Update app.py

app.py

@@ -5,12 +5,13 @@ import re
 import gradio as gr
 import os
 import urllib3
+import pickle
 from functools import lru_cache
 from dotenv import load_dotenv
 from fastapi import FastAPI, Request
 from fastapi.responses import JSONResponse
-from queue import Queue
-import pickle  #Para persistencia
+import time
+from tqdm import tqdm # Importando tqdm
 
 urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)
 
@@ -37,11 +38,11 @@ global_data = {
 }
 
 response_cache = {}
-model_cache_dir = "model_cache" # Directorio para guardar modelos en disco
+model_cache_dir = "model_cache"
 os.makedirs(model_cache_dir, exist_ok=True)
 
 class ModelManager:
-    def __init__(self, max_models=10):
+    def __init__(self, max_models=2):
         self.models = {}
         self.max_models = max_models
         self.model_cache_dir = model_cache_dir
@@ -99,9 +100,13 @@ def remove_duplicates(text):
 @lru_cache(maxsize=128)
 def generate_model_response(model, inputs):
     try:
+        start_time = time.time()
         response = model(inputs, max_tokens=150)
+        end_time = time.time()
+        print(f"Tiempo de generación del modelo: {end_time - start_time:.4f} segundos")
         return remove_duplicates(response['choices'][0]['text'])
-    except Exception:
+    except Exception as e:
+        print(f"Error en la generación del modelo: {e}")
         return ""
 
 async def process_message(message):
@@ -110,8 +115,9 @@ async def process_message(message):
         return response_cache[inputs]
 
     responses = {}
+    start_time = time.time()
     with ThreadPoolExecutor(max_workers=model_manager.max_models) as executor:
-        futures = [executor.submit(model_manager.load_model, config) for config in global_data['model_configs']]
+        futures = [executor.submit(model_manager.load_model, config) for config in tqdm(global_data['model_configs'], desc="Cargando modelos")]
         for future in as_completed(futures):
             future.result()
 
@@ -121,6 +127,8 @@ async def process_message(message):
                 responses[config['name']] = generate_model_response(model, inputs)
                 model_manager.unload_model(config['name'])
 
+    end_time = time.time()
+    print(f"Tiempo total de procesamiento: {end_time - start_time:.4f} segundos")
     formatted_response = "\n\n".join([f"**{model}:**\n{response}" for model, response in responses.items()])
     response_cache[inputs] = formatted_response
     return formatted_response