bellman¶

API reference documentation for the Bellman toolbox.

The top level package structure aims to mirror the package structure in TF-Agents.

To get started, check out the bellman.agents.* package.

`bellman.agents`	This package provides model-based reinforcement learning agents.
`bellman.benchmark`	This package provides helper functions for running experiments and benchmarks of the RL agents which have been implemented in this toolbox.
`bellman.distributions`	This package provides functionality for manipulating probability distributions.
`bellman.drivers`	This package defines drivers which manage the interaction of RL agents and environments.
`bellman.environments`	This package defines a model of a real environment.
`bellman.harness`	This package provides a harness for running experiments with agents.
`bellman.networks`	This package contains layers for transition models implemented using Keras.
`bellman.policies`	This package contains policies which are specific to model-based reinforcement learning.
`bellman.training`	This package provides helper wrappers for training RL agents which have trainable components.
`bellman.trajectory_optimisers`	This package defines methods for trajectory optimisation.