Update app.py

app.py

@@ -744,15 +744,15 @@ def get_mteb_average():
     # Debugging:
     # DATA_OVERALL.to_csv("overall.csv")
     
-    DATA_OVERALL.insert(1, f"Average ({len(TASK_LIST_EN)} datasets)", DATA_OVERALL[TASK_LIST_EN].mean(axis=1, skipna=False))
-    DATA_OVERALL.insert(2, f"Classification Average ({len(TASK_LIST_CLASSIFICATION)} datasets)", DATA_OVERALL[TASK_LIST_CLASSIFICATION].mean(axis=1, skipna=False))
-    DATA_OVERALL.insert(3, f"Clustering Average ({len(TASK_LIST_CLUSTERING)} datasets)", DATA_OVERALL[TASK_LIST_CLUSTERING].mean(axis=1, skipna=False))
-    DATA_OVERALL.insert(4, f"Pair Classification Average ({len(TASK_LIST_PAIR_CLASSIFICATION)} datasets)", DATA_OVERALL[TASK_LIST_PAIR_CLASSIFICATION].mean(axis=1, skipna=False))
-    DATA_OVERALL.insert(5, f"Reranking Average ({len(TASK_LIST_RERANKING)} datasets)", DATA_OVERALL[TASK_LIST_RERANKING].mean(axis=1, skipna=False))
-    DATA_OVERALL.insert(6, f"Retrieval Average ({len(TASK_LIST_RETRIEVAL)} datasets)", DATA_OVERALL[TASK_LIST_RETRIEVAL].mean(axis=1, skipna=False))
-    DATA_OVERALL.insert(7, f"STS Average ({len(TASK_LIST_STS)} datasets)", DATA_OVERALL[TASK_LIST_STS].mean(axis=1, skipna=False))
-    DATA_OVERALL.insert(8, f"Summarization Average ({len(TASK_LIST_SUMMARIZATION)} dataset)", DATA_OVERALL[TASK_LIST_SUMMARIZATION].mean(axis=1, skipna=False))
-    DATA_OVERALL.sort_values(f"Average ({len(TASK_LIST_EN)} datasets)", ascending=False, inplace=True)
+    DATA_OVERALL.insert(1, f"Average", DATA_OVERALL[TASK_LIST_EN].mean(axis=1, skipna=False))
+    DATA_OVERALL.insert(2, f"Classification Average", DATA_OVERALL[TASK_LIST_CLASSIFICATION].mean(axis=1, skipna=False))
+    DATA_OVERALL.insert(3, f"Clustering Average", DATA_OVERALL[TASK_LIST_CLUSTERING].mean(axis=1, skipna=False))
+    DATA_OVERALL.insert(4, f"Pair Classification Average", DATA_OVERALL[TASK_LIST_PAIR_CLASSIFICATION].mean(axis=1, skipna=False))
+    DATA_OVERALL.insert(5, f"Reranking Average", DATA_OVERALL[TASK_LIST_RERANKING].mean(axis=1, skipna=False))
+    DATA_OVERALL.insert(6, f"Retrieval Average", DATA_OVERALL[TASK_LIST_RETRIEVAL].mean(axis=1, skipna=False))
+    DATA_OVERALL.insert(7, f"STS Average", DATA_OVERALL[TASK_LIST_STS].mean(axis=1, skipna=False))
+    DATA_OVERALL.insert(8, f"Summarization Average", DATA_OVERALL[TASK_LIST_SUMMARIZATION].mean(axis=1, skipna=False))
+    DATA_OVERALL.sort_values(f"Average", ascending=False, inplace=True)
     # Start ranking from 1
     DATA_OVERALL.insert(0, "Rank", list(range(1, len(DATA_OVERALL) + 1)))
 
@@ -769,14 +769,12 @@ def get_mteb_average():
     # Fill NaN after averaging
     DATA_OVERALL.fillna("", inplace=True)
 
-    DATA_OVERALL = DATA_OVERALL[["Rank", "Model", "Model Size (GB)", "Embedding Dimensions", "Max Tokens", f"Average ({len(TASK_LIST_EN)} datasets)", f"Classification Average ({len(TASK_LIST_CLASSIFICATION)} datasets)", f"Clustering Average ({len(TASK_LIST_CLUSTERING)} datasets)", f"Pair Classification Average ({len(TASK_LIST_PAIR_CLASSIFICATION)} datasets)", f"Reranking Average ({len(TASK_LIST_RERANKING)} datasets)", f"Retrieval Average ({len(TASK_LIST_RETRIEVAL)} datasets)", f"STS Average ({len(TASK_LIST_STS)} datasets)", f"Summarization Average ({len(TASK_LIST_SUMMARIZATION)} dataset)"]]
+    DATA_OVERALL = DATA_OVERALL[["Rank", "Model", "Model Size (GB)", "Embedding Dimensions", "Max Tokens", f"Average", f"Classification Average", f"Clustering Average", f"Pair Classification Average", f"Reranking Average", f"Retrieval Average", f"STS Average", f"Summarization Average"]]
     DATA_OVERALL = DATA_OVERALL[DATA_OVERALL.iloc[:, 5:].ne("").any(axis=1)]
 
     return DATA_OVERALL
 
 DATA_OVERALL=get_mteb_average()
-print(DATA_OVERALL)
-print(DATA_OVERALL.columns)
                 
 import unicodedata
 
@@ -798,9 +796,28 @@ for column in DATA_OVERALL.columns:
             if DATA_OVERALL[column].dtype == 'object':
                 DATA_OVERALL[column] = DATA_OVERALL[column].apply(remove_invalid_unicode)
 
+DATA_OVERALL_COLUMN_TO_DATATYPE = [
+    "Rank", "number",
+    "Model Size (GB)", "number",
+    "Embedding Dimensions", "number",
+    "Max Tokens", "number",
+    "Average", "number",
+    "Classification Average", "number",
+    "Classification Average", "number",
+    "Pair Classification Average", "number",
+    "Reranking Average", "number",
+    "Retrieval Average", "number",
+    "STS Average", "number",
+    "Summarization Average", "number"
+]
+
+COLS = [col.name for col in DATA_OVERALL_COLUMN_TO_DATATYPE]
+TYPES = [col.type for col in DATA_OVERALL_COLUMN_TO_DATATYPE]
+
 data_overall = gr.components.Dataframe(
-    headers=DATA_OVERALL.columns.tolist(),
-    datatype=DATA_OVERALL.values.tolist(),
+    DATA_OVERALL,
+    headers=COLS,
+    datatype=TYPES,
     visible=False,
     line_breaks=False,
     interactive=False