jax.random.triangular

jax.random.triangular(key, left, mode, right, shape=None, dtype=None, *, out_sharding=None)

Sample Triangular random values with given shape and float dtype.

The values are returned according to the probability density function:

\[\begin{split}f(x; a, b, c) = \frac{2}{c-a} \left\{ \begin{array}{ll} \frac{x-a}{b-a} & a \leq x \leq b \\ \frac{c-x}{c-b} & b \leq x \leq c \end{array} \right.\end{split}\]

on the domain \(a \leq x \leq c\).

Parameters:

• key (ArrayLike) – a PRNG key used as the random key.

• left (RealArray) – a float or array of floats broadcast-compatible with shape representing the lower limit parameter of the distribution.

• mode (RealArray) – a float or array of floats broadcast-compatible with shape representing the peak value parameter of the distribution, value must fulfill the condition left <= mode <= right.

• right (RealArray) – a float or array of floats broadcast-compatible with shape representing the upper limit parameter of the distribution, must be larger than left.

• shape (Shape | None) – optional, a tuple of nonnegative integers specifying the result shape. Must be broadcast-compatible with left,``mode`` and right. The default (None) produces a result shape equal to left.shape, mode.shape and right.shape.

• dtype (DTypeLikeFloat | None) – optional, a float dtype for the returned values (default float64 if jax_enable_x64 is true, otherwise float32).

• out_sharding (NamedSharding | P | None) – optional, specifies how the output array should be sharded across devices in multi-device computation. Can be a NamedSharding, a PartitionSpec (P), or None (default). When specified, the output will be sharded according to the given sharding specification. Primarily used in explicit sharding mode. See the explicit sharding tutorial for more details.

Returns:

A random array with the specified dtype and with shape given by shape if shape is not None, or else by left.shape, mode.shape and right.shape.

Return type:

Array