Update app.py

app.py

@@ -63,10 +63,7 @@ def chatbot_submit(message, chat_history, system_message, max_tokens_val, temper
  print("Updating chatbot...")
 
  # Atualiza o histórico do chat com a mensagem do usuário
- if not chat_history or (chat_history and chat_history[-1]["role"] != "user"):
- chat_history = user(message, chat_history, system_message)
- else:
- chat_history = user(message, chat_history)
+ chat_history.append({"role": "user", "content": message})
 
  # Chama a API da NVIDIA para gerar uma resposta
  chat_history = call_nvidia_api(chat_history, max_tokens_val, temperature_val, top_p_val)
@@ -77,20 +74,19 @@ def chatbot_submit(message, chat_history, system_message, max_tokens_val, temper
  else:
  assistant_message = "Desculpe, ocorreu um erro ao gerar a resposta."
 
- return assistant_message
-
-
+ return assistant_message, chat_history
+ 
+chat_history_state = gr.State([])
+system_msg = gr.Textbox(BASE_SYSTEM_MESSAGE, label="System Message", placeholder="System prompt.", lines=5)
+max_tokens = gr.Slider(20, 1024, label="Max Tokens", step=20, value=1024)
+temperature = gr.Slider(0.0, 1.0, label="Temperature", step=0.1, value=0.2)
+top_p = gr.Slider(0.0, 1.0, label="Top P", step=0.05, value=0.7)
 with gr.Blocks() as demo:
- chat_history_state = gr.State([])
- system_msg = gr.Textbox(BASE_SYSTEM_MESSAGE, label="System Message", placeholder="System prompt.", lines=5)
- max_tokens = gr.Slider(20, 1024, label="Max Tokens", step=20, value=1024)
- temperature = gr.Slider(0.0, 1.0, label="Temperature", step=0.1, value=0.2)
- top_p = gr.Slider(0.0, 1.0, label="Top P", step=0.05, value=0.7)
-
  chatbot = gr.ChatInterface(
  fn=chatbot_submit,
  additional_inputs=[system_msg, max_tokens, temperature, top_p],
  title="LLAMA 70B Free Demo",
+ state=chat_history_state, 
  description="""
  <div style="text-align: center; font-size: 1.5em; margin-bottom: 20px;">
  <strong>Explore the Capabilities of LLAMA 2 70B</strong>