Lightning-AI/pytorch-lightning · error · RuntimeError

f"Logging inside `{fx_name}` is not implemented." " Please,

Error message

f"Logging inside `{fx_name}` is not implemented." " Please, open an issue in `https://github.com/Lightning-AI/pytorch-lightning/issues`."

What it means

The _FxValidator maintains a whitelist of hooks where self.log() is permitted (e.g. training_step, validation_step). check_logging raises a RuntimeError when the hook name is entirely absent from that registry, meaning Lightning hit a logging call in a code path it does not account for — this is effectively an internal invariant/Lightning bug rather than a user config error.

Source

Thrown at src/lightning/pytorch/trainer/connectors/logger_connector/fx_validator.py:155

        "predict_step": None,
        "configure_optimizers": None,
        "train_dataloader": None,
        "val_dataloader": None,
        "test_dataloader": None,
        "prepare_data": None,
        "configure_callbacks": None,
        "on_validation_model_zero_grad": None,
        "on_validation_model_eval": None,
        "on_test_model_eval": None,
        "on_validation_model_train": None,
        "on_test_model_train": None,
    }

    @classmethod
    def check_logging(cls, fx_name: str) -> None:
        """Check if the given hook is allowed to log."""
        if fx_name not in cls.functions:
            raise RuntimeError(
                f"Logging inside `{fx_name}` is not implemented."
                " Please, open an issue in `https://github.com/Lightning-AI/pytorch-lightning/issues`."
            )

        if cls.functions[fx_name] is None:
            raise MisconfigurationException(
                f"You can't `self.log()` inside `{fx_name}`. HINT: You can still log directly to the logger by using"
                " `self.logger.experiment`."
            )

    @classmethod
    def get_default_logging_levels(
        cls, fx_name: str, on_step: Optional[bool], on_epoch: Optional[bool]
    ) -> tuple[bool, bool]:
        """Return default logging levels for given hook."""
        fx_config = cls.functions[fx_name]
        assert fx_config is not None
        on_step = fx_config["default_on_step"] if on_step is None else on_step

View on GitHub (pinned to 9fed5c27d2)

Solutions

1. Update/align Lightning versions so internal hook names match the registry (pip install -U pytorch-lightning)
2. If you monkey-patched or subclassed Trainer loops, stop passing custom fx_name values into the logging API; use standard hooks like training_step
3. Open an issue at https://github.com/Lightning-AI/pytorch-lightning/issues with a reproduction, as the message requests
4. As a workaround, log directly via self.logger.experiment instead of self.log()

from lightning.pytorch.trainer.connectors.logger_connector.fx_validator import _FxValidator
assert hook_name in _FxValidator.functions, f"unregistered hook {hook_name}"

try:
    self.log(name, value)
except RuntimeError as e:
    if "not implemented" in str(e):
        self.logger.experiment.log_metric(name, value)  # fallback
    else:
        raise

When it happens

Trigger: self.log(...) executing in a context whose fx_name is not a key in _FxValidator.functions, typically from a custom loop/callback triggering logging under an unrecognized hook name, or after renaming/adding hooks in a fork or outdated monkey-patch of Lightning internals.

Common situations: Upgrading PyTorch Lightning where hook names changed while a custom subclass/monkey-patch still passes old names; calling the logging result machinery manually with an arbitrary fx_name string; forks of Lightning adding new hooks without registering them.

Related errors

• you tried to log {v} which is currently not supported. Try a
• You are trying to `self.log()` but the loop's result collect
• You are trying to `self.log()` but it is not managed by the
• ReduceLROnPlateau conditioned on metric {monitor_key} which
• f"You can't `self.log()` inside `{fx_name}`. HINT: You can s

AI-assisted analysis of Lightning-AI/pytorch-lightning@9fed5c27d2 (2026-08-28). Data as JSON: /api/errors/3848e878c3be7276. Report an issue: GitHub.