|View source on GitHub|
Create a random variable for Mixture.
tfp.edward2.Mixture( *args, **kwargs )
See Mixture for more details.
Original Docstring for Distribution
Initialize a Mixture distribution.
Mixture is defined by a
cat, representing the
mixture probabilities) and a list of
all having matching dtype, batch shape, event shape, and continuity
properties (the components).
cat must be possible to infer at graph construction
time and match
Categoricaldistribution instance, representing the probabilities of
components: A list or tuple of
Distributioninstances. Each instance must have the same type, be defined on the same domain, and have matching
True, raise a runtime error if batch or event ranks are inconsistent between cat and any of the distributions. This is only checked if the ranks cannot be determined statically at graph construction time.
allow_nan_stats: Boolean, default
False, raise an exception if a statistic (e.g. mean/mode/etc...) is undefined for any batch member. If
True, batch members with valid parameters leading to undefined statistics will return NaN for this statistic.
use_static_graph: Calls to
samplewill not rely on dynamic tensor indexing, allowing for some static graph compilation optimizations, but at the expense of sampling all underlying distributions in the mixture. (Possibly useful when running on TPUs). Default value:
False(i.e., use dynamic indexing).
name: A name for this distribution (optional).
TypeError: If cat is not a
componentsis not a list or tuple, or the elements of
componentsare not instances of
Distribution, or do not have matching
componentsis an empty list or tuple, or its elements do not have a statically known event rank. If
cat.num_classescannot be inferred at graph creation time, or the constant value of
cat.num_classesis not equal to
len(components), or all
catdo not have matching static batch shapes, or all components do not have matching static event shapes.