Log benchmark name

benchmarks/benchmark.py

@@ -106,6 +106,7 @@ def is_trained_model_valid(self, test_acc):
 
     def run(self):
         print(f"Using device {self.device}")
+        benchmark_name = self.__class__.__name__
         all_explanations = defaultdict(list)
         all_runtimes = defaultdict(list)
         for experiment_i in tq(range(self.sample_count)):
@@ -142,7 +143,7 @@ def time_wrapper(*args, **kwargs):
 
                 time_wrapper.explain_function = explain_function
                 accs = self.evaluate_explanation(time_wrapper, model, test_dataset, explain_name)
-                print(f'Run #{experiment_i + 1}, Explain Method: {explain_name}, Accuracy: {np.mean(accs)}')
+                print(f'Benchmark:{benchmark_name} Run #{experiment_i + 1}, Explain Method: {explain_name}, Accuracy: {np.mean(accs)}')
                 all_explanations[explain_name].append(list(accs))
                 all_runtimes[explain_name].extend(duration_samples)
                 metrics = {
@@ -154,7 +155,7 @@ def time_wrapper(*args, **kwargs):
                     json.dump(all_explanations, open(file_path, 'w'), indent=2)
                     mlflow.log_artifact(file_path)
                 mlflow.log_metrics(metrics, step=experiment_i)
-            print(f'Run #{experiment_i + 1} finished. Average Explanation Accuracies for each method:')
+            print(f'Benchmark:{benchmark_name} Run #{experiment_i + 1} finished. Average Explanation Accuracies for each method:')
             accuracies_summary = {}
             for name, run_accs in all_explanations.items():
                 run_accs = [np.mean(single_run_acc) for single_run_acc in run_accs]