microsoft/qlib · error · SeedIteratorNotAvailable

Seed iterator for training is not available.

Error message

Seed iterator for training is not available.

What it means

SeedIteratorNotAvailable raised by the default `Vessel.train_seed_iterator` (qlib/rl/trainer/vessel.py:56). A vessel is abstract: subclasses must override the seed-iterator hooks so the trainer knows which simulator initial states to train on. Calling the base implementation signals the override is missing.

Source

Thrown at qlib/rl/trainer/vessel.py:56

    The ship also defines the most important logic of the core training part,
    and (optionally) some callbacks to insert customized logics at specific events.
    """

    simulator_fn: Callable[[InitialStateType], Simulator[InitialStateType, StateType, ActType]]
    state_interpreter: StateInterpreter[StateType, ObsType]
    action_interpreter: ActionInterpreter[StateType, PolicyActType, ActType]
    policy: BasePolicy
    reward: Reward
    trainer: Trainer

    def assign_trainer(self, trainer: Trainer) -> None:
        self.trainer = weakref.proxy(trainer)  # type: ignore

    def train_seed_iterator(self) -> ContextManager[Iterable[InitialStateType]] | Iterable[InitialStateType]:
        """Override this to create a seed iterator for training.
        If the iterable is a context manager, the whole training will be invoked in the with-block,
        and the iterator will be automatically closed after the training is done."""
        raise SeedIteratorNotAvailable("Seed iterator for training is not available.")

    def val_seed_iterator(self) -> ContextManager[Iterable[InitialStateType]] | Iterable[InitialStateType]:
        """Override this to create a seed iterator for validation."""
        raise SeedIteratorNotAvailable("Seed iterator for validation is not available.")

    def test_seed_iterator(self) -> ContextManager[Iterable[InitialStateType]] | Iterable[InitialStateType]:
        """Override this to create a seed iterator for testing."""
        raise SeedIteratorNotAvailable("Seed iterator for testing is not available.")

    def train(self, vector_env: BaseVectorEnv) -> Dict[str, Any]:
        """Implement this to train one iteration. In RL, one iteration usually refers to one collect."""
        raise NotImplementedError()

    def validate(self, vector_env: FiniteVectorEnv) -> Dict[str, Any]:
        """Implement this to validate the policy once."""
        raise NotImplementedError()

    def test(self, vector_env: FiniteVectorEnv) -> Dict[str, Any]:

View on GitHub (pinned to 79633dd950)

Solutions

1. Override `train_seed_iterator` in your vessel to return an iterable of initial states (e.g. a generator over order/date combinations), optionally a context-manager iterable for resource lifecycle.
2. Model it on `OrderExecutionVessel` / existing vessels in the codebase that yield seeds from an order list.
3. If training is genuinely unsupported for this vessel, catch `SeedIteratorNotAvailable` at the call site and skip the fit phase with a clear log.

Example fix

// before
class MyVessel(Vessel):
    def train(self, venv): ...
    # train_seed_iterator missing -> SeedIteratorNotAvailable
// after
class MyVessel(Vessel):
    def train_seed_iterator(self):
        return iter(self.order_list)  # iterable of initial states
    def train(self, venv): ...

from qlib.rl.trainer.vessel import Vessel

def train_seeds_available(vessel: Vessel) -> bool:
    return type(vessel).train_seed_iterator is not Vessel.train_seed_iterator

def is_trainable_vessel(v) -> bool:
    from qlib.rl.trainer.vessel import Vessel
    return (
        type(v).train_seed_iterator is not Vessel.train_seed_iterator
        and type(v).train is not Vessel.train
    )

from qlib.rl.trainer.vessel import SeedIteratorNotAvailable
try:
    seeds = vessel.train_seed_iterator()
except SeedIteratorNotAvailable:
    raise RuntimeError("override train_seed_iterator before calling Trainer.fit") from None

When it happens

Trigger: Implementing a custom `Vessel` subclass that provides `train`/`validate`/`test` but not `train_seed_iterator`; passing a vessel instance whose seed methods were left at defaults into `Trainer.fit(train_vessel, ...)`.

Common situations: Adapting the RL workflow to a new task and copying only the train loop; refactor renaming the method so the override is lost; using the vessel in a context where training is not intended (then use a vessel that explicitly raises or skip fit).

Related errors

• Seed iterator for validation is not available.
• Seed iterator for testing is not available.
• Implement reward calculation recipe in `reward()`.
• Unsupported earlystopping mode: {mode}
• Render is not implemented in EnvWrapper.

AI-assisted analysis of microsoft/qlib@79633dd950 (2026-08-15). Data as JSON: /api/errors/4e48033b065df9c8. Report an issue: GitHub.