Update app.py

app.py

@@ -95,12 +95,12 @@ def chat_document(file, question):
     Question: {question}
     Answer:"""
 
-    stream = client.text_generation(message, max_new_tokens=2048, stream=True, details=True, return_full_text=False)
+    stream = client.text_generation(message, max_new_tokens=4096, stream=True, details=True, return_full_text=False)
     output = ""
     for response in stream:
         if not response.token.text == "</s>":
             output += response.token.text
-    return output
+        yield output
 
 
 with gr.Blocks() as demo: