d3rlpy.dataset.create_fifo_replay_buffer

d3rlpy.dataset.create_fifo_replay_buffer(limit, episodes=None, transition_picker=None, trajectory_slicer=None, writer_preprocessor=None, env=None, write_at_termination=False)[source]

Builds FIFO replay buffer.

This function is a shortcut alias to build replay buffer with FIFOBuffer.

Parameters:

  • limit (int) – Maximum capacity of FIFO buffer.

  • episodes (Optional[Sequence[EpisodeBase]]) – List of episodes to initialize replay buffer.

  • transition_picker (Optional[TransitionPickerProtocol]) – Transition picker implementation for Q-learning-based algorithms. If None is given, BasicTransitionPicker is used by default.

  • trajectory_slicer (Optional[TrajectorySlicerProtocol]) – Trajectory slicer implementation for Transformer-based algorithms. If None is given, BasicTrajectorySlicer is used by default.

  • writer_preprocessor (Optional[WriterPreprocessProtocol]) – Writer preprocessor implementation. If None is given, BasicWriterPreprocess is used by default.

  • env (Optional[Union[Env[Any, Any], Env[Any, Any]]]) – Gym environment to extract shapes of observations and action.

  • write_at_termination (bool) – Flag to write experiences to the buffer at the end of an episode all at once.

Returns:

Replay buffer.

Return type:

ReplayBuffer