Base agent

mighty.mighty_agents.base_agent

Base agent template.

MightyAgent

MightyAgent(
    output_dir,
    env: MIGHTYENV,
    seed: int | None = None,
    eval_env: MIGHTYENV | None = None,
    learning_rate: float = 0.01,
    epsilon: float = 0.1,
    batch_size: int = 64,
    learning_starts: int = 1,
    n_gradient_steps: int = 1,
    render_progress: bool = True,
    log_wandb: bool = False,
    wandb_kwargs: dict | None = None,
    replay_buffer_class: str
    | DictConfig
    | type[MightyReplay]
    | type[MightyRolloutBuffer]
    | None = None,
    replay_buffer_kwargs: TypeKwargs | None = None,
    meta_methods: list[str | type] | None = None,
    meta_kwargs: list[TypeKwargs] | None = None,
    verbose: bool = True,
)

Bases: ABC

Base agent for RL implementations.

Creates all relevant class variables and calls agent-specific init function

:param env: Train environment :param eval_env: Evaluation environment :param learning_rate: Learning rate for training :param epsilon: Exploration factor for training :param batch_size: Batch size for training :param render_progress: Render progress :param log_tensorboard: Log to tensorboard as well as to file :param log_wandb: Whether to log to wandb :param wandb_kwargs: Kwargs for wandb.init, e.g. including the project name :param replay_buffer_class: Replay buffer class from coax replay buffers :param replay_buffer_kwargs: Arguments for the replay buffer :param tracer_class: Reward tracing class from coax tracers :param tracer_kwargs: Arguments for the reward tracer :param meta_methods: Class names or types of mighty meta learning modules to use :param meta_kwargs: List of kwargs for the meta learning modules :return:

Source code in mighty/mighty_agents/base_agent.py

def __init__(  # noqa: PLR0915, PLR0912
    self,
    output_dir,
    env: MIGHTYENV,  # type: ignore
    seed: int | None = None,
    eval_env: MIGHTYENV | None = None,  # type: ignore
    learning_rate: float = 0.01,
    epsilon: float = 0.1,
    batch_size: int = 64,
    learning_starts: int = 1,
    n_gradient_steps: int = 1,
    render_progress: bool = True,
    log_wandb: bool = False,
    wandb_kwargs: dict | None = None,
    replay_buffer_class: str
    | DictConfig
    | type[MightyReplay]
    | type[MightyRolloutBuffer]
    | None = None,
    replay_buffer_kwargs: TypeKwargs | None = None,
    meta_methods: list[str | type] | None = None,
    meta_kwargs: list[TypeKwargs] | None = None,
    verbose: bool = True,
):
    """Base agent initialization.

    Creates all relevant class variables and calls agent-specific init function

    :param env: Train environment
    :param eval_env: Evaluation environment
    :param learning_rate: Learning rate for training
    :param epsilon: Exploration factor for training
    :param batch_size: Batch size for training
    :param render_progress: Render progress
    :param log_tensorboard: Log to tensorboard as well as to file
    :param log_wandb: Whether to log to wandb
    :param wandb_kwargs: Kwargs for wandb.init, e.g. including the project name
    :param replay_buffer_class: Replay buffer class from coax replay buffers
    :param replay_buffer_kwargs: Arguments for the replay buffer
    :param tracer_class: Reward tracing class from coax tracers
    :param tracer_kwargs: Arguments for the reward tracer
    :param meta_methods: Class names or types of mighty meta learning modules to use
    :param meta_kwargs: List of kwargs for the meta learning modules
    :return:
    """
    if meta_kwargs is None:
        meta_kwargs = []
    if meta_methods is None:
        meta_methods = []
    if wandb_kwargs is None:
        wandb_kwargs = {}
    self.learning_rate = learning_rate
    self._epsilon = epsilon
    self._batch_size = batch_size
    self._learning_starts = learning_starts
    self.n_gradient_steps = n_gradient_steps

    self.buffer: MightyReplay | None = None
    self.policy: MightyExplorationPolicy | None = None

    self.seed = seed
    if self.seed is not None:
        self.rng = np.random.default_rng(seed=seed)
        torch.manual_seed(seed)
    else:
        self.rng = np.random.default_rng()

    # Replay Buffer
    replay_buffer_class = retrieve_class(
        cls=replay_buffer_class,
        default_cls=MightyReplay,  # type: ignore
    )
    if replay_buffer_kwargs is None or len(replay_buffer_kwargs) == 0:
        if issubclass(replay_buffer_class, MightyReplay):
            replay_buffer_kwargs = {  # type: ignore
                "capacity": 1_000_000,
            }
        else:
            replay_buffer_kwargs = {}

    self.buffer_class = replay_buffer_class
    self.buffer_kwargs = replay_buffer_kwargs

    self.output_dir = output_dir
    self.verbose = verbose

    self.env = env
    if eval_env is None:
        self.eval_env = self.env
    else:
        self.eval_env = eval_env

    self.render_progress = render_progress
    self.output_dir = output_dir
    if self.output_dir is not None:
        self.model_dir = Path(self.output_dir) / Path("models")

    # Create meta modules
    self.meta_modules = {}
    for i, m in enumerate(meta_methods):
        meta_class = retrieve_class(cls=m, default_cls=None)  # type: ignore
        assert meta_class is not None, (
            f"Class {m} not found, did you specify the correct loading path?"
        )
        kwargs: Dict = {}
        if len(meta_kwargs) > i:
            kwargs = meta_kwargs[i]
        self.meta_modules[meta_class.__name__] = meta_class(**kwargs)

    self.last_state = None
    self.total_steps = 0

    self.result_buffer = {
        "seed": [],
        "step": [],
        "reward": [],
        "action": [],
        "state": [],
        "next_state": [],
        "terminated": [],
        "truncated": [],
        "mean_episode_reward": [],
    }

    self.eval_buffer = {
        "step": [],
        "seed": [],
        "eval_episodes": [],
        "mean_eval_step_reward": [],
        "mean_eval_reward": [],
        "instance": [],
    }

    self.hp_buffer = {
        "step": [],
        "hp/lr": [],
        "hp/pi_epsilon": [],
        "hp/batch_size": [],
        "hp/learning_starts": [],
        "meta_modules": [],
    }
    self.loss_buffer = None
    starting_hps = {
        "step": 0,
        "hp/lr": self.learning_rate,
        "hp/pi_epsilon": self._epsilon,
        "hp/batch_size": self._batch_size,
        "hp/learning_starts": self._learning_starts,
        "meta_modules": list(self.meta_modules.keys()),
    }
    self.hp_buffer = update_buffer(self.hp_buffer, starting_hps)

    self.log_wandb = log_wandb
    if log_wandb:
        wandb.init(**wandb_kwargs)
        wandb.log(starting_hps)

    self.initialize_agent()
    self.steps = 0

__del__

__del__() -> None

Close wandb upon deletion.

Source code in mighty/mighty_agents/base_agent.py

def __del__(self) -> None:
    """Close wandb upon deletion."""
    self.env.close()  # type: ignore
    if self.log_wandb:
        wandb.finish()

adapt_hps

adapt_hps(metrics: Dict) -> None

Set hyperparameters.

Source code in mighty/mighty_agents/base_agent.py

def adapt_hps(self, metrics: Dict) -> None:
    """Set hyperparameters."""
    old_hps = {
        "step": self.steps,
        "hp/lr": self.learning_rate,
        "hp/pi_epsilon": self._epsilon,
        "hp/batch_size": self._batch_size,
        "hp/learning_starts": self._learning_starts,
        "meta_modules": list(self.meta_modules.keys()),
    }
    self.learning_rate = metrics["hp/lr"]
    self._epsilon = metrics["hp/pi_epsilon"]
    self._batch_size = metrics["hp/batch_size"]
    self._learning_starts = metrics["hp/learning_starts"]

    updated_hps = {
        "step": self.steps,
        "hp/lr": self.learning_rate,
        "hp/pi_epsilon": self._epsilon,
        "hp/batch_size": self._batch_size,
        "hp/learning_starts": self._learning_starts,
        "meta_modules": list(self.meta_modules.keys()),
    }
    # FIXME: EWRL: Can we naively compare dictoinaries like this?
    if old_hps != updated_hps:
        self.hp_buffer = update_buffer(self.hp_buffer, updated_hps)

apply_config

apply_config(config: Dict) -> None

Apply config to agent.

Source code in mighty/mighty_agents/base_agent.py

def apply_config(self, config: Dict) -> None:
    """Apply config to agent."""
    for n in config:
        algo_name = n.split(".")[-1]
        if hasattr(self, algo_name):
            setattr(self, algo_name, config[n])
        elif hasattr(self, "_" + algo_name):
            setattr(self, "_" + algo_name, config[n])
        elif n in ["architecture", "n_units", "n_layers", "size"]:
            pass
        else:
            print(f"Trying to set hyperparameter {algo_name} which does not exist.")

evaluate

evaluate(eval_env: MIGHTYENV | None = None) -> Dict

Eval agent on an environment. (Full rollouts).

:param env: The environment to evaluate on :param episodes: The number of episodes to evaluate :return:

Source code in mighty/mighty_agents/base_agent.py

def evaluate(self, eval_env: MIGHTYENV | None = None) -> Dict:  # type: ignore
    """Eval agent on an environment. (Full rollouts).

    :param env: The environment to evaluate on
    :param episodes: The number of episodes to evaluate
    :return:
    """

    terminated, truncated = False, False
    options: Dict = {}
    if eval_env is None:
        eval_env = self.eval_env

    state, _ = eval_env.reset(options=options)  # type: ignore
    rewards = np.zeros(eval_env.num_envs)  # type: ignore
    steps = np.zeros(eval_env.num_envs)  # type: ignore
    mask = np.zeros(eval_env.num_envs)  # type: ignore
    while not np.all(mask):
        action = self.policy(state, evaluate=True)  # type: ignore
        state, reward, terminated, truncated, _ = eval_env.step(action)  # type: ignore
        rewards += reward * (1 - mask)
        steps += 1 * (1 - mask)
        dones = np.logical_or(terminated, truncated)
        mask = np.where(dones, 1, mask)

    eval_env.close()  # type: ignore

    if isinstance(self.eval_env, DACENV) or isinstance(self.env, CARLENV):
        instance = eval_env.instance  # type: ignore
    else:
        instance = "None"

    eval_metrics = {
        "step": self.steps,
        "seed": self.seed,
        "eval_episodes": np.array(rewards) / steps,
        "mean_eval_step_reward": np.mean(rewards) / steps,
        "mean_eval_reward": np.mean(rewards),
        "instance": instance,
        "eval_rewards": rewards,
    }
    self.eval_buffer = update_buffer(self.eval_buffer, eval_metrics)

    if self.log_wandb:
        wandb.log(eval_metrics)

    return eval_metrics

initialize_agent

initialize_agent() -> None

General initialization of tracer and buffer for all agents.

Algorithm specific initialization like policies etc. are done in _initialize_agent

Source code in mighty/mighty_agents/base_agent.py

def initialize_agent(self) -> None:
    """General initialization of tracer and buffer for all agents.

    Algorithm specific initialization like policies etc.
    are done in _initialize_agent
    """
    self._initialize_agent()
    self.buffer = self.buffer_class(**self.buffer_kwargs)  # type: ignore

make_checkpoint_dir

make_checkpoint_dir(t: int) -> None

Checkpoint model.

:param T: Current timestep :return:

Source code in mighty/mighty_agents/base_agent.py

def make_checkpoint_dir(self, t: int) -> None:
    """Checkpoint model.

    :param T: Current timestep
    :return:
    """
    self.upper_checkpoint_dir = Path(self.output_dir) / Path("checkpoints")
    if not self.upper_checkpoint_dir.exists():
        Path(self.upper_checkpoint_dir).mkdir()
    self.checkpoint_dir = self.upper_checkpoint_dir / f"{t}"
    if not self.checkpoint_dir.exists():
        Path(self.checkpoint_dir).mkdir()

process_transition

process_transition(
    curr_s,
    action,
    reward,
    next_s,
    dones,
    log_prob=None,
    metrics=None,
) -> Dict

Agent/algorithm specific transition operations.

Source code in mighty/mighty_agents/base_agent.py

def process_transition(  # type: ignore
    self, curr_s, action, reward, next_s, dones, log_prob=None, metrics=None
) -> Dict:
    """Agent/algorithm specific transition operations."""
    raise NotImplementedError

run

run(
    n_steps: int,
    eval_every_n_steps: int = 1000,
    save_model_every_n_steps: int | None = 5000,
    env: MIGHTYENV = None,
) -> Dict

Run agent.

Source code in mighty/mighty_agents/base_agent.py

def run(  # noqa: PLR0915
    self,
    n_steps: int,
    eval_every_n_steps: int = 1_000,
    save_model_every_n_steps: int | None = 5000,
    env: MIGHTYENV = None,  # type: ignore
) -> Dict:
    """Run agent."""
    episodes = 0
    if env is not None:
        self.env = env

    logging_layout = Layout()
    logging_layout.split_column(
        Layout(name="upper"), Layout(name="middle"), Layout(name="lower")
    )
    logging_layout["middle"].split_row(
        Layout(name="left"),
        Layout(name="right"),
    )
    logging_layout["lower"].split_row(
        Layout(name="left"),
        Layout(name="right"),
    )

    metrics_table = self.make_logging_table(0, [0], [0], [0], [0])  # type: ignore
    training_detail_table = Table(title="Training Details")
    training_detail_table.add_column("Name", style="cyan", justify="center")
    training_detail_table.add_column("Value", style="cyan", justify="center")
    training_detail_table.add_row("Training Steps", str(n_steps))
    training_detail_table.add_row("Seed", str(self.seed))
    training_detail_table.add_row("Learning Rate", str(self.learning_rate))
    training_detail_table.add_row("Batch Size", str(self._batch_size))

    logging_layout["middle"]["left"].update(metrics_table)
    logging_layout["middle"]["right"].update(training_detail_table)

    progress = Progress(
        "[progress.description]{task.description}",
        BarColumn(),
        "[progress.percentage]{task.percentage:>3.0f}%",
        "Remaining:",
        TimeRemainingColumn(),
        "Elapsed:",
        TimeElapsedColumn(),
        disable=not self.render_progress,
    )
    steps_task = progress.add_task(
        "Train Steps",
        total=n_steps - self.steps,
        start=False,
        visible=False,
    )
    progress.start_task(steps_task)

    progress_table = Table.grid(expand=True)
    progress_table.add_row(
        Panel(
            progress,
            title="Training Progress",
            border_style="green",
            padding=(2, 2),
        ),
    )
    logging_layout["upper"].update(progress_table)
    update_multiplier = 0

    with Live(logging_layout, refresh_per_second=10, vertical_overflow="visible"):
        steps_since_eval = 0
        # FIXME: this is more of a question: are there cases where we don't want to reset this completely?
        # I can't think of any, can you? If yes, we should maybe add this as an optional argument
        metrics = {
            "env": self.env,
            "vf": self.value_function,  # type: ignore
            "policy": self.policy,
            "step": self.steps,
            "hp/lr": self.learning_rate,
            "hp/pi_epsilon": self._epsilon,
            "hp/batch_size": self._batch_size,
            "hp/learning_starts": self._learning_starts,
        }

        # Reset env and initialize reward sum
        curr_s, _ = self.env.reset()  # type: ignore
        if len(curr_s.squeeze().shape) == 0:
            episode_reward = [0]
        else:
            episode_reward = np.zeros(curr_s.squeeze().shape[0])  # type: ignore

        last_episode_reward = episode_reward
        if not torch.is_tensor(last_episode_reward):
            last_episode_reward = torch.tensor(last_episode_reward).float()
        progress.update(steps_task, visible=True)
        recent_episode_reward = []
        recent_step_reward = []
        recent_actions = []
        evaluation_reward = []
        eval_curve = [0]
        learning_curve = [0]
        curve_xs = [0]
        logging_layout["lower"]["left"].update(
            self.get_plot(curve_xs, learning_curve, "Training Reward")
        )
        logging_layout["lower"]["right"].update(
            self.get_plot(curve_xs, eval_curve, "Evaluation Reward")
        )

        # Main loop: rollouts, training and evaluation
        while self.steps < n_steps:
            metrics["episode_reward"] = episode_reward

            action, log_prob = self.step(curr_s, metrics)

            next_s, reward, terminated, truncated, _ = self.env.step(action)  # type: ignore
            dones = np.logical_or(terminated, truncated)

            episode_reward += reward

            # Log everything
            t = {
                "seed": self.seed,
                "step": self.steps,
                "reward": reward,
                "action": action,
                "state": curr_s,
                "next_state": next_s,
                "terminated": terminated.astype(int),
                "truncated": truncated.astype(int),
                "dones": dones.astype(int),
                "mean_episode_reward": last_episode_reward.mean().cpu().numpy().item(),
            }
            metrics["log_prob"] = log_prob.detach().cpu().numpy()
            metrics["episode_reward"] = episode_reward
            metrics["transition"] = t

            recent_actions.append(np.mean(action))
            if len(recent_actions) > 100:
                recent_actions.pop(0)

            for k in self.meta_modules:
                self.meta_modules[k].post_step(metrics)

            transition_metrics = self.process_transition(
                metrics["transition"]["state"],
                metrics["transition"]["action"],
                metrics["transition"]["reward"],
                metrics["transition"]["next_state"],
                metrics["transition"]["dones"],
                metrics["log_prob"],
                metrics,
            )
            metrics.update(transition_metrics)

            self.result_buffer = update_buffer(self.result_buffer, t)

            if self.log_wandb:
                wandb.log(t)

            self.steps += len(action)
            metrics["step"] = self.steps
            steps_since_eval += len(action)
            for _ in range(len(action)):
                progress.advance(steps_task)

            # Update agent
            if (
                len(self.buffer) >= self._batch_size  # type: ignore
                and self.steps >= self._learning_starts
            ):
                update_kwargs = {"next_s": next_s, "dones": dones}

                metrics = self.update(metrics, update_kwargs)

            # End step
            self.last_state = curr_s
            curr_s = next_s

            # Evaluate
            if eval_every_n_steps and steps_since_eval >= eval_every_n_steps:
                steps_since_eval = 0
                eval_metrics = self.evaluate()
                evaluation_reward = eval_metrics["eval_rewards"]

            # Log to command line
            if self.steps >= 1000*update_multiplier:
                metrics_table = self.make_logging_table(
                    self.steps,
                    recent_episode_reward,
                    recent_step_reward,
                    evaluation_reward,
                    recent_actions,
                )
                logging_layout["middle"]["left"].update(metrics_table)
                eval_curve.append(np.mean(evaluation_reward))
                learning_curve.append(np.mean(recent_episode_reward))
                curve_xs.append(self.steps)

                logging_layout["lower"]["left"].update(
                    self.get_plot(curve_xs, learning_curve, "Training Reward")
                )
                logging_layout["lower"]["right"].update(
                    self.get_plot(curve_xs, eval_curve, "Evaluation Reward")
                )
                update_multiplier += 1

            # Save
            if (
                save_model_every_n_steps
                and self.steps % save_model_every_n_steps == 0
            ):
                self.save(self.steps)
                log_to_file(
                    self.output_dir,
                    self.result_buffer,
                    self.hp_buffer,
                    self.eval_buffer,
                    self.loss_buffer,
                )

            if np.any(dones):
                last_episode_reward = np.where(  # type: ignore
                    dones, episode_reward, last_episode_reward
                )
                recent_episode_reward.append(np.mean(last_episode_reward))
                recent_step_reward.append(
                    np.mean(last_episode_reward) / len(last_episode_reward)
                )
                last_episode_reward = torch.tensor(last_episode_reward).float()
                if len(recent_episode_reward) > 10:
                    recent_episode_reward.pop(0)
                    recent_step_reward.pop(0)
                episode_reward = np.where(dones, 0, episode_reward)  # type: ignore
                # End episode
                if isinstance(self.env, DACENV) or isinstance(self.env, CARLENV):
                    instance = self.env.instance  # type: ignore
                else:
                    instance = None
                metrics["instance"] = instance
                episodes += 1
                for k in self.meta_modules:
                    self.meta_modules[k].post_episode(metrics)

                # Meta Module hooks
                for k in self.meta_modules:
                    self.meta_modules[k].pre_episode(metrics)
    log_to_file(
        self.output_dir,
        self.result_buffer,
        self.hp_buffer,
        self.eval_buffer,
        self.loss_buffer,
    )
    return metrics

update

update(metrics: Dict, update_kwargs: Dict) -> Dict

Update agent.

Source code in mighty/mighty_agents/base_agent.py

def update(self, metrics: Dict, update_kwargs: Dict) -> Dict:
    """Update agent."""
    for k in self.meta_modules:
        self.meta_modules[k].pre_update(metrics)

    batches = []
    for batches_left in reversed(range(self.n_gradient_steps)):
        batch = self.buffer.sample(self._batch_size)
        agent_update_metrics = self.update_agent(
            transition_batch=batch, batches_left=batches_left, **update_kwargs
        )

        # FIXME: EWRL: Move logging out of gradient steps

        metrics.update(agent_update_metrics)

        metrics = {k: np.array(v) for k, v in metrics.items()}
        metrics["step"] = self.steps

        # Wandb logging
        if self.log_wandb:
            import json

            import wandb

            # Only log relevant, serializable keys
            log_keys = [
                "step",
                "episode_reward",
                "Update/policy_loss",
                "Update/value_loss",
                "Update/entropy",
                "Update/approx_kl",
            ]
            serializable_metrics = {}
            for k in log_keys:
                if k in metrics:
                    v = metrics[k]
                    # Convert numpy arrays to scalars or lists
                    if isinstance(v, np.ndarray):
                        if v.size == 1:
                            v = v.item()
                        else:
                            v = v.tolist()
                    # Convert torch tensors to scalars or lists
                    if isinstance(v, torch.Tensor):
                        if v.numel() == 1:
                            v = v.item()
                        else:
                            v = v.cpu().numpy().tolist()
                    # Try to serialize, skip if not possible
                    try:
                        json.dumps(v)
                        serializable_metrics[k] = v
                    except TypeError:
                        print(f"Skipping non-serializable metric: {k}")

            wandb.log(serializable_metrics, step=self.steps)

        metrics["env"] = self.env
        metrics["vf"] = self.value_function  # type: ignore
        metrics["policy"] = self.policy
        batches.append(batch)

    metrics["update_batches"] = batches
    for k in self.meta_modules:
        self.meta_modules[k].post_update(metrics)
    del metrics["update_batches"]
    return metrics

update_agent

update_agent() -> Dict

Policy/value function update.

Source code in mighty/mighty_agents/base_agent.py

def update_agent(self) -> Dict:
    """Policy/value function update."""
    raise NotImplementedError