tensortrade.agents.agent module¶

class tensortrade.agents.agent.Agent[source]¶

Bases: tensortrade.core.base.Identifiable

abstract get_action(state: numpy.ndarray, **kwargs) → int [source]¶: Get an action for a specific state in the environment.

abstract restore(path: str, **kwargs)[source]¶: Restore the agent from the file specified in path.

abstract save(path: str, **kwargs)[source]¶: Save the agent to the directory specified in path.

abstract train(n_steps: int = None, n_episodes: int = 10000, save_every: int = None, save_path: str = None, callback: callable = None, **kwargs) → float [source]¶: Train the agent in the environment and return the mean reward.