add api

app.py

@@ -12,7 +12,7 @@ os.environ['CUDA_VISIBLE_DEVICES'] = "0"
 MODEL_NAME = "RegularizedSelfPlay/sppo_forward1reverse5-0.1-PromptABC-Mistral-7B-Instruct-SPPO-Iter3"  # Example: "meta-llama/Llama-2-7b-chat-hf"
 
 # Load model and tokenizer
-tokenizer = LlamaTokenizer.from_pretrained("mistralai/Mistral-7B-Instruct-v0.2")
+tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-Instruct-v0.2")
 llm = LLM(
         model=MODEL_NAME, 
         # revision="1296dc8fd9b21e6424c9c305c06db9ae60c03ace",