Welcome BDI 2018 attendees! If you're here at my talk, Machine Learning from Scratch, then you've come to the right place. I've prepared a jupyter notebook with some starter code to follow along with. If you already have Anaconda installed on your computer, please feel free to download or clone the notebook here, and run it on your own machine. If not, there are two simple ways to access and execute the code straight from your browser.
-
Head to colab.research.google.com. (open this link in a new tab, then proceed)
-
Once the notebook loads, in the top left corner, click
file, chooseupload notebook
-
In the panel across the top, click
GITHUB, then paste the following link in the search bar:https://github.com/collinprather/BDI-2018-JupyterHub/blob/master/SVM_scratch.ipynb
-
Click on the
SVM_scratch.ipynbnotebook, and wait for you notebook to load. You should be all set to start executing your code!
Binder is powered by Docker and JupyterHub. It allows you to run code in a notebook without any installation whatsoever. If you're willing to wait about 1-3 minutes for it to load, it gives you a true Jupyter feeling notebook (Google's colab just feels a bit different.)
The JupyterHub for our repo will be launched by clicking the following "badge":
If you are interested in the notebooks where I performed my data preparation, they can be found here. The SVM_scratch.ipynb notebook that has the full stochastic gradient descent algorithm can be found here.