Update app.py

app.py

@@ -15,12 +15,9 @@ headers = {
 
 BASE_SYSTEM_MESSAGE = "I carefully provide accurate, factual, thoughtful, nuanced answers and am brilliant at reasoning."
 
-def call_nvidia_api(history_api, system_message, max_tokens, temperature, top_p):
+def call_nvidia_api(message, history_api, system_message, max_tokens, temperature, top_p):
     messages = [{"role": "system", "content": system_message}] if system_message else []
-    for msg in history_api:
-        if msg[1]:  # Se existe uma resposta do assistente
-            messages.append({"role": "user", "content": msg[0]})
-            messages.append({"role": "assistant", "content": msg[1]})
+    messages.extend([{"role": "user", "content": msg[0]}, {"role": "assistant", "content": msg[1]} for msg in history_api])
 
     payload = {
         "messages": messages,
@@ -41,37 +38,27 @@ def call_nvidia_api(history_api, system_message, max_tokens, temperature, top_p)
 
     if response_body.get("choices"):
         assistant_message = response_body["choices"][0]["message"]["content"]
-        return assistant_message, response_body["choices"][0]
+        return assistant_message
     else:
-        return "Desculpe, ocorreu um erro ao gerar a resposta.", None
+        return "Desculpe, ocorreu um erro ao gerar a resposta."
 
-def chatbot_submit(message, chat_history_ui, chat_history_api, system_message, max_tokens_val, temperature_val, top_p_val):
-    assistant_message, api_response = call_nvidia_api(chat_history_api, system_message, max_tokens_val, temperature_val, top_p_val)
-
-    chat_history_ui.append([message, assistant_message])
-
-    if api_response:
-        chat_history_api.append([message, assistant_message])
-
-    return assistant_message, chat_history_ui, chat_history_api
+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)
-# Gradio interface setup
+
 with gr.Blocks() as demo:
-    chat_history_state_ui = gr.State([])
-    chat_history_state_api = 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,
-        inputs=[gr.Textbox(label="Your Message"), chat_history_state_ui, chat_history_state_api, system_msg, max_tokens, temperature, top_p],
-        outputs=[gr.Text(label="Assistant Response"), chat_history_state_ui, chat_history_state_api],
-        title="Chatbot Interface"
+    chat_history_state = gr.State([])
+    chat_interface = gr.ChatInterface(
+        fn=chatbot_function,
+        chatbot=gr.Chatbot(history=chat_history_state),
+        additional_inputs=[system_msg, max_tokens, temperature, top_p],
+        title="LLAMA 70B Free Demo",
     )
 
-demo.launch()
+demo.launch()