fix acc

src/leaderboard/read_evals.py

@@ -80,7 +80,7 @@ class EvalResult:
             if accs.size == 0 or any([acc is None for acc in accs]):
                 continue
 
-            mean_acc = np.mean(accs) * 100.0
+            mean_acc = np.mean(accs) if len(accs) > 0 else 0
             results[task.benchmark] = mean_acc
 
         return self(

src/populate.py

@@ -13,6 +13,7 @@ def get_leaderboard_df(results_path: str, requests_path: str, dynamic_path: str,
     """Creates a dataframe from all the individual experiment results"""
     raw_data = get_raw_eval_results(results_path, requests_path, dynamic_path)
     all_data_json = [v.to_dict() for v in raw_data]
+    print("all_data_json", all_data_json)
     #all_data_json.append(baseline_row)
     filter_models_flags(all_data_json)