Update app.py

app.py

@@ -1,29 +1,26 @@
-from huggingface_hub import InferenceClient
 import gradio as gr
+from huggingface_hub import InferenceClient
+import concurrent.futures
 
-client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")
-system_prompt = "Deberás proporcinar código limpio, resumido, investigando cada línea en internet en foros, revisa stackoverflow.com para consultas, elimina comentarios siempre, concatena el funcionamiento de los bloques y bibliotecas, y esquematiza el funcionamiento global del código. Preveé posibles errores y complementa al final con una tabla explicando el funcionamiento, propon alternativas de bibliotecas para solucionar errores, siempre consulta internet para posibles resoluciones."
+system_prompt = ""
 system_prompt_sent = False
 
 def format_prompt(message, history):
     global system_prompt_sent
-    prompt = "<s>"
+    prompt = "[INST] "
 
-    if not any(f"[INST] {system_prompt} [/INST]" in user_prompt for user_prompt, _ in history):
-        prompt += f"[INST] {system_prompt} [/INST]"
+    if not any(f"[INST] {system_prompt} " in user_prompt for user_prompt, _ in history):
+        prompt += f"{system_prompt} "
         system_prompt_sent = True
 
     for user_prompt, bot_response in history:
-        prompt += f"[INST] {user_prompt} [/INST]"
-        prompt += f" {bot_response}</s> "
+        prompt += f"[INST] {user_prompt} [{bot_response}]"
+        prompt += " "
 
-    prompt += f"[INST] {message} [/INST]"
+    prompt += f"[INST] {message} "
     return prompt
 
-def generate(
-    prompt, history, temperature=0.9, max_new_tokens=4096, top_p=0.95, repetition_penalty=1.0,
-):
-    global system_prompt_sent
+def generate(prompt, history, temperature=0.9, max_new_tokens=4096, top_p=0.95, repetition_penalty=1.0):
     temperature = float(temperature)
     if temperature < 1e-2:
         temperature = 1e-2
@@ -39,7 +36,7 @@ def generate(
     )
 
     formatted_prompt = format_prompt(prompt, history)
-    
+
     stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=True)
     output = ""
 
@@ -49,8 +46,15 @@ def generate(
 
     return output
 
+def run_chatbot(prompt, history, temperature, max_new_tokens, top_p, repetition_penalty):
+    global client
+    client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")
+    with concurrent.futures.ProcessPoolExecutor() as executor:
+        result = executor.submit(generate, prompt, history, temperature, max_new_tokens, top_p, repetition_penalty)
+        return result
+
 chat_interface = gr.ChatInterface(
-    fn=generate,
+    fn=run_chatbot,
     chatbot=gr.Chatbot(show_label=False, show_share_button=False, show_copy_button=False, likeable=False, layout="vertical", height=900),
     concurrency_limit=9,
     theme="soft",