s

src/leaderboard/read_evals.py

@@ -108,6 +108,7 @@ class EvalResult:
     def to_dict(self):
         """Converts the Eval Result to a dict compatible with our dataframe display"""
         average = sum([v for v in self.results.values() if v is not None]) / len(Tasks)
+        print('average', average)
         data_dict={}
         # data_dict = {
         #     "eval_name": self.eval_name,  # not a column, just a save name,
@@ -198,7 +199,10 @@ class EvalResult:
 
 
         for task in Tasks:
-            data_dict[task.value.col_name] = self.results[task.value.benchmark]
+            try:
+                data_dict[task.value.col_name] = self.results[task.value.benchmark]
+            except KeyError:
+                print(f"Could not find {task.value.col_name}")
 
         return data_dict