Update app.py

app.py

@@ -11,7 +11,7 @@ model_path = 'THUDM/BPO'
 device = 'cuda'
 
 tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True, add_prefix_space=True)
-model = AutoModelForCausalLM.from_pretrained(model_path, trust_remote_code=True).to(device)
+model = AutoModelForCausalLM.from_pretrained(model_path, trust_remote_code=True, load_in_8bit=True).to(device)
 model = model.eval()
 
 prompt_template = "[INST] You are an expert prompt engineer. Please help me improve this prompt to get a more helpful and harmless response:\n{} [/INST]"