Lightning-AI/pytorch-lightning · critical · ModuleNotFoundError

raise ModuleNotFoundError(str(_XLA_AVAILABLE))

Error message

raise ModuleNotFoundError(str(_XLA_AVAILABLE))

What it means

SingleDeviceXLAStrategy (single-process XLA) requires torch_xla. Its __init__ raises ModuleNotFoundError with Lightning's _XLA_AVAILABLE message when the package is missing, before any XLA setup is attempted.

Source

Thrown at src/lightning/pytorch/strategies/single_xla.py:45

from lightning.pytorch.plugins.precision.xla import XLAPrecision
from lightning.pytorch.strategies.single_device import SingleDeviceStrategy
from lightning.pytorch.trainer.states import TrainerFn
from lightning.pytorch.utilities import find_shared_parameters, set_shared_parameters


class SingleDeviceXLAStrategy(SingleDeviceStrategy):
    """Strategy for training on a single XLA device."""

    def __init__(
        self,
        device: _DEVICE,
        accelerator: Optional["pl.accelerators.Accelerator"] = None,
        checkpoint_io: Optional[Union[XLACheckpointIO, _WrappingCheckpointIO]] = None,
        precision_plugin: Optional[XLAPrecision] = None,
        debug: bool = False,
    ):
        if not _XLA_AVAILABLE:
            raise ModuleNotFoundError(str(_XLA_AVAILABLE))
        if isinstance(device, torch.device):
            # unwrap the `torch.device` in favor of `xla_device`
            device = device.index
        import torch_xla.core.xla_model as xm

        super().__init__(
            accelerator=accelerator,
            device=xm.xla_device(device),
            checkpoint_io=checkpoint_io,
            precision_plugin=precision_plugin,
        )
        self.debug = debug

    @property
    @override
    def checkpoint_io(self) -> Union[XLACheckpointIO, _WrappingCheckpointIO]:
        plugin = self._checkpoint_io
        if plugin is not None:

View on GitHub (pinned to 9fed5c27d2)

Solutions

1. pip install lightning[xla] or a torch_xla wheel matching your torch/torchvision versions
2. Verify import torch_xla succeeds
3. If XLA wasn't intended, use SingleDeviceStrategy (CPU/GPU) instead

Example fix

# before
strategy = SingleDeviceXLAStrategy()  # ModuleNotFoundError

# after
# pip install lightning[xla]
strategy = SingleDeviceXLAStrategy()

from lightning.fabric.utilities.imports import _XLA_AVAILABLE
assert _XLA_AVAILABLE, "pip install lightning[xla] before using SingleDeviceXLAStrategy"

When it happens

Trigger: Instantiating SingleDeviceXLAStrategy (e.g. accelerator='tpu' with devices=1, or explicitly) without torch_xla installed; environment with plain torch only.

Common situations: Trying single-TPU/GPU-with-XLA setups locally; CI environments lacking the xla extra.

Understand the failure class

Background: "X is not installed. Please install it with pip install Y": missing optional dependency errors — ImportError/ValueError raised when a library's optional extra was never installed — this error's family across 22 libraries.

Related errors

• str(_XLA_AVAILABLE)
• {str(_XLA_AVAILABLE)}
• {str(_XLA_AVAILABLE)}
• raise ModuleNotFoundError(str(_XLA_AVAILABLE))
• raise ModuleNotFoundError(str(_XLA_AVAILABLE))

AI-assisted analysis of Lightning-AI/pytorch-lightning@9fed5c27d2 (2026-08-28). Data as JSON: /api/errors/1da873629de8006f. Report an issue: GitHub.