first commit

app.py

@@ -46,8 +46,8 @@ def color_model_type_column(df, color_map):
         return f'background-color: {color}'
     
     # Format for different columns
-    format_dict = {col: "{:.2f}" for col in df.columns if col not in ['Average', 'Model', 'Model Type']}
-    format_dict['Average'] = "{:.2f}"
+    format_dict = {col: "{:.2f}" for col in df.columns if col not in ['Avg.', 'Model', 'Model Type']}
+    format_dict['Avg.'] = "{:.2f}"
     format_dict[''] = "{:d}"
 
     return df.style.applymap(apply_color, subset=['Model Type']).format(format_dict, na_rep='')
@@ -96,7 +96,7 @@ repo = snapshot_download(
 with zipfile.ZipFile(os.path.join(eval_set_dir, 'videos.zip'), 'r') as zip_ref:
     zip_ref.extractall(eval_set_dir)
 
-rewardbench_data = load_all_data(eval_results_dir).sort_values(by='Average', ascending=False)
+rewardbench_data = load_all_data(eval_results_dir).sort_values(by='Avg.', ascending=False)
 col_types_rewardbench = ["number"] + ["markdown"]+ ["str"] + ["number"] * (len(rewardbench_data.columns) - 1)
 # for showing random samples
 eval_set = pd.read_csv(os.path.join(eval_set_dir, 'videogen-rewardbench.csv'))

src/utils.py

@@ -33,9 +33,9 @@ def load_all_data(data_dir):
                     data_list.append({
                         "Model": model_hyperlink(model_link, model),
                         "Model Type": data.get("model_type", ""),
-                        "Average": avg,
-                        "Average (w/o Ties)": avg_without_tie,
-                        "Average (w/ Ties)": avg_with_tie,
+                        "Avg.": avg,
+                        "Avg. (w/o Ties)": avg_without_tie,
+                        "Avg. (w/ Ties)": avg_with_tie,
                         "Overall (w/o Ties)": round(without_tie.get("overall", 0), 2),
                         "VQ (w/o Ties)": round(without_tie.get("vq", 0), 2),
                         "MQ (w/o Ties)": round(without_tie.get("mq", 0), 2),