Update app.py

app.py

@@ -1,29 +1,29 @@
-import concurrent.futures
-import gradio as gr
-from dogpile.cache import make_region
 from huggingface_hub import InferenceClient
+import gradio as gr
 
-cache = make_region().configure('dogpile.cache.memory', thread_local=True)
-
-system_prompt = ""
+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_sent = False
 
 def format_prompt(message, history):
     global system_prompt_sent
-    prompt = "[INST] "
+    prompt = "<s>"
 
-    if not any(f"[INST] {system_prompt} " in user_prompt for user_prompt, _ in history):
-        prompt += f"{system_prompt} "
+    if not any(f"[INST] {system_prompt} [/INST]" in user_prompt for user_prompt, _ in history):
+        prompt += f"[INST] {system_prompt} [/INST]"
         system_prompt_sent = True
 
     for user_prompt, bot_response in history:
-        prompt += f"[INST] {user_prompt} [{bot_response}]"
-        prompt += " "
+        prompt += f"[INST] {user_prompt} [/INST]"
+        prompt += f" {bot_response}</s> "
 
-    prompt += f"[INST] {message} "
+    prompt += f"[INST] {message} [/INST]"
     return prompt
 
-def generate(prompt, history, temperature=0.9, max_new_tokens=4096, top_p=0.95, repetition_penalty=1.0):
+def generate(
+    prompt, history, temperature=0.9, max_new_tokens=4096, top_p=0.95, repetition_penalty=1.0,
+):
+    global system_prompt_sent
     temperature = float(temperature)
     if temperature < 1e-2:
         temperature = 1e-2
@@ -39,11 +39,7 @@ def generate(prompt, history, temperature=0.9, max_new_tokens=4096, top_p=0.95,
     )
 
     formatted_prompt = format_prompt(prompt, history)
-    cache_key = f"generate:{formatted_prompt}:{temperature}:{max_new_tokens}:{top_p}:{repetition_penalty}"
-    cached_response = cache.get(cache_key)
-    if cached_response is not None:
-        return cached_response
-
+    
     stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=True)
     output = ""
 
@@ -51,12 +47,8 @@ def generate(prompt, history, temperature=0.9, max_new_tokens=4096, top_p=0.95,
         output += response.token.text
         yield output
 
-    cache.set(cache_key, output)
-
     return output
 
-client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")
-
 chat_interface = gr.ChatInterface(
     fn=generate,
     chatbot=gr.Chatbot(show_label=False, show_share_button=False, show_copy_button=False, likeable=False, layout="vertical", height=900),