Skip model loading when using Inference API

bot.py

@@ -231,8 +231,9 @@ class RAGBot:
         self.tokenizer = None
         self.vector_retriever = None
         
-        # Load model
-        self._load_model()
+        # Load model (unless skipping for Inference API)
+        if not hasattr(args, 'skip_model_loading') or not args.skip_model_loading:
+            self._load_model()
         
         # Initialize vector retriever
         self._setup_vector_retriever()