Lightning-AI/pytorch-lightning · error · KeyError

Found invalid table_kwargs key

Error message

Found invalid table_kwargs key: {key}. Should be within {valid_table_keys}.

What it means

After excluding 'sort_by' and 'row_limit', remaining table_kwargs keys are validated against the signature of torch.autograd.EventList.table. Any key not in that signature raises KeyError, since it would be silently ignored or crash later when passed through.

Solutions

1. Inspect allowed keys: import inspect, torch; print(inspect.signature(torch.autograd.EventList.table).parameters.keys()) and use only those
2. Upgrade/downgrade Lightning or PyTorch so the kwarg set matches what you pass
3. Drop the unsupported key if it's not essential

Example fix

# before
profiler = PyTorchProfiler(table_kwargs={"n_cols": 5})  # invalid key

# after
profiler = PyTorchProfiler(table_kwargs={"max_src_column_width": 100})  # real EventList.table kwarg

import inspect, torch
valid = set(inspect.signature(torch.autograd.EventList.table).parameters) - {"self", "sort_by", "row_limit"}
bad = set(table_kwargs) - valid
assert not bad, f"invalid table_kwargs: {bad}"

def valid_table_kwargs(kwargs: dict) -> dict:
    import inspect, torch
    valid = set(inspect.signature(torch.autograd.EventList.table).parameters) - {"self", "sort_by", "row_limit"}
    return {k: v for k, v in kwargs.items() if k in valid}

When it happens

Trigger: PyTorchProfiler(table_kwargs={"max_group_row_limit": 5}) where the key is not a parameter of EventList.table in your installed PyTorch version.

Common situations: Using kwargs valid for a different PyTorch version's EventList.table (signature changed across releases); guessing kwarg names from other profiling APIs.

Understand the failure class

Background: "Must be a positive integer", "Invalid value", "Unsupported": the invalid-argument-value error family, when a library rejects the value you pass — this error's family across 35 libraries.

Related errors

• Found invalid table_kwargs key
• Found sort_by_key: . Should be within .
• You requested to find
• Attempting to stop recording an action
• f"`gradient_clip_algorithm

AI-assisted analysis of Lightning-AI/pytorch-lightning@9fed5c27d2 (2026-08-28). Data as JSON: /api/errors/b877a1c8de6b4132. Report an issue: GitHub.

Appendix: source

Thrown at src/lightning/pytorch/profilers/pytorch.py:333

            self._init_kineto(profiler_kwargs)

        if self._sort_by_key not in self.AVAILABLE_SORT_KEYS:
            raise MisconfigurationException(
                f"Found sort_by_key: {self._sort_by_key}. Should be within {self.AVAILABLE_SORT_KEYS}. "
            )

        for key in self._table_kwargs:
            if key in {"sort_by", "row_limit"}:
                raise KeyError(
                    f"Found invalid table_kwargs key: {key}. This is already a positional argument of the Profiler."
                )
            valid_table_keys = set(inspect.signature(EventList.table).parameters.keys()) - {
                "self",
                "sort_by",
                "row_limit",
            }
            if key not in valid_table_keys:
                raise KeyError(f"Found invalid table_kwargs key: {key}. Should be within {valid_table_keys}.")

    def _init_kineto(self, profiler_kwargs: Any) -> None:
        has_schedule = "schedule" in profiler_kwargs
        self._has_on_trace_ready = "on_trace_ready" in profiler_kwargs

        schedule = profiler_kwargs.get("schedule", None)
        if schedule is not None:
            if not callable(schedule):
                raise MisconfigurationException(f"Schedule should be a callable. Found: {schedule}")
            action = schedule(0)
            if not isinstance(action, ProfilerAction):
                raise MisconfigurationException(
                    f"Schedule should return a `torch.profiler.ProfilerAction`. Found: {action}"
                )
        self._default_schedule()
        schedule = schedule if has_schedule else self._default_schedule()
        self._schedule = ScheduleWrapper(schedule) if schedule is not None else schedule
        self._profiler_kwargs["schedule"] = self._schedule

View on GitHub (pinned to 9fed5c27d2)