Intro-notebooks

These notebooks aim to develop basic skills of ML. Pick and choose - depending on what your self-assessment of your skills.

You are encouraged to fork this repository and then clone your fork e.g.:

git clone git@github.com:<your github handle>/Intro-notebooks.git

Initial setup

Gives you a quick run through on how to set up your working environment that you will be using this week - this may also prove useful for your workflows in the future.

If you are a Beginner - it would probably serve you best to go through the notebooks below. If you already did some ML projects - pick and choose according to where you might have a gap in experience.

Jupyter tips and tricks.ipynb

If you plan to use jupyter you might find some useful tricks here.

tensors_broadcasting_graph_backpropagation.ipynb

In this notebook we explore how tensors are treated in pytorch - and how to efficiently structure your code through the application of broadcasting. We alo look at the computational graph and backpropagation.

deep_learning_building_blocks.ipynb

Here we get closer to actual deep learning practice - how to set up most basic models and run a training loop. How would one implement basic layers in deep networks.

fundamentals_in_numpy.ipynb

This is only for the interested - how we might go about backprobagation 'by hand' using numpy only. In practice some ML library would always be used.