jax.random.double_sided_maxwell
-
jax.random.double_sided_maxwell(key, loc, scale, shape=(), dtype=<class 'numpy.float64'>)[source]
Sample from a double sided Maxwell distribution.
- Samples using:
loc + scale* sgn(U-0.5)* one_sided_maxwell U~Unif;
- Parameters
key (Union
[Any
, PRNGKeyArray
]) β a PRNG key.
loc (Any
) β The location parameter of the distribution.
scale (Any
) β The scale parameter of the distribution.
shape (Sequence
[int
]) β The shape added to the parameters loc and scale broadcastable shape.
dtype (Any
) β The type used for samples.
- Return type
ndarray
- Returns
A jnp.array of samples.