Module: tf.contrib.stateless

Defined in tensorflow/contrib/stateless/__init__.py.

Stateless random ops which take seed as a tensor input.

DEPRECATED: Use tf.random.stateless_uniform rather than tf.contrib.stateless.stateless_random_uniform, and similarly for the other routines.

Instead of taking seed as an attr which initializes a mutable state within the op, these random ops take seed as an input, and the random numbers are a deterministic function of shape and seed.

WARNING: These ops are in contrib, and are not stable. They should be consistent across multiple runs on the same hardware, but only for the same version of the code.

Functions

stateless_multinomial(...): Draws deterministic pseudorandom samples from a multinomial distribution. (deprecated)

stateless_random_normal(...): Outputs deterministic pseudorandom values from a normal distribution.

stateless_random_uniform(...): Outputs deterministic pseudorandom values from a uniform distribution.

stateless_truncated_normal(...): Outputs deterministic pseudorandom values, truncated normally distributed.