update scripts

src/leaderboard/read_evals.py

@@ -75,8 +75,9 @@ class EvalResult:
 
             #print(task.value)
             if isinstance(task.metric, str):
-                accs = np.array([v.get(task.metric, None) for k, v in data["results"].items() if
-                                 task.benchmark == k and isinstance(v, dict)])
+                # accs = np.array([v.get(task.metric, None) for k, v in data["results"].items() if
+                #                  task.benchmark == k and isinstance(v, dict)])
+                accs = np.array([v for k, v in data["results"].items() if task.benchmark == k])
 
                 # 过滤掉 None 值，确保 accs 只包含有效的数值
                 accs = accs[accs != None]