Update app.py

app.py

@@ -9,7 +9,6 @@ import tempfile
 app = Flask(__name__)
 logging.basicConfig(level=logging.INFO)
 
-# CONFIGURACIÓN DE TOKENS
 MAX_CONTEXT_TOKENS = 1024 * 8
 MAX_GENERATION_TOKENS = 1024 * 4
 
@@ -236,6 +235,10 @@ def chat_completions():
         # Aplicar límite de tokens si no se especifica
         if 'max_tokens' not in kwargs:
             kwargs['max_tokens'] = MAX_GENERATION_TOKENS
+        else:
+            # Validar que max_tokens no exceda el máximo permitido
+            if kwargs['max_tokens'] > MAX_GENERATION_TOKENS:
+                kwargs['max_tokens'] = MAX_GENERATION_TOKENS
         
         result = llm_manager.chat_completion(model_name, messages, **kwargs)