Update app.py

app.py

@@ -44,11 +44,13 @@ class LLMManager:
                 actual_gb = actual_size / (1024*1024*1024)
                 logging.info(f"📊 Tamaño descargado para {model_name}: {actual_gb:.2f} GB")
 
+                n_batch = model_config.get("n_batch", 96)
+
                 logging.info(f"🔄 Cargando {model_name} en RAM…")
                 llm_instance = Llama(
                     model_path=temp_path,
                     n_ctx=MAX_CONTEXT_TOKENS,
-                    n_batch=96,
+                    n_batch=n_batch,
                     n_threads=2,
                     n_threads_batch=2,
                     use_mlock=True,