Update app.py

app.py

@@ -16,17 +16,10 @@ headers = {
 BASE_SYSTEM_MESSAGE = "I carefully provide accurate, factual, thoughtful, nuanced answers and am brilliant at reasoning."
 
 def call_nvidia_api(message, history_api, system_message, max_tokens, temperature, top_p):
-    messages = []
-    if system_message:  # Adiciona a mensagem do sistema, se houver
-        messages.append({"role": "system", "content": system_message})
-    # Adiciona as mensagens do histórico
+    messages = [{"role": "system", "content": system_message}] if system_message else []
+    messages.extend([{"role": "user", "content": message}])
     for msg in history_api:
-        messages.append({"role": "user", "content": msg[0]})
-        if msg[1]:  # Garante que não adicionamos respostas vazias
-            messages.append({"role": "assistant", "content": msg[1]})
-
-    # Adiciona a mensagem atual do usuário
-    messages.append({"role": "user", "content": message})
+        messages.extend([{"role": "user", "content": msg[0]}, {"role": "assistant", "content": msg[1]}])
 
     payload = {
         "messages": messages,
@@ -51,24 +44,25 @@ def call_nvidia_api(message, history_api, system_message, max_tokens, temperatur
     else:
         return "Desculpe, ocorreu um erro ao gerar a resposta."
 
-
 def chatbot_function(message, history_api, system_message, max_tokens, temperature, top_p):
     assistant_message = call_nvidia_api(message, history_api, system_message, max_tokens, temperature, top_p)
     history_api.append([message, assistant_message])
     return assistant_message, history_api
 
-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)
+system_msg = gr.Textbox(value=BASE_SYSTEM_MESSAGE, label="System Message", placeholder="System prompt.", lines=5)
+max_tokens = gr.Slider(minimum=20, maximum=1024, label="Max Tokens", step=20, value=1024)
+temperature = gr.Slider(minimum=0.0, maximum=1.0, label="Temperature", step=0.1, value=0.2)
+top_p = gr.Slider(minimum=0.0, maximum=1.0, label="Top P", step=0.05, value=0.7)
 
 with gr.Blocks() as demo:
     chat_history_state = gr.State([])
     chat_interface = gr.ChatInterface(
         fn=chatbot_function,
-        chatbot=gr.Chatbot(value=chat_history_state),
-        additional_inputs=[system_msg, max_tokens, temperature, top_p],
+        inputs=["message", "history_api", system_msg, max_tokens, temperature, top_p],
+        outputs=["assistant_message", "history_api"],
         title="LLAMA 70B Free Demo",
+        description="Explore the capabilities of LLAMA 2 70B",
+        additional_inputs=[system_msg, max_tokens, temperature, top_p]
     )
 
-demo.launch()
+demo.launch()