Update utils.py

utils.py

@@ -12,14 +12,15 @@ SUBJECTS = ["Biology", "Business", "Chemistry", "Computer Science", "Economics",
             "Health", "History", "Law", "Math", "Philosophy", "Physics", "Psychology", "Other"]
 
 MODEL_INFO = [
-    "Models", "Data Source",
+    "Models", "Model Size(B)", "Data Source",
     "Overall",
     "Biology", "Business", "Chemistry", "Computer Science", "Economics", "Engineering",
     "Health", "History", "Law", "Math", "Philosophy", "Physics", "Psychology", "Other"]
 
-DATA_TITLE_TYPE = ['markdown', 'markdown', 'number', 'number', 'number', 'number', 'number', 'number',
+DATA_TITLE_TYPE = ['markdown', 'str', 'markdown', 'number',
+                   'number', 'number', 'number', 'number', 'number', 'number',
                    'number', 'number', 'number', 'number', 'number', 'number', 'number',
-                   'number', 'number']
+                   'number']
 
 SUBMISSION_NAME = "mmlu_pro_leaderboard_submission"
 SUBMISSION_URL = os.path.join("https://huggingface.co/datasets/TIGER-Lab/", SUBMISSION_NAME)
@@ -148,16 +149,14 @@ def refresh_data():
 def search_and_filter_models(df, query, min_size, max_size):
     if query:
         df = df[df['Models'].str.contains(query, case=False, na=False)]
-
+    
     numeric_mask = df['Model Size(B)'].apply(lambda x: isinstance(x, (int, float)))
     size_filtered = df[numeric_mask & 
-                         (df['Model Size(B)'] >= min_size) & 
-                         (df['Model Size(B)'] <= max_size)]
+                      (df['Model Size(B)'] >= min_size) & 
+                      (df['Model Size(B)'] <= max_size)]
     unknown_entries = df[df['Model Size(B)'] == 'unknown']
-    return pd.concat([size_filtered, unknown_entries])[COLUMN_NAMES]
-    # df = df[(df['Model Size(B)'] >= min_size) & (df['Model Size(B)'] <= max_size)]
     
-    # return df[COLUMN_NAMES]
+    return pd.concat([size_filtered, unknown_entries])[COLUMN_NAMES]
 
 
 def search_models(df, query):
@@ -167,10 +166,10 @@ def search_models(df, query):
 
 
 def get_size_range(df):
-    sizes = df['Model Size(B)'].dropna()
-    if len(sizes) > 0:
-        return float(sizes.min()), float(sizes.max())
-    return 0, 1000 
+    numeric_sizes = df[df['Model Size(B)'].apply(lambda x: isinstance(x, (int, float)))]['Model Size(B)']
+    if len(numeric_sizes) > 0:
+        return float(numeric_sizes.min()), float(numeric_sizes.max())
+    return 0, 1000
 
 
 def process_model_size(size):