NumPy

NumPy foundations

Notebook (download/open locally): /kicsigeze/notebooks/02_numpy_foundations.ipynb

What you should get out of this

• Think in arrays and shapes (and stop writing Python loops)
• Master broadcasting and indexing (including boolean masks)
• Use einsum and basic linear algebra confidently
• Recognize common numerical issues (dtype, overflow, stability)

Exercises (quick list)

• Implement batch cosine similarity without loops
• Implement softmax in a numerically stable way
• Implement PCA with np.linalg.svd

Readings/videos: Reading / Videos