# tensorplay.distributions

Source: https://www.tensorplay.cn/docs/distributions.html

## Classes

| tensorplay.distributions.AbsTransform |Transform via the mapping $y = \|x\|$. |
| --- | --- |
| tensorplay.distributions.AffineTransform | Transform via the pointwise affine mapping $y = \text{loc} + \text{scale} \times x$. |
| tensorplay.distributions.Bernoulli | Creates a Bernoulli distribution parameterized by probs or logits (but not both). |
| tensorplay.distributions.Beta | Beta distribution parameterized by concentration1 and concentration0. |
| tensorplay.distributions.Binomial | Creates a Binomial distribution parameterized by total_count and either probs or logits (but not both). |
| tensorplay.distributions.CatTransform | Transform functor that applies a sequence of transforms tseq component-wise to each submatrix at dim, of length lengths[dim], in a way compatible with tensorplay.cat(). |
| tensorplay.distributions.Categorical | Creates a categorical distribution parameterized by either probs or logits (but not both). |
| tensorplay.distributions.Cauchy | Samples from a Cauchy (Lorentz) distribution. |
| tensorplay.distributions.Chi2 | Creates a Chi-squared distribution parameterized by shape parameter df. |
| tensorplay.distributions.ComposeTransform | Composes multiple transforms in a chain. |
| tensorplay.distributions.ContinuousBernoulli | Creates a continuous Bernoulli distribution parameterized by probs or logits (but not both). |
| tensorplay.distributions.CorrCholeskyTransform | Transforms an unconstrained real vector $x$ with length $D*(D-1)/2$ into the Cholesky factor of a D-dimension correlation matrix. |
| tensorplay.distributions.CumulativeDistributionTransform | Transform via the cumulative distribution function of a probability distribution. |
| tensorplay.distributions.Dirichlet | Creates a Dirichlet distribution parameterized by concentration concentration. |
| tensorplay.distributions.Distribution | Distribution is the abstract base class for probability distributions. |
| tensorplay.distributions.ExpTransform | Transform via the mapping $y = \exp(x)$. |
| tensorplay.distributions.Exponential | Creates an Exponential distribution parameterized by rate. |
| tensorplay.distributions.ExponentialFamily | ExponentialFamily is the abstract base class for probability distributions belonging to an exponential family, whose probability mass/density function is defined below |
| tensorplay.distributions.FisherSnedecor | Creates a Fisher-Snedecor distribution parameterized by df1 and df2. |
| tensorplay.distributions.Gamma | Creates a Gamma distribution parameterized by shape concentration and rate. |
| tensorplay.distributions.GeneralizedPareto | Creates a Generalized Pareto distribution parameterized by loc, scale, and concentration. |
| tensorplay.distributions.Geometric | Creates a Geometric distribution parameterized by probs, where probs is the probability of success of Bernoulli trials. |
| tensorplay.distributions.Gumbel | Samples from a Gumbel Distribution. |
| tensorplay.distributions.HalfCauchy | Creates a half-Cauchy distribution parameterized by scale where. |
| tensorplay.distributions.HalfNormal | Creates a half-normal distribution parameterized by scale where. |
| tensorplay.distributions.Independent | Reinterprets some of the batch dims of a distribution as event dims. |
| tensorplay.distributions.IndependentTransform | Wrapper around another transform to treat reinterpreted_batch_ndims-many extra of the right most dimensions as dependent. |
| tensorplay.distributions.InverseGamma | Creates an inverse gamma distribution parameterized by concentration and rate where. |
| tensorplay.distributions.Kumaraswamy | Samples from a Kumaraswamy distribution. |
| tensorplay.distributions.LKJCholesky | LKJ distribution for lower Cholesky factor of correlation matrices. The distribution is controlled by concentration parameter $\eta$ to make the probability of the correlation matrix $M$ generated from a Cholesky factor proportional to $\det(M)^{\eta - 1}$. Because of that, when concentration == 1, we have a uniform distribution over Cholesky factors of correlation matrices::. |
| tensorplay.distributions.Laplace | Creates a Laplace distribution parameterized by loc and scale. |
| tensorplay.distributions.LogNormal | Creates a log-normal distribution parameterized by loc and scale where. |
| tensorplay.distributions.LogisticNormal | Creates a logistic-normal distribution parameterized by loc and scale that define the base Normal distribution transformed with the StickBreakingTransform such that. |
| tensorplay.distributions.LowRankMultivariateNormal | Creates a multivariate normal distribution with covariance matrix having a low-rank form parameterized by cov_factor and cov_diag. |
| tensorplay.distributions.LowerCholeskyTransform | Transform from unconstrained matrices to lower-triangular matrices with nonnegative diagonal entries. |
| tensorplay.distributions.MixtureSameFamily | The MixtureSameFamily distribution implements a (batch of) mixture distribution where all components are from different parameterizations of the same distribution type. |
| tensorplay.distributions.Multinomial | Creates a Multinomial distribution parameterized by total_count and either probs or logits (but not both). |
| tensorplay.distributions.MultivariateNormal | Creates a multivariate normal (also called Gaussian) distribution parameterized by a mean vector and a covariance matrix. |
| tensorplay.distributions.NegativeBinomial | Creates a Negative Binomial distribution, i.e. distribution of the number of successful independent and identical Bernoulli trials before total_count failures are achieved. |
| tensorplay.distributions.Normal | Creates a normal (also called Gaussian) distribution parameterized by loc and scale. |
| tensorplay.distributions.OneHotCategorical | Creates a one-hot categorical distribution parameterized by probs or logits. |
| tensorplay.distributions.OneHotCategoricalStraightThrough | Creates a reparameterizable OneHotCategorical distribution based on the straight- through gradient estimator from [1]. |
| tensorplay.distributions.Pareto | Samples from a Pareto Type 1 distribution. |
| tensorplay.distributions.Poisson | Creates a Poisson distribution parameterized by rate, the rate parameter. |
| tensorplay.distributions.PositiveDefiniteTransform | Transform from unconstrained matrices to positive-definite matrices. |
| tensorplay.distributions.PowerTransform | Transform via the mapping $y = x^{\text{exponent}}$. |
| tensorplay.distributions.RelaxedBernoulli | Creates a RelaxedBernoulli distribution, parameterized by temperature, and either probs or logits (but not both). |
| tensorplay.distributions.RelaxedOneHotCategorical | Creates a RelaxedOneHotCategorical distribution parameterized by temperature, and either probs or logits. |
| tensorplay.distributions.ReshapeTransform | Unit Jacobian transform to reshape the rightmost part of a tensor. |
| tensorplay.distributions.SigmoidTransform | Transform via the mapping $y = \frac{1}{1 + \exp(-x)}$ and $x = \text{logit}(y)$. |
| tensorplay.distributions.SoftmaxTransform | Transform from unconstrained space to the simplex via $y = \exp(x)$ then normalizing. |
| tensorplay.distributions.SoftplusTransform | Transform via the mapping $\text{Softplus}(x) = \log(1 + \exp(x))$. |
| tensorplay.distributions.StackTransform | Transform functor that applies a sequence of transforms tseq component-wise to each submatrix at dim in a way compatible with tensorplay.stack(). |
| tensorplay.distributions.StickBreakingTransform | Transform from unconstrained space to the simplex of one additional dimension via a stick-breaking process. |
| tensorplay.distributions.StudentT | Creates a Student's t-distribution parameterized by degree of freedom df, mean loc and scale scale. |
| tensorplay.distributions.TanhTransform | Transform via the mapping $y = \tanh(x)$. |
| tensorplay.distributions.Transform | Abstract class for invertible transformations with computable log det jacobians. |
| tensorplay.distributions.TransformedDistribution | Extension of the Distribution class, which applies a sequence of Transforms to a base distribution. Let f be the composition of transforms applied::. |
| tensorplay.distributions.Uniform | Generates uniformly distributed random samples from the half-open interval [low, high). |
| tensorplay.distributions.VonMises | A circular von Mises distribution. |
| tensorplay.distributions.Weibull | Samples from a two-parameter Weibull distribution. |
| tensorplay.distributions.Wishart | Creates a Wishart distribution parameterized by a symmetric positive definite matrix $\Sigma$, or its Cholesky decomposition $\mathbf{\Sigma} = \mathbf{L}\mathbf{L}^\top$ |

## Functions

| tensorplay.distributions.kl_divergence |Compute Kullback-Leibler divergence $KL(p \\| q)$ between two distributions. |
| --- | --- |
| tensorplay.distributions.register_kl | Decorator to register a pairwise function with kl_divergence(). Usage::. |