agrag.eval.verifier.verdict_report

verdict_report(gold:Sequence[str], predicted:Sequence[str]) -> VerdictReport

Score predicted verdicts against gold verdicts.

Parameters:

• gold (Sequence[str]) – The gold label of each item.
• predicted (Sequence[str]) – The label of each item from run_verifier.

Returns:

• VerdictReport – Per-class scores, macro F1, the confusion matrix and the error count.