Update utils.py

utils.py

@@ -389,11 +389,10 @@ def query(api_llm, payload):
 def llm_chain2(prompt, context):
     full_prompt = RAG_CHAIN_PROMPT.format(context=context, question=prompt)
     inputs = tokenizer_rag(full_prompt, return_tensors="pt", max_length=1024, truncation=True)
-    attention_mask = (inputs != tokenizer_rag.pad_token_id).long()
     #Generiere die Antwort
     outputs = modell_rag.generate(
         inputs.input_ids, 
-        attention_mask=attention_mask,
+        attention_mask=inputs.attention_mask,
         max_new_tokens=1024, 
         do_sample=True, 
         temperature=0.9,