Update app.py

app.py

@@ -565,8 +565,8 @@ def generate(message, temperature):
 
     # The reason why the library method is not used here is that if an error occurs,
     #    the returned data will not be a stream, and using the official library will result in an error.
-    for chunk in str(response).encode():
-        chunk = chunk.decode("utf-8")
+    for chunk in response:
+        chunk = chunk.decode("utf-8") if isinstance(chunk, bytes) else chunk
         data_chunks = parse_data_chunk(chunk)
         try:
             for data_chunk in data_chunks: