Skip to content
LearnPY
NumPyfunction reference

np.random.default_rng

Create a reproducible random number generator.

Signature

np.random.default_rng(seed=None)

What it does

default_rng returns a Generator with methods such as integers, normal, and permutation. Passing a seed recreates the same starting stream for repeatable experiments.

Parameters

seedint, array_like, or Noneoptional
Initial state for the generator. None uses fresh entropy.

Returns

Generator

Example

pythonnp.random.default_rng example
import numpy as np

rng = np.random.default_rng(7)
print(rng.integers(1, 7, size=5))

Watch out for

  • Re-seeding inside a loop can produce repeated blocks of values.
  • A seed makes a run repeatable, not cryptographically secure.