evidence

Identity contracts for post-MIP evaluation evidence.

Functions

`checkpoint_fingerprint`	Return the checkpoint fingerprint used by evaluation identities.
`collect_kd_exposure`	Combine configured exposure with durable token and GPU-hour evidence.
`downstream_evaluation_identity`	Bind one comparison to its checkpoints, training, data, and evaluator.
`evaluation_contract`	Validate and project the fields that must match across evaluations.
`kd_exposure_metrics`	Publish the comparable numeric exposure fields as node metrics.

checkpoint_fingerprint(checkpoint)

Return the checkpoint fingerprint used by evaluation identities.

Parameters:: checkpoint (str | Path)
Return type:: str

collect_kd_exposure(output, configured, *, max_steps, elapsed_gpu_hours, resumed_completed_milestone)

Combine configured exposure with durable token and GPU-hour evidence.

Parameters:

output (Path)
configured (Mapping[str, Any])
max_steps (int)
elapsed_gpu_hours (float)
resumed_completed_milestone (bool)

Return type:

dict[str, Any]

downstream_evaluation_identity(*, source, reference_checkpoint, profile, evaluator_revision, settings, candidate, reference_checkpoint_fingerprint=None)

Bind one comparison to its checkpoints, training, data, and evaluator.

Parameters:

source (CandidateRevision)
reference_checkpoint (str | Path)
profile (Any)
evaluator_revision (Any)
settings (Mapping[str, Any])
candidate (Mapping[str, Any])
reference_checkpoint_fingerprint (str | None)

Return type:

dict[str, Any]

evaluation_contract(identity, *, label, expected_profile, expected_manifest_sha256, expected_reference_fingerprint)

Validate and project the fields that must match across evaluations.

Parameters:

identity (Any)
label (str)
expected_profile (str)
expected_manifest_sha256 (str)
expected_reference_fingerprint (str)

Return type:

dict[str, Any]

kd_exposure_metrics(exposure)

Publish the comparable numeric exposure fields as node metrics.

Parameters:: exposure (Mapping[str, Any])
Return type:: dict[str, float]