Update app.py

app.py

@@ -5,20 +5,11 @@ from concurrent.futures import ThreadPoolExecutor, as_completed
 import uvicorn
 import re
 from dotenv import load_dotenv
-import gradio as gr
-from spaces import ZeroGPU
 
 load_dotenv()
 
 app = FastAPI()
 
-# Inicializar ZeroGPU
-try:
-    ZeroGPU.initialize()
-except Exception as e:
-    print(f"ZeroGPU initialization failed: {e}")
-
-# Diccionario global para almacenar modelos y tokens
 global_data = {
     'models': {},
     'tokens': {
@@ -69,22 +60,18 @@ class ModelManager:
     def load_all_models(self):
         if self.loaded:
             return global_data['models']
-        
-        try:
-            with ThreadPoolExecutor() as executor:
-                futures = [executor.submit(self.load_model, config) for config in model_configs]
-                models = []
-                for future in as_completed(futures):
-                    model = future.result()
-                    if model:
-                        models.append(model)
-            
-            global_data['models'] = {model['name']: model['model'] for model in models}
-            self.loaded = True
-            return global_data['models']
-        except Exception as e:
-            print(f"Error loading models: {e}")
-            return {}
+
+        with ThreadPoolExecutor() as executor:
+            futures = [executor.submit(self.load_model, config) for config in model_configs]
+            models = []
+            for future in as_completed(futures):
+                model = future.result()
+                if model:
+                    models.append(model)
+
+        global_data['models'] = {model['name']: model['model'] for model in models}
+        self.loaded = True
+        return global_data['models']
 
 model_manager = ModelManager()
 model_manager.load_all_models()
@@ -111,7 +98,6 @@ def remove_duplicates(text):
             seen_lines.add(line)
     return '\n'.join(unique_lines)
 
-@spaces.GPU(duration=0)
 def generate_model_response(model, inputs, top_k, top_p, temperature):
     try:
         response = model.generate(inputs, top_k=top_k, top_p=top_p, temperature=temperature)
@@ -124,11 +110,12 @@ def generate_model_response(model, inputs, top_k, top_p, temperature):
 async def generate(request: ChatRequest):
     try:
         inputs = normalize_input(request.message)
-        futures = [
-            executor.submit(generate_model_response, model, inputs, request.top_k, request.top_p, request.temperature)
-            for model in global_data['models'].values()
-        ]
-        responses = [{'model': model_name, 'response': future.result()} for model_name, future in zip(global_data['models'].keys(), as_completed(futures))]
+        with ThreadPoolExecutor() as executor:
+            futures = [
+                executor.submit(generate_model_response, model, inputs, request.top_k, request.top_p, request.temperature)
+                for model in global_data['models'].values()
+            ]
+            responses = [{'model': model_name, 'response': future.result()} for model_name, future in zip(global_data['models'].keys(), as_completed(futures))]
         unique_responses = remove_repetitive_responses(responses)
         return unique_responses
     except Exception as e:
@@ -152,4 +139,4 @@ def remove_repetitive_responses(responses):
     return unique_responses
 
 if __name__ == "__main__":
-    uvicorn.run(app, host="0.0.0.0", port=8000)
+    uvicorn.run(app, host="0.0.0.0", port=7860)