diff --git a/litellm/proxy/hooks/tag_rate_limiter.py b/litellm/proxy/hooks/tag_rate_limiter.py index ab810ae836a..01876511b03 100644 --- a/litellm/proxy/hooks/tag_rate_limiter.py +++ b/litellm/proxy/hooks/tag_rate_limiter.py @@ -473,7 +473,7 @@ class _PendingConcurrencyKeys: __slots__ = ("keys",) def __init__(self) -> None: - self.keys: list[tuple[str, "_PartitionKey"]] = [] # mutable-ok: shared across forks by design; see docstring + self.keys: list[tuple[str, _PartitionKey]] = [] # mutable-ok: shared across forks by design; see docstring _pending_concurrency_keys: Final[contextvars.ContextVar[_PendingConcurrencyKeys | None]] = contextvars.ContextVar(