Update app.py

app.py

@@ -9,7 +9,7 @@ MODEL_NAME = "warleagle/medical_chat_saiga"
 config = PeftConfig.from_pretrained(MODEL_NAME)
 model = AutoModelForCausalLM.from_pretrained(
     config.base_model_name_or_path,
-    load_in_4bit=True,
+    load_in_8bit=True,
     torch_dtype=torch.float16,
     device_map="auto"
 )
@@ -42,7 +42,7 @@ def predict(input_data, temp):
     
     conversation = Conversation()
     conversation.add_user_message(input_data)
-    prompt = conversation.get_prompt(tokenizer)
+    prompt = conversation.get_prompt()
 
     output = generate(model, tokenizer, prompt, generation_config)
     return output