Update app.py

app.py

@@ -119,7 +119,7 @@ def respond(
     messages.append({"role": "user", "content": message})
     input_ids = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt").to(accelerator.device) #.to('cuda')
     
-    input_ids2 = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True, return_tensors="pt").to(accelerator.device) #.to('cuda')
+    input_ids2 = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True, return_tensors="pt") #.to('cuda')
 
     input_str= str(input_ids2)
     print('input str = ', input_str)