Implementations from scratch (if numpy counts as scratch as well as the occaisonal, vastly more efficient, TensorFlow script). They say you really get to know a model when you get more intimate with it, knowing it's pros/cons, when or how best to apply it and so on. That's the main motivation for this repo.
-
Clone the repo:
$ git clone https://github.com/eltonlaw/machine_learning.git
-
To satisfy dependencies, run the following:
$ virtualenv ~/machine_learning $ source ~/machine_learning/activate $ pip3 install -r requirements.txt
-
And for a quick sanity check:
$ python3 -m unittest discover
-
If you're having any trouble, submit an issue.