mindspore.nn.probability.distribution.Categorical

class mindspore.nn.probability.distribution.Categorical(probs=None, seed=None, dtype=mstype.int32, name='Categorical')[source]

Categorical distribution. A Categorical Distribution is a discrete distribution with the range \(\{1, 2, ..., k\}\) and the probability mass function as \(P(X = i) = p_i, i = 1, ..., k\).

Parameters

probs (Tensor, list, numpy.ndarray) – Event probabilities. Default: None.
seed (int) – The global seed is used in sampling. Global seed is used if it is None. Default: None.
dtype (mindspore.dtype) – The type of the event samples. Default: mstype.int32.
name (str) – The name of the distribution. Default: Categorical.

Note

probs must have rank at least 1, values are proper probabilities and sum to 1.

Raises: ValueError – When the sum of all elements in probs is not 1.

Supported Platforms:: Ascend GPU

Examples

>>> import mindspore
>>> import mindspore.nn as nn
>>> import mindspore.nn.probability.distribution as msd
>>> from mindspore import Tensor
>>> # To initialize a Categorical distribution of probs [0.5, 0.5]
>>> ca1 = msd.Categorical(probs=[0.2, 0.8], dtype=mindspore.int32)
>>> # A Categorical distribution can be initialized without arguments.
>>> # In this case, `probs` must be passed in through arguments during function calls.
>>> ca2 = msd.Categorical(dtype=mindspore.int32)
>>> # Here are some tensors used below for testing
>>> value = Tensor([1, 0], dtype=mindspore.int32)
>>> probs_a = Tensor([0.5, 0.5], dtype=mindspore.float32)
>>> probs_b = Tensor([0.35, 0.65], dtype=mindspore.float32)
>>> # Private interfaces of probability functions corresponding to public interfaces, including
>>> # `prob`, `log_prob`, `cdf`, `log_cdf`, `survival_function`, and `log_survival`, are the same as follows.
>>> # Args:
>>> #     value (Tensor): the value to be evaluated.
>>> #     probs (Tensor): event probabilities. Default: self.probs.
>>> # Examples of `prob`.
>>> # Similar calls can be made to other probability functions
>>> # by replacing `prob` by the name of the function.
>>> ans = ca1.prob(value)
>>> print(ans.shape)
(2,)
>>> # Evaluate `prob` with respect to distribution b.
>>> ans = ca1.prob(value, probs_b)
>>> print(ans.shape)
(2,)
>>> # `probs` must be passed in during function calls.
>>> ans = ca2.prob(value, probs_a)
>>> print(ans.shape)
(2,)
>>> # Functions `mean`, `sd`, `var`, and `entropy` have the same arguments.
>>> # Args:
>>> #     probs (Tensor): event probabilities. Default: self.probs.
>>> # Examples of `mean`. `sd`, `var`, and `entropy` are similar.
>>> ans = ca1.mean() # return 0.8
>>> print(ans.shape)
(1,)
>>> ans = ca1.mean(probs_b)
>>> print(ans.shape)
(1,)
>>> # `probs` must be passed in during function calls.
>>> ans = ca2.mean(probs_a)
>>> print(ans.shape)
(1,)
>>> # Interfaces of `kl_loss` and `cross_entropy` are the same as follows:
>>> # Args:
>>> #     dist (str): the name of the distribution. Only 'Categorical' is supported.
>>> #     probs_b (Tensor): event probabilities of distribution b.
>>> #     probs (Tensor): event probabilities of distribution a. Default: self.probs.
>>> # Examples of `kl_loss`, `cross_entropy` is similar.
>>> ans = ca1.kl_loss('Categorical', probs_b)
>>> print(ans.shape)
()
>>> ans = ca1.kl_loss('Categorical', probs_b, probs_a)
>>> print(ans.shape)
()
>>> # An additional `probs` must be passed in.
>>> ans = ca2.kl_loss('Categorical', probs_b, probs_a)
>>> print(ans.shape)
()

property probs

Return the event probability.

Returns: Tensor, the event probability.

cdf(value, probs)

Compute the cumulatuve distribution function(CDF) of the given value.

Parameters

value (Tensor) - the value to compute.
probs (Tensor) - the event probability. Default value: None.

Returns

Tensor, the value of the cumulatuve distribution function for the given input.

cross_entropy(dist, probs_b, probs)

Compute the cross entropy of two distribution.

Parameters

dist (str) - the type of the other distribution.
probs_b (Tensor) - the event probability of the other distribution.
probs (Tensor) - the event probability. Default value: None.

Returns

Tensor, the value of the cross entropy.

entropy(probs)

Compute the value of the entropy.

Parameters

probs (Tensor) - the event probability. Default value: None.

Returns

Tensor, the value of the entropy.

kl_loss(dist, probs_b, probs)

Compute the value of the K-L loss between two distribution, namely KL(a||b).

Parameters

dist (str) - the type of the other distribution.
probs_b (Tensor) - the event probability of the other distribution.
probs (Tensor) - the event probability. Default value: None.

Returns

Tensor, the value of the K-L loss.

log_cdf(value, probs)

Compute the log value of the cumulatuve distribution function.

Parameters

value (Tensor) - the value to compute.
probs (Tensor) - the event probability. Default value: None.

Returns

Tensor, the log value of the cumulatuve distribution function.

log_prob(value, probs)

the log value of the probability.

Parameters

value (Tensor) - the value to compute.
probs (Tensor) - the event probability. Default value: None.

Returns

Tensor, the log value of the probability.

log_survival(value, probs)

Compute the log value of the survival function.

Parameters

value (Tensor) - the value to compute.
probs (Tensor) - the event probability. Default value: None.

Returns

Tensor, the value of the K-L loss.

mean(probs)

Compute the mean value of the distribution.

Parameters

probs (Tensor) - the event probability. Default value: None.

Returns

Tensor, the mean of the distribution.

mode(probs)

Compute the mode value of the distribution.

Parameters

probs (Tensor) - the event probability. Default value: None.

Returns

Tensor, the mode of the distribution.

prob(value, probs)

The probability of the given value. For the discrete distribution, it is the probability mass function(pmf).

Parameters

value (Tensor) - the value to compute.
probs (Tensor) - the event probability. Default value: None.

Returns

Tensor, the value of the probability.

sample(shape, probs)

Generate samples.

Parameters

shape (tuple) - the shape of the sample.
probs (Tensor) - the event probability. Default value: None.

Returns

Tensor, the sample following the distribution.

sd(probs)

The standard deviation.

Parameters

probs (Tensor) - the event probability. Default value: None.

Returns

Tensor, the standard deviation of the distribution.

survival_function(value, probs)

Compute the value of the survival function.

Parameters

value (Tensor) - the value to compute.
probs (Tensor) - the event probability. Default value: None.

Returns

Tensor, the value of the survival function.

var(probs)

Compute the variance of the distribution.

Parameters

probs (Tensor) - the event probability. Default value: None.

Returns

Tensor, the variance of the distribution.