add caching

src/rag/retriever.py

@@ -30,13 +30,13 @@ class Retriever:
         ])
 
         # Load SentenceTransformer model for semantic similarity
-        self.semantic_model = SentenceTransformer(semantic_model)
+        self.semantic_model = SentenceTransformer(semantic_model, cache_folder="cache/")
         self.semantic_chunk_embeddings = self.semantic_model.encode(
             self.text_chunks, convert_to_tensor=False
         )
 
         # Load CrossEncoder model for reranking
-        self.reranker = CrossEncoder(reranker_model)
+        self.reranker = CrossEncoder(reranker_model, cache_dir="cache/")
         
 
     def _load_chunks_from_directory(self, chunked_dir: str) -> list[str]: