Lightning-AI/pytorch-lightning · error · ValueError

accelerator set through both strategy class and accelerator

Error message

accelerator set through both strategy class and accelerator flag, choose one

What it means

When a Strategy instance (e.g. DDPStrategy) is passed to Fabric and that instance already carries an `_accelerator`, the connector adopts it — but only if the separate `accelerator=` flag is still the default "auto". Setting both a strategy with an embedded accelerator and an explicit `accelerator=` flag is ambiguous and rejected.

Source

Thrown at src/lightning/fabric/connector.py:265

                raise ValueError(
                    f"Received multiple values for {', '.join(duplicated_plugin_key)} flags in `plugins`."
                    " Expected one value for each type at most."
                )

            if plugins_flags_types.get(Precision.__name__) and precision_input is not None:
                raise ValueError(
                    f"Received both `precision={precision_input}` and `plugins={self._precision_instance}`. Choose one."
                )

        self._precision_input = "32-true" if precision_input is None else precision_input

        # 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 isinstance(self._strategy_flag, Strategy):
            if self._strategy_flag._accelerator:
                if self._accelerator_flag != "auto":
                    raise ValueError("accelerator set through both strategy class and accelerator flag, choose one")
                self._accelerator_flag = self._strategy_flag._accelerator
            if self._strategy_flag._precision:
                # [RFC] handle precision plugin set up conflict?
                if self._precision_instance:
                    raise ValueError("precision set through both strategy class and plugins, choose one")
                self._precision_instance = self._strategy_flag._precision
            if self._strategy_flag._checkpoint_io:
                if self.checkpoint_io:
                    raise ValueError("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 ValueError("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"):

View on GitHub (pinned to 9fed5c27d2)

Solutions

1. Remove the `accelerator=` flag from Fabric(...) and let the strategy decide
2. Or construct the Strategy without an accelerator (e.g. DDPStrategy() ) and pass accelerator= separately

Example fix

# before
strategy = DDPStrategy(accelerator=CUDAAccelerator())
fabric = Fabric(strategy=strategy, accelerator="gpu")

# after
strategy = DDPStrategy(accelerator=CUDAAccelerator())
fabric = Fabric(strategy=strategy)

from lightning.fabric.strategies import Strategy

if isinstance(strategy, Strategy) and strategy._accelerator and accelerator != "auto":
    raise SystemExit("Strategy already has an accelerator; drop the accelerator flag")

When it happens

Trigger: strategy = DDPStrategy(accelerator=CUDAAccelerator()); Fabric(strategy=strategy, accelerator="cpu") — any Strategy instance whose _accelerator is set combined with accelerator != "auto".

Common situations: Reusing a preconfigured strategy object across scripts that also set accelerator="gpu"/"cpu"; refactoring older code where accelerator was set in both places; sharing strategy instances from a config factory.

Related errors

• precision set through both strategy class and plugins, choos
• checkpoint_io set through both strategy class and plugins, c
• cluster_environment set through both strategy class and plug
• CPU parallel_devices set through {self._strategy_flag.__clas
• GPU parallel_devices set through {self._strategy_flag.__clas

AI-assisted analysis of Lightning-AI/pytorch-lightning@9fed5c27d2 (2026-08-28). Data as JSON: /api/errors/020dab32a7c2f8b4. Report an issue: GitHub.