Expand description
Random number generation, like np.random.
The engine uses xoshiro256** seeded by SplitMix64, so the same seed produces the same
sequence on every platform. The streams are not NumPy’s: Rng::seed(42) does not reproduce
np.random.default_rng(42) (which uses PCG64).
use rustnum as np;
let mut rng = np::random::Rng::seed(7);
let a = rng.random(&[2, 3]);
assert_eq!(a.shape(), &[2, 3]);
assert_eq!(a, np::random::Rng::seed(7).random(&[2, 3]));§Architecture
engine(private): raw generator, distributions andVec-producing functions.api: the publicRngwrapper, which returnsArray.
Structs§
- Rng
- Random number generator using xoshiro256** seeded by SplitMix64, like
np.random.default_rng.