Update app.py

app.py

@@ -16,14 +16,14 @@ else:
 @spaces.GPU
 def generate_response(passage: str, question: str) -> str:
     # Prepare the input text by combining the passage and question
-    message = [f"Passage: {passage}\nQuestion: {question}\nAnswer: "]
+    message = [f"Passage: {passage}\nQuestion: {question}\nAnswer:"]
     inputs = tokenizer(message, return_tensors='pt', return_token_type_ids=False).to('cuda')
 
     response = model.generate(**inputs, max_new_tokens=100)
     
     response = tokenizer.batch_decode(response, skip_special_tokens=True)[0]
 
-    response = response[len(message[0]):].strip()
+    response = response[len(message[0]):].strip().split('\n')[0]
     
     return response