet. There's darch (package for deep architectures and Restricted-Bolzmann-Machines) [1], deepnet [2] which is incorporated into caret which is nice, and gputools [3] which has some of the GPU operations found in Theano.
The exact quote:
> IPython 3.x was the last monolithic release of IPython, containing the notebook server, qtconsole, etc. As of IPython 4.0, the language-agnostic parts of the project: the notebook format, message protocol, qtconsole, notebook web application, etc. have moved to new projects under the name Jupyter. IPython itself is focused on interactive Python, part of which is providing a Python kernel for Jupyter.
