Update app.py

app.py

@@ -32,8 +32,7 @@ def respond(
         yield "Error: No model loaded. Please load a model first."
         return
 
-    response = ""
-    yield " "
+    yield "..."
     
     generation_kwargs = dict(
         max_length=8192,
@@ -50,7 +49,8 @@ def respond(
         messages = [chatglm_cpp.ChatMessage(role="system", content=system_message)]
 
     messages.append(chatglm_cpp.ChatMessage(role="user", content=message))
-    
+
+    response = ""
     for chunk in pipeline.chat(messages, **generation_kwargs):
         response += chunk.content
         yield response