Update app.py

app.py

@@ -21,21 +21,6 @@ tokenizer = AutoTokenizer.from_pretrained(
     pad_token=''
 )
 
-#单轮
-@spaces.GPU
-def single_turn_chat(question):
-    prompt = f"Question: {question}"
-    messages = [
-        {"role": "system", "content": "You are a helpful TCM medical assistant named 仲景中医大语言模型, created by 医哲未来 of Fudan University."},
-        {"role": "user", "content": prompt}
-    ]
-    input = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
-    model_inputs = tokenizer([input], return_tensors="pt").to(device)
-    generated_ids = model.generate( model_inputs.input_ids,max_new_tokens=512) 
-    generated_ids = [output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)]
-    response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
-    return response
-
 #多轮  
 @spaces.GPU
 def multi_turn_chat(question, chat_history=None):
@@ -80,15 +65,6 @@ def multi_turn_chat(question, chat_history=None):
 def clear_history():
     return [], []
 
-# 单轮界面
-single_turn_interface = gr.Interface(
-    fn=single_turn_chat, 
-    inputs=["text"],
-    outputs="text",
-    title="仲景GPT-V2-1.8B 单轮对话",
-    description="博极医源,精勤不倦。Unlocking the Wisdom of Traditional Chinese Medicine with AI."
-)
-
 # 多轮界面  
 multi_turn_interface = gr.Interface(
     lambda x, y: multi_turn_chat(x, y)[0],
@@ -98,8 +74,4 @@ multi_turn_interface = gr.Interface(
     allow_flagging="never",
 )
 
-# 组合单轮和多轮界面
-main_ui = gr.TabbedInterface([single_turn_interface, multi_turn_interface], ["单轮对话", "多轮对话"])
-
-# 启动界面
-main_ui.launch(debug=True, server_name='0.0.0.0', server_port=6006)
+multi_turn_interface.queue(concurrency_count=10, max_size=20).launch(debug=True, server_name='0.0.0.0', server_port=6006)