def categorical(self,
name,
logits,
n_samples=None,
group_ndims=0,
dtype=tf.int32,
**kwargs):
"""
Add a stochastic node in this :class:`BayesianNet` that follows the
Categorical distribution.
:param name: The name of the stochastic node. Must be unique in a
:class:`BayesianNet`.
See
:class:`~zhusuan.distributions.univariate.Categorical` for more
information about the other arguments.
:return: A :class:`StochasticTensor` instance.
"""
dist = distributions.Categorical(logits,
group_ndims=group_ndims,
dtype=dtype,
**kwargs)
return self.stochastic(name, dist, n_samples=n_samples, **kwargs)
def __init__(self,
name,
logits,
n_samples=None,
group_ndims=0,
dtype=None,
**kwargs):
cat = distributions.Categorical(logits,
group_ndims=group_ndims,
dtype=dtype,
**kwargs)
super(Categorical, self).__init__(name, cat, n_samples)
def __init__(self, logits, dtype=None):
"""
Construct the :class:`Categorical`.
Args:
logits: An N-D (N >= 1) `float` Tensor of shape
``(..., n_categories)``. Each slice `[i, j,..., k, :]`
represents the un-normalized log probabilities for all
categories. :math:`\\mathrm{logits} \\propto \\log p`
dtype: The value type of samples from the distribution.
(default ``tf.int32``)
"""
if dtype is None:
dtype = tf.int32
super(Categorical, self).__init__(
zd.Categorical(logits=logits, dtype=dtype))