add system prompt

app.py

@@ -1,6 +1,7 @@
 from transformers import AutoTokenizer, AutoModelForCausalLM, TextIteratorStreamer
 from threading import Thread
 import gradio as gr
+import json
 
 # load model and tokenizer
 model_name = "inclusionAI/Ling-mini-2.0"
@@ -25,12 +26,17 @@ def respond(
     """
     #client = InferenceClient(token=hf_token.token, model="openai/gpt-oss-20b")
 
+    if len(system_message) == 0:
+        system_message = "## 你是谁\n\n我是百灵(Ling)，一个由蚂蚁集团(Ant Group) 开发的AI智能助手"
+
     messages = [{"role": "system", "content": system_message}]
 
     messages.extend(history)
 
     messages.append({"role": "user", "content": message})
 
+    print(f"system_prompt: {json.dumps(messages, ensure_ascii=False, indent=2)}")
+
     text = tokenizer.apply_chat_template(
         messages,
         tokenize=False,

requirements.txt

@@ -3,3 +3,4 @@ transformers
 torch
 accelerate
 openai
+json