<br />

|---|
| [![](https://www.tensorflow.org/images/GitHub-Mark-32px.png) View source on GitHub](https://github.com/tensorflow/probability/blob/v0.25.0/tensorflow_probability/python/bijectors/__init__.py) |

Bijective transformations.

## Classes

[`class AbsoluteValue`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/AbsoluteValue): Computes `Y = g(X) = Abs(X)`, element-wise.

[`class Ascending`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Ascending): Maps unconstrained R\^n to R\^n in ascending order.

[`class AutoCompositeTensorBijector`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/AutoCompositeTensorBijector): Base for `CompositeTensor` bijectors with auto-generated `TypeSpec`s.

[`class AutoregressiveNetwork`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/AutoregressiveNetwork): Masked Autoencoder for Distribution Estimation \[Germain et al. (2015)\]\[1\].

[`class BatchNormalization`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/BatchNormalization): Compute `Y = g(X) s.t. X = g^-1(Y) = (Y - mean(Y)) / std(Y)`.

[`class Bijector`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Bijector): Interface for transformations of a `Distribution` sample.

[`class Blockwise`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Blockwise): Bijector which applies a list of bijectors to blocks of a `Tensor`.

[`class Chain`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Chain): Bijector which applies a composition of bijectors.

[`class CholeskyOuterProduct`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/CholeskyOuterProduct): Compute `g(X) = X @ X.T`; X is lower-triangular, positive-diagonal matrix.

[`class CholeskyToInvCholesky`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/CholeskyToInvCholesky): Maps the Cholesky factor of `M` to the Cholesky factor of `M^{-1}`.

[`class Composition`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Composition): Base class for Composition bijectors (Chain, JointMap).

[`class CorrelationCholesky`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/CorrelationCholesky): Maps unconstrained reals to Cholesky-space correlation matrices.

[`class Cumsum`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Cumsum): Computes the cumulative sum of a tensor along a specified axis.

[`class DiscreteCosineTransform`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/DiscreteCosineTransform): Compute `Y = g(X) = DCT(X)`, where DCT type is indicated by the `type` arg.

[`class Exp`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Exp): Compute `Y = g(X) = exp(X)`.

[`class Expm1`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Expm1): Compute `Y = g(X) = exp(X) - 1`.

[`class FFJORD`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/FFJORD): Implements a continuous normalizing flow X-\>Y defined via an ODE.

[`class FillScaleTriL`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/FillScaleTriL): Transforms unconstrained vectors to TriL matrices with positive diagonal.

[`class FillTriangular`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/FillTriangular): Transforms vectors to triangular.

[`class FrechetCDF`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/FrechetCDF): The Frechet cumulative density function.

[`class GeneralizedExtremeValueCDF`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/GeneralizedExtremeValueCDF): Compute the GeneralizedExtremeValue CDF.

[`class GeneralizedPareto`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/GeneralizedPareto): Bijector mapping R\*\*n to non-negative reals.

[`class Glow`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Glow): Implements the Glow Bijector from Kingma \& Dhariwal (2018)\[1\].

[`class GlowDefaultExitNetwork`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/GlowDefaultExitNetwork): Default network for the glow exit bijector.

[`class GlowDefaultNetwork`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/GlowDefaultNetwork): Default network for the glow bijector.

[`class GompertzCDF`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/GompertzCDF): Compute `Y = g(X) = 1 - exp(-c * (exp(rate * X) - 1)`, the Gompertz CDF.

[`class GumbelCDF`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/GumbelCDF): Compute `Y = g(X) = exp(-exp(-(X - loc) / scale))`, the Gumbel CDF.

[`class Householder`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Householder): Compute the reflection of a vector across a hyperplane.

[`class Identity`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Identity): Compute Y = g(X) = X.

[`class Inline`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Inline): Bijector constructed from custom callables.

[`class Invert`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Invert): Bijector which inverts another Bijector.

[`class IteratedSigmoidCentered`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/IteratedSigmoidCentered): Bijector which applies a Stick Breaking procedure.

[`class JointMap`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/JointMap): Bijector which applies a structure of bijectors in parallel.

[`class KumaraswamyCDF`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/KumaraswamyCDF): Compute `Y = g(X) = (1 - X**a)**b, X in [0, 1]`.

[`class LambertWTail`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/LambertWTail): LambertWTail transformation for heavy-tail Lambert W x F random variables.

[`class Log`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Log): Compute `Y = log(X)`. This is `Invert(Exp())`.

[`class Log1p`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Log1p): Compute `Y = log1p(X)`. This is `Invert(Expm1())`.

[`class MaskedAutoregressiveFlow`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/MaskedAutoregressiveFlow): Affine MaskedAutoregressiveFlow bijector.

[`class MatrixInverseTriL`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/MatrixInverseTriL): Computes `g(L) = inv(L)`, where `L` is a lower-triangular matrix.

[`class MatvecLU`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/MatvecLU): Matrix-vector multiply using LU decomposition.

[`class MoyalCDF`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/MoyalCDF): Compute `Y = g(X) = erfc(exp(- 1/2 * (X - loc) / scale) / sqrt(2))`.

[`class NormalCDF`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/NormalCDF): Compute `Y = g(X) = NormalCDF(x)`.

[`class Pad`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Pad): Pads a value to the `event_shape` of a `Tensor`.

[`class Permute`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Permute): Permutes the rightmost dimension of a `Tensor`.

[`class Power`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Power): Compute `g(X) = X ** power`; where X is a non-negative real number.

[`class PowerTransform`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/PowerTransform): Compute `Y = g(X) = (1 + X * c)**(1 / c), X >= -1 / c`.

[`class RationalQuadraticSpline`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/RationalQuadraticSpline): A piecewise rational quadratic spline, as developed in \[1\].

[`class RayleighCDF`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/RayleighCDF): Compute `Y = g(X) = 1 - exp( -(X/scale)**2 / 2 ), X >= 0`.

[`class RealNVP`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/RealNVP): RealNVP 'affine coupling layer' for vector-valued events.

[`class Reciprocal`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Reciprocal): A `Bijector` that computes the reciprocal `b(x) = 1. / x` entrywise.

[`class Reshape`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Reshape): Reshapes the `event_shape` of a `Tensor`.

[`class Restructure`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Restructure): Converts between nested structures of Tensors.

[`class Scale`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Scale): Compute `Y = g(X; scale) = scale * X`.

[`class ScaleMatvecDiag`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/ScaleMatvecDiag): Compute `Y = g(X; scale) = scale @ X`.

[`class ScaleMatvecLU`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/ScaleMatvecLU): Matrix-vector multiply using LU decomposition.

[`class ScaleMatvecLinearOperator`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/ScaleMatvecLinearOperator): Compute `Y = g(X; scale) = scale @ X`.

[`class ScaleMatvecLinearOperatorBlock`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/ScaleMatvecLinearOperatorBlock): Compute `Y = g(X; scale) = scale @ X` for blockwise `X` and `scale`.

[`class ScaleMatvecTriL`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/ScaleMatvecTriL): Compute `Y = g(X; scale) = scale @ X`.

[`class Shift`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Shift): Compute `Y = g(X; shift) = X + shift`.

[`class ShiftedGompertzCDF`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/ShiftedGompertzCDF): Compute `Y = g(X) = (1 - exp(-rate * X)) * exp(-c * exp(-rate * X))`.

[`class Sigmoid`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Sigmoid): Bijector that computes the logistic sigmoid function.

[`class Sinh`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Sinh): Bijector that computes `Y = sinh(X)`.

[`class SinhArcsinh`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/SinhArcsinh): `Y = g(X) = Sinh( (Arcsinh(X) + skewness) * tailweight ) * multiplier`.

[`class SoftClip`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/SoftClip): Bijector that approximates clipping as a continuous, differentiable map.

[`class Softfloor`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Softfloor): Compute a differentiable approximation to [`tf.math.floor`](https://www.tensorflow.org/api_docs/python/tf/math/floor).

[`class SoftmaxCentered`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/SoftmaxCentered): Bijector which computes `Y = g(X) = exp([X 0]) / sum(exp([X 0]))`.

[`class Softplus`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Softplus): Bijector which computes `Y = g(X) = Log[1 + exp(X)]`.

[`class Softsign`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Softsign): Bijector which computes `Y = g(X) = X / (1 + |X|)`.

[`class Split`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Split): Split a `Tensor` event along an axis into a list of `Tensor`s.

[`class Square`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Square): Compute `g(X) = X^2`; X is a positive real number.

[`class Tanh`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Tanh): Bijector that computes `Y = tanh(X)`, therefore `Y in (-1, 1)`.

[`class TransformDiagonal`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/TransformDiagonal): Applies a Bijector to the diagonal of a matrix.

[`class Transpose`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/Transpose): Compute `Y = g(X) = transpose_rightmost_dims(X, rightmost_perm)`.

[`class UnitVector`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/UnitVector): Bijector mapping vectors onto the unit sphere.

[`class WeibullCDF`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/WeibullCDF): Compute `Y = g(X) = 1 - exp( -( X / scale) ** concentration), X >= 0`.

## Functions

[`masked_autoregressive_default_template(...)`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/masked_autoregressive_default_template): Build the Masked Autoregressive Density Estimator (Germain et al., 2015). (deprecated)

[`masked_dense(...)`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/masked_dense): A autoregressively masked dense layer. Analogous to `tf.layers.dense`.

[`pack_sequence_as(...)`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/pack_sequence_as): Returns a Bijector variant of tf.nest.pack_sequence_as.

[`real_nvp_default_template(...)`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/real_nvp_default_template): Build a scale-and-shift function using a multi-layer neural network.

[`tree_flatten(...)`](https://www.tensorflow.org/probability/api_docs/python/tfp/bijectors/tree_flatten): Returns a Bijector variant of tf.nest.flatten.