OOP concepts in Python

Here’s my preliminary (incomplete) attempt at going over some basic OOP concepts using ML object models (@jeremy had mentioned on twitter that OOP would be useful for PyTorch). Hoping to add more this, feel free to give me suggestions.

I primarily referred to the Learning Python book: Found this quite comprehensive, hands-on and a bit repetitive(which could be good or bad)


Here’s a good intro to python, with some basic practical exercises:

Automate the boring stuff (with python)

Here’s a more thorough python treatment, though some of the content is locked behind a paywall.

Learn python the hard way (free sample, pay for full course)


Agree, Automate the boring stuff… is really useful to get started, especially the Regex and File Manipulation parts. It introduces functions, no classes though.

1 Like

Thanks for making a start on this. Note that where you refer to __init__ you need to surround it in backticks in your Jupyter markdown. As it is at the moment, the underscores are causing it to format as bold, and are not being displayed.


A former online student of Part 1 & 2, @Jason, recommends this track in his 4 months of Deep Learning post on Medium.

It contains 20 courses, each around 3-4 hours long with a browser UI. Costs 29.95$ a month (1st course is free, no payment info required).
I started it Friday and completed the first 2 courses already, though I expect the difficulty to scale.


hii @EricPB , I have been also taken datacamp for 2 month. I think it’s very good for Machine learning and Data science purpose but not so good for deep learning. Also you can access all 87 courses.

1 Like

Quick update on the DataCamp “DS with Python track” linked above:

I’ve done the first 6 courses (approx 1 course of 3-4h per day) out of the 20 and I really like it.

When doing Fast.ai Part 1 & 2 online last year, this would have really helped me then so I’d recommend it to anyone looking for a crash-course on Python for Machine Learning (once you know the basics of Python).
Little theory but lots of code snippets to re-use in practice.

Here’s a few selection of the 3-hours courses:

  • Intro and intermediate Python for DS
  • Python DS Toolbox with Numpy and Pandas, 1 & 2
  • Cleaning Data in Python
  • Manipulating DataFrames in Pandas
  • Merging DataFrames in Pandas
  • Statistical Thinking in Python 1 & 2
  • ML with experts: School Budgets
  • Supervised Learning with Scikit-Learn
  • Unsupervised Learning with Python
  • Deep Learning with Python and Keras

Here’s some I’m not sure yet:

  • Relational databases in Python
  • Interactive DataViz with Bokeh
  • Network Analysis with Python

Each course has 4 sub-modules.
Each sub-module has 3 videos of 3 mins each, each video followed by 5 exercises via browser (no install, no data download, 1 hint + 1 solution links) so super-fast to run.

The instructors speak clearly, the exercises scale very well, little touches of reinforcement learning here and there.
Plus the possibility to do daily mini-quizzes on previous courses.
You can download the slides and the videos on your PC for later use.

One month subscription is 29.95$ and should be plenty enough to cover the 10-15 courses you want.

Great deal, thanks @Jason