Lesson 4
Random Arrays
Random module:
rng = np.random.default_rng(42) rng.random((3,)) # 3 floats rng.integers(1, 7, 5) # 5 ints 1-6 rng.normal(0, 1, (2,3)) # normal dist
A seed makes results reproducible.
Why this matters
Train/validation splits, simulations, and initialized weights all need randomness—but reproducible randomness when you debug or share results. The Generator API is the modern, explicit way to do that.
Without a seed, "it worked yesterday" is often just a different random draw.
ResourcesDocs, references & more — opens in a new tab
🎯 Your Task
Seed 0. Generate rolls — 10 ints from 1 to 6.
rng = np.random.default_rng(0), rng.integers(1, 7, 10)
⌘⏎ run · ⌘← → nav
▶ Output
Run your code to see output here.