AttackEvaluator classes give back empty dictionary

[foesine]

Hello,

I am facing problems with the AttackEvaluator class.
When running any of the inheriting classes, it always gives back an empty dict.

I am running:

evaluators = [DataLeakageLinear(), DataLeakageMLP(), DataLeakageXGB()]

for evaluator in evaluators:
  evaluator_name = type(evaluator).__name__
  score_dict[evaluator_name] = []

  for loader in loaders_syndata:
    score = evaluator.evaluate(X_gt=loader_cps, X_syn=loader)
    score_dict[evaluator_name].append(score)

DataLeakage_scores = pd.DataFrame(score_dict)
DataLeakage_scores = DataLeakage_scores.T
DataLeakage_scores.columns = [df.name for df in gen_data]
DataLeakage_scores.index.name = "Evaluator"

This produces a data frame which contains a bunch of empty dicts instead of scores.

Using the same loaders for other evaluation metrics gives me valid results, so I don't think they are the problem.

I'd be very thankful for suggestions or solutions!