remove chunking

app.py

@@ -13,6 +13,7 @@ def get_history_messages():
 
 
 def predict(prompt):
+    print("Predict:", prompt)
     pipe = pipeline("conversational", model="cognitivecomputations/TinyDolphin-2.8-1.1b")
     response = pipe(
         [
@@ -20,12 +21,9 @@ def predict(prompt):
             {"role": "user", "content": prompt}
         ],
     )
-    history.append((prompt, ""))
-    message = ""
-    for chunk in response:
-        message += chunk.messages[-1]["content"]
-        history[-1] = (prompt, message)
-        yield "", history
+    history.append((prompt, response.messages[-1]["content"])) 
+    print("Predict done")
+    return "", history
 
 with gradio.Blocks(fill_height=True) as demo:
     chat = gradio.Chatbot(scale=1)