Update README.md

README.md

@@ -100,7 +100,13 @@ reranker = pipeline(
 )
 
 similarity = reranker(
-    [dict(text=ii, text_pair=query) for ii in context_list]
+    [
+        dict(
+            text=ii,        # the model was trained with context in text argument
+            text_pair=query # and query in text_pair argument.
+        )
+        for ii in context_list
+    ]
 )
 context_reranked = sorted(
     filter(