OakEmu-RL v0.1.0

Appearance

TrainerRuntime ​

Minimal trainer state required for DQN checkpoint capture and restore.

Source

Properties ​

replay_buffer ​

python
replay_buffer: ReplayBuffer

Source

steps ​

python
steps: int

Source

total_episodes ​

python
total_episodes: int

Source

total_losses ​

python
total_losses: int

Source

total_truncated ​

python
total_truncated: int

Source

total_wins ​

python
total_wins: int

Source

trajectory_runtime ​

python
trajectory_runtime: TrajectoryEpisodeCoordinator

Source

Methods ​

replace_replay_buffer ​

python
replace_replay_buffer(replay_buffer: ReplayBuffer) -> None

Source

restore_runtime_state ​

python
restore_runtime_state(state: ActorRuntimeState) -> None

Source