LostTech.TensorFlow : API Documentation

Type ReparameterizationType

Namespace tensorflow.distributions

Parent PythonObjectContainer

Interfaces IReparameterizationType

Instances of this class represent how sampling is reparameterized.

Two static instances exist in the distributions library, signifying one of two possible properties for samples from a distribution:

`FULLY_REPARAMETERIZED`: Samples from the distribution are fully reparameterized, and straight-through gradients are supported.

`NOT_REPARAMETERIZED`: Samples from the distribution are not fully reparameterized, and straight-through gradients are either partially unsupported or are not supported at all. In this case, for purposes of e.g. RL or variational inference, it is generally safest to wrap the sample results in a `stop_gradients` call and use policy gradients / surrogate loss instead.

Properties

Public properties

object PythonObject get;