update scripts

src/display/utils.py

@@ -29,7 +29,11 @@ auto_eval_column_dict.append(["model_type_symbol", ColumnContent,
 auto_eval_column_dict.append(["model", ColumnContent,
                             ColumnContent("Model", "markdown", True, never_hidden=True)])
 for task in Tasks:
-    auto_eval_column_dict.append([task.name, ColumnContent, ColumnContent(task.value.col_name, "number", True)])
+    if task.value.col_name == "Overall Humanlike %" or task.value.col_name == "Overall CI":
+        auto_eval_column_dict.append([task.name, ColumnContent, ColumnContent(task.value.col_name, "number", True)])
+    else:
+        auto_eval_column_dict.append([task.name, ColumnContent, ColumnContent(task.value.col_name, "number", False)])
+
 #
 # Model information
 auto_eval_column_dict.append(["model_type", ColumnContent, ColumnContent("Type", "str", False)])
@@ -95,7 +99,7 @@ class WeightType(Enum):
 
 class Precision(Enum):
     float16 = ModelDetails("float16")
-    bfloat16 = ModelDetails("bfloat16")
+    bfloat16 = ModelDetails("BF16")
     qt_8bit = ModelDetails("8bit")
     qt_4bit = ModelDetails("4bit")
     qt_GPTQ = ModelDetails("GPTQ")

src/populate.py

@@ -16,7 +16,7 @@ def get_leaderboard_df(results_path: str, requests_path: str, cols: list, benchm
     
     #print(all_data_json)
     df = pd.DataFrame.from_records(all_data_json)
-    print("all results:",df)
+    print("all results:",df.to_string())
     # exit()
     try:
         df = df.sort_values(by=[utils.AutoEvalColumn.hallucination_rate.name], ascending=True)