View Source Rein.Environment behaviour (rein v0.1.0)
Defines an environment to be passed to Rein.
Summary
Types
The full state of the current Reinforcement Learning process, as stored in the Rein struct
An arbitrary Nx.Container that holds metadata for the environment
Callbacks
Applies the selected action to the environment.
Initializes the environment state with the given enviroment-specific options.
Resets any values that vary between sessions (which would be episodes for episodic tasks, epochs for non-episodic tasks) for the environment state.
Types
@type rl_state() :: Rein.t()
The full state of the current Reinforcement Learning process, as stored in the Rein struct
@type t() :: Nx.Container.t()
An arbitrary Nx.Container that holds metadata for the environment
Callbacks
Applies the selected action to the environment.
Returns the updated environment, also updated with the reward and a flag indicating whether the new state is terminal.
Initializes the environment state with the given enviroment-specific options.
Should be implemented in a way that the result would be semantically
the same as if reset/2 was called in the end of the function.
As a suggestion, the implementation should only initialize fixed
values here, that is values that don't change between sessions
(epochs for non-episodic tasks, episodes for episodic tasks). Then,
call reset/2 internally to initialize the rest of variable values.
Resets any values that vary between sessions (which would be episodes for episodic tasks, epochs for non-episodic tasks) for the environment state.