Lightning-AI/pytorch-lightning · error · MisconfigurationException

checkpoint_io set through both strategy class and plugins, c

Error message

checkpoint_io set through both strategy class and plugins, choose one

What it means

If the strategy instance already has checkpoint_io configured (strategy._checkpoint_io), passing another CheckpointIO via Trainer(plugins=[...]) (or checkpoint_io argument) is rejected. Only one source of checkpoint IO is allowed.

Source

Thrown at src/lightning/pytorch/trainer/connectors/accelerator_connector.py:284

        # handle the case when the user passes in a strategy instance which has an accelerator, precision,
        # checkpoint io or cluster env set up
        # TODO: improve the error messages below
        if self._strategy_flag and isinstance(self._strategy_flag, Strategy):
            if self._strategy_flag._accelerator:
                if self._accelerator_flag != "auto":
                    raise MisconfigurationException(
                        "accelerator set through both strategy class and accelerator flag, choose one"
                    )
                self._accelerator_flag = self._strategy_flag._accelerator
            if self._strategy_flag._precision_plugin:
                # [RFC] handle precision plugin set up conflict?
                if self._precision_plugin_flag:
                    raise MisconfigurationException("precision set through both strategy class and plugins, choose one")
                self._precision_plugin_flag = self._strategy_flag._precision_plugin
            if self._strategy_flag._checkpoint_io:
                if self.checkpoint_io:
                    raise MisconfigurationException(
                        "checkpoint_io set through both strategy class and plugins, choose one"
                    )
                self.checkpoint_io = self._strategy_flag._checkpoint_io
            if getattr(self._strategy_flag, "cluster_environment", None):
                if self._cluster_environment_flag:
                    raise MisconfigurationException(
                        "cluster_environment set through both strategy class and plugins, choose one"
                    )
                self._cluster_environment_flag = getattr(self._strategy_flag, "cluster_environment")

            if hasattr(self._strategy_flag, "parallel_devices") and self._strategy_flag.parallel_devices:
                if self._strategy_flag.parallel_devices[0].type == "cpu":
                    if self._accelerator_flag and self._accelerator_flag not in ("auto", "cpu"):
                        raise MisconfigurationException(
                            f"CPU parallel_devices set through {self._strategy_flag.__class__.__name__} class,"
                            f" but accelerator set to {self._accelerator_flag}, please choose one device type"
                        )
                    self._accelerator_flag = "cpu"

View on GitHub (pinned to 9fed5c27d2)

Solutions

1. Keep checkpoint IO in exactly one place: either on the strategy instance or in plugins/Trainer argument
2. Audit helper functions that build strategies for hidden checkpoint_io defaults

Example fix

# before
strategy = DDPStrategy(checkpoint_io=AsyncCheckpointIO())
trainer = Trainer(strategy=strategy, plugins=[AsyncCheckpointIO()])
# after
strategy = DDPStrategy(checkpoint_io=AsyncCheckpointIO())
trainer = Trainer(strategy=strategy)

from lightning.pytorch.plugins import CheckpointIO
if isinstance(strategy, Strategy) and strategy._checkpoint_io:
    plugins = [p for p in plugins if not isinstance(p, CheckpointIO)]
    checkpoint_io = None
trainer = Trainer(strategy=strategy, plugins=plugins, checkpoint_io=checkpoint_io)

def checkpoint_io_conflict(strategy, plugins) -> bool:
    from lightning.pytorch.plugins import CheckpointIO
    return bool(getattr(strategy, "_checkpoint_io", None)) and any(isinstance(p, CheckpointIO) for p in (plugins or []))

When it happens

Trigger: Trainer(strategy=DDPStrategy(checkpoint_io=MyCheckpointIO()), plugins=[MyCheckpointIO()]) or the checkpoint_io Trainer argument combined with a pre-configured strategy.

Common situations: Using custom checkpoint IO (e.g. for cloud storage) configured both in a shared strategy builder and in the Trainer call site.

Related errors

• accelerator set through both strategy class and accelerator
• precision set through both strategy class and plugins, choos
• cluster_environment set through both strategy class and plug
• The `{type(self).__name__}` does not use the `CheckpointIO`
• You selected an invalid strategy name: `strategy={strategy!r

AI-assisted analysis of Lightning-AI/pytorch-lightning@9fed5c27d2 (2026-08-28). Data as JSON: /api/errors/2fa35402bed91a48. Report an issue: GitHub.