Module 12

NumPy

arrays, shapes, axes

When you finish this module you can say:

I can compute with numpy arrays by reasoning about shape, and I can train a model by hand with gradient descent.

By the end of this module you'll be able to train a linear model with gradient descent, using nothing but NumPy — no machine learning library at all.

Lessons

7 lessons · ≈ 69 min

1. Not a list, an array9 min
2. Shape: shape and reshape9 min
3. Element-wise ops and broadcasting10 min
4. Indexing and masks9 min
5. Matrix multiplication: @10 min
6. axis and argmax10 min
7. Project: Linear regression by hand12 min · project

Practice exercises

Short writing exercises, independent of the lessons: Go to the exercises

1. Sum of squares
2. Multiplication table
3. Exam summary
4. Weights stored the other way round
5. Exam leaders