aip_sdk.MetricConfigRef
aip_sdk.MetricConfigRef
A stored, versioned MetricConfig a metric runs with, named by that config's own identity.
(config_name, config_version) is the storage layer's key for a config row — independent of
the metric it configures, so toxicity-strict and toxicity-lenient can both configure
llm.toxicity and either is pinnable. Resolved within one workspace, which owns every stored
config and is why no scope is named here. Lengths
match MetricConfigRecord.config_name/config_version, so what passes here is what the
lookup could have stored. frozen to stay hashable, as the dataclass this replaced was.
The pin names a version; it does not freeze one. A stored config's params and prompts
can be rewritten in place at the same key until some run has scored against it, so bump
config_version for a content change wherever reproducibility is the point of pinning.
aip_sdk.MetricConfigRef.config_name
aip_sdk.MetricConfigRef.config_name: str = Field(min_length=1, max_length=255)
No docstring is defined in the source.
aip_sdk.MetricConfigRef.config_version
aip_sdk.MetricConfigRef.config_version: str = Field(min_length=1, max_length=64)
No docstring is defined in the source.
aip_sdk.MetricConfigRef.model_config
aip_sdk.MetricConfigRef.model_config = ConfigDict(extra='forbid', frozen=True)
No docstring is defined in the source.