Sonnet

Sonnet is an object-oriented library written in Python. It was released by DeepMind in 2017. Sonnet intends to cleanly separate the following two aspects of building computation graphs from objects:

• The configuration of objects called modules
• The connection of objects to computation graphs

Sonnet can be installed in Python 3 with the following command:

pip3 install dm-sonnet

Note

Sonnet can be installed from the source by following directions from the following link: https://github.com/deepmind/sonnet/blob/master/docs/INSTALL.md.

The modules are defined as sub-classes of the abstract class sonnet.AbstractModule. At the time of writing this book, the following modules are available in Sonnet:

Basic modules	AddBias, BatchApply, BatchFlatten, BatchReshape, FlattenTrailingDimensions, Linear, MergeDims, SelectInput, SliceByDim, TileByDim, and TrainableVariable
Recurrent modules	DeepRNN, ModelRNN, VanillaRNN, BatchNormLSTM, GRU, and LSTM
Recurrent + ConvNet modules	Conv1DLSTM and Conv2DLSTM
ConvNet modules...