Update app.py

app.py

@@ -77,11 +77,11 @@ def reply(message: str, history: list[str]) -> str:
     # Preprocess message
 
     pipe = transformers.pipeline(
-        "text2text-generation", model="Qwen/Qwen2.5-7B-Instruct", device="cuda", return_full_text=False
+        "text2text-generation", model="Qwen/Qwen2.5-7B-Instruct", device="cuda"
     )
 
     message = preprocess(message, PUBLICATIONS_TO_RETRIEVE)
-    return pipe(message, max_new_tokens=512)[0]["generated_text"]
+    return pipe(message, max_new_tokens=512, return_full_text=False)[0]["generated_text"]
 
 
 # Example Queries for Interface