<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/experimental/nn/__init__.py) |

Tools for building neural networks.

## Modules

[`initializers`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/initializers) module: Initializer functions for building neural networks.

[`losses`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/losses) module: Loss functions for neural networks.

[`util`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/util) module: Utilitity functions for building neural networks.

## Classes

[`class Affine`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/Affine): Basic affine layer.

[`class AffineVariationalFlipout`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/AffineVariationalFlipout): Densely-connected layer class with Flipout estimator.

[`class AffineVariationalReparameterization`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/AffineVariationalReparameterization): Densely-connected layer class with reparameterization estimator.

[`class AffineVariationalReparameterizationLocal`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/AffineVariationalReparameterizationLocal): Densely-connected layer class with local reparameterization estimator.

[`class Convolution`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/Convolution): Convolution layer.

[`class ConvolutionTranspose`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/ConvolutionTranspose): ConvolutionTranspose layer.

[`class ConvolutionTransposeVariationalFlipout`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/ConvolutionTransposeVariationalFlipout): ConvolutionTranspose layer class with Flipout estimator.

[`class ConvolutionTransposeVariationalReparameterization`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/ConvolutionTransposeVariationalReparameterization): ConvolutionTranspose layer class with reparameterization estimator.

[`class ConvolutionV2`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/ConvolutionV2): Convolution layer.

[`class ConvolutionVariationalFlipout`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/ConvolutionVariationalFlipout): Convolution layer class with Flipout estimator.

[`class ConvolutionVariationalFlipoutV2`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/ConvolutionVariationalFlipoutV2): Convolution layer class with Flipout estimator.

[`class ConvolutionVariationalReparameterization`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/ConvolutionVariationalReparameterization): Convolution layer class with reparameterization estimator.

[`class ConvolutionVariationalReparameterizationV2`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/ConvolutionVariationalReparameterizationV2): Convolution layer class with reparameterization estimator.

[`class Layer`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/Layer): A `callable` [`tf.Module`](https://www.tensorflow.org/api_docs/python/tf/Module).

[`class Sequential`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/Sequential): A `Layer` characterized by iteratively given functions.

[`class VariationalLayer`](https://www.tensorflow.org/probability/api_docs/python/tfp/experimental/nn/VariationalLayer): Base class for all variational layers.