Lightning-AI/pytorch-lightning · error · RuntimeError

The `{type(self._strategy).__name__}` requires the model and

Error message

The `{type(self._strategy).__name__}` requires the model and optimizer(s) to be set up jointly through `.setup(model, optimizer, ...)`.

What it means

DeepSpeedStrategy and XLAStrategy couple the model and optimizer into a single engine/session, so optimizers cannot be set up standalone. _validate_setup_optimizers raises RuntimeError for these strategies, directing you to the joint .setup(model, optimizer, ...) call.

Source

Thrown at src/lightning/fabric/fabric.py:1226

        if isinstance(self._strategy, FSDPStrategy) and any(
            _has_meta_device_parameters_or_buffers(optimizer) for optimizer in optimizers
        ):
            raise RuntimeError(
                "The optimizer has references to the model's meta-device parameters. Materializing them is"
                " is currently not supported unless you to set up the model and optimizer(s) separately."
                " Create and set up the model first through `model = fabric.setup_module(model)`. Then create the"
                " optimizer and set it up: `optimizer = fabric.setup_optimizers(optimizer)`."
            )

    def _validate_setup_module(self, module: nn.Module) -> None:
        self._validate_launched()
        if isinstance(module, _FabricModule):
            raise ValueError("A model should be passed only once to the `setup_module` method.")

    def _validate_setup_optimizers(self, optimizers: Sequence[Optimizer]) -> None:
        self._validate_launched()
        if isinstance(self._strategy, (DeepSpeedStrategy, XLAStrategy)):
            raise RuntimeError(
                f"The `{type(self._strategy).__name__}` requires the model and optimizer(s) to be set up jointly"
                " through `.setup(model, optimizer, ...)`."
            )

        if not optimizers:
            raise ValueError("`setup_optimizers` requires at least one optimizer as input.")

        if any(isinstance(opt, _FabricOptimizer) for opt in optimizers):
            raise ValueError("An optimizer should be passed only once to the `setup_optimizers` method.")

        if any(_has_meta_device_parameters_or_buffers(optimizer) for optimizer in optimizers):
            raise RuntimeError(
                "The optimizer has references to the model's meta-device parameters. Materializing them is"
                " is currently not supported. Create the optimizer after setting up the model, then call"
                " `fabric.setup_optimizers(optimizer)`."
            )

    def _validate_setup_dataloaders(self, dataloaders: Sequence[DataLoader]) -> None:

View on GitHub (pinned to 9fed5c27d2)

Solutions

1. Use the joint call: model, optimizer = fabric.setup(model, optimizer)
2. Branch your code: for deepspeed/XLA use joint setup, otherwise separate setup is fine

Example fix

# before
model = fabric.setup_module(model)
optimizer = fabric.setup_optimizers(optimizer)  # deepspeed
# after
model, optimizer = fabric.setup(model, optimizer)

from lightning.fabric.strategies import DeepSpeedStrategy, XLAStrategy
if isinstance(fabric.strategy, (DeepSpeedStrategy, XLAStrategy)):
    model, optimizer = fabric.setup(model, optimizer)
else:
    model = fabric.setup_module(model)
    optimizer = fabric.setup_optimizers(optimizer)

When it happens

Trigger: fabric = Fabric(strategy='deepspeed') then fabric.setup_optimizers(optimizer) after separately setting up the model — the separated flow is unsupported for these strategies.

Common situations: Writing strategy-agnostic code with separate setup_module/setup_optimizers calls, then switching strategy='deepspeed' or running on TPU/XLA.

Related errors

• No models were set up for backward. Did you forget to call `
• When using multiple models + deepspeed, please provide the m
• To spawn processes with the `{type(self.strategy).__name__}`
• A model should be passed only once to the `setup` method.
• An optimizer should be passed only once to the `setup` metho

AI-assisted analysis of Lightning-AI/pytorch-lightning@9fed5c27d2 (2026-08-28). Data as JSON: /api/errors/b0f25afbe5086d39. Report an issue: GitHub.