My experiments with πŸ”₯

Practical Mojo for ML

This repo is inspired by fastai’s approach but does everything slowly, in Mojo πŸ”₯.

Dependencies

We will use only the packages available in the Mojo Playground. Those are:

  • The Python standard library
  • matplotlib
  • numpy

Content

Experiments with Mojo: 1. Read in the MNIST dataset from numpy
2. Define a linear layer
3. Backprop from scratch
4. Iterators