Expand description
Reinforcement learning agents
More agents can be found in crate::torch::agents
.
Re-exports
pub use buffers::HistoryDataBound;
pub use buffers::WriteExperience;
pub use buffers::WriteExperienceError;
pub use buffers::WriteExperienceIncremental;
Modules
Structs
A pair of agents / actors / configs for a two-agent environment.
Configuration for BetaThompsonSamplingAgent
An agent that samples actions at random.
Configuration for a random agent.
Lifts a regular agent to act on a meta environment (agent interface).
Configuration for a ResettingMetaAgent
.
A serial combined actor-agent.
Configuration of an Epsilon-Greedy Tabular Q Learning Agent.
Configuration for a UCB1Agent
Enums
Behaviour mode of an actor.
Error building an agent
Traits
Take actions in an environment.
A reinforcement learning agent. Provides environment actors.
Update an agent with steps collected into history buffers.
Build an agent instance for a given environment structure.
Type Definitions
A Thompson sampling agent for Bernoulli rewards with a Beta prior.
An epsilon-greedy tabular Q learning agent.
A UCB1 Agent