Update app.py

app.py

@@ -54,8 +54,9 @@ def chat(history, system_message, max_tokens, temperature, top_p, top_k, repetit
     )
     for tokens in prediction:
         tokens = re.findall(r'(.*?)(\s|$)', tokens)
-        for s in tokens:
-            answer = s
+        for subtoken in tokens:
+            subtoken = "".join(subtoken)
+            answer = subtoken
             history[-1][1] += answer
             # stream the response
             yield history, history, ""