Update space

app.py

@@ -105,8 +105,8 @@ def init_leaderboard(dataframe):
 # model_result_path = "./src/results/models_2024-10-08-17:39:21.001582.jsonl"
 # model_result_path = "./src/results/models_2024-10-09-05:17:38.810960.json"
 # model_result_path = "./src/results/models_2024-10-09-06:22:21.122422.json"
-# model_result_path = "./src/results/models_2024-10-10-06:18:54.263527.json"
-model_result_path = "./src/results/models_2024-10-18-14:06:13.588399.json"
+model_result_path = "./src/results/models_2024-10-10-06:18:54.263527.json"
+# model_result_path = "./src/results/models_2024-10-18-14:06:13.588399.json"
 # model_leaderboard_df = get_model_leaderboard_df(model_result_path)
 
 
@@ -170,13 +170,13 @@ with demo:
                         # AutoEvalColumn.rank_overall.name,
                         AutoEvalColumn.model.name, 
                         AutoEvalColumn.rank_overall.name,
-                        # AutoEvalColumn.rank_math_algebra.name,
-                        # AutoEvalColumn.rank_math_geometry.name,
+                        AutoEvalColumn.rank_math_algebra.name,
+                        AutoEvalColumn.rank_math_geometry.name,
                         AutoEvalColumn.rank_math_probability.name,
                         AutoEvalColumn.rank_reason_logical.name,
-                        # AutoEvalColumn.rank_reason_social.name,
+                        AutoEvalColumn.rank_reason_social.name,
                         AutoEvalColumn.rank_chemistry.name,
-                        # AutoEvalColumn.rank_cpp.name,
+                        AutoEvalColumn.rank_cpp.name,
                         ],
                     rank_col=[],
                 )
@@ -265,7 +265,7 @@ with demo:
                             AutoEvalColumn.rank_math_probability.name,
                             AutoEvalColumn.model.name, 
                             AutoEvalColumn.score_math_probability.name,
-                            # AutoEvalColumn.sd_math_probability.name,
+                            AutoEvalColumn.sd_math_probability.name,
                             AutoEvalColumn.license.name,
                             AutoEvalColumn.organization.name,
                             AutoEvalColumn.knowledge_cutoff.name,

src/populate.py

@@ -19,7 +19,7 @@ def get_model_leaderboard_df(results_path: str, requests_path: str="", cols: lis
     df = pd.DataFrame.from_records(all_data_json)
     
     df = df[benchmark_cols]
-    print(df.head())
+    # print(df.head())
     
     if rank_col: # if there is one col in rank_col, sort by that column and remove NaN values
         df = df.dropna(subset=benchmark_cols)
@@ -48,7 +48,8 @@ def get_model_leaderboard_df(results_path: str, requests_path: str="", cols: lis
         #         df[col] = (df[col]).map('{:.2f}'.format)
         #     else:   
         #         df[col] = (df[col]*100).map('{:.2f}'.format)
-            if "Chemistry" in col or "C++" in col or "Overall" in col or "Probability" in col or "Logical" in col:
+            if "Chemistry" in col or "C++" in col:
+            # if "Chemistry" in col or "C++" in col or "Overall" in col or "Probability" in col or "Logical" in col:
                 df[col] = (df[col]).map('{:.2f}'.format)
             else:
                 df[col] = (df[col]*100).map('{:.2f}'.format)