Update app.py

app.py

@@ -389,6 +389,8 @@ def create_prompt(question, passage):
     return prompt.format(passage=passage, question=question)
 
 def generate_answer(prompt, max_length=860, temperature=0.2):
+    tokenizer_f = models['llm_tokenizer']
+    model_f = models['llm_model']
     inputs = tokenizer_f(prompt, return_tensors="pt", truncation=True)
     # Start timing
     start_time = time.time()