transformer_rankers.eval.evaluation.evaluate_models

transformer_rankers.eval.evaluation.evaluate_models(results)[source]

Calculate METRICS for each model in the results dict

Parameters
  • results – dict containing one key for each model and inside them pred and label keys.

  • example (For) –

    results = {
    ‘model_1’: {

    ‘preds’: [[1,2],[1,2]], ‘labels’: [[1,2],[1,2]]

    }

    }.

Returns: dict with the METRIC results per model and query.