diff --git a/gnes/indexer/base.py b/gnes/indexer/base.py
index 8fd89876..b7dac4c4 100644
--- a/gnes/indexer/base.py
+++ b/gnes/indexer/base.py
@@ -36,8 +36,10 @@ def __init__(self,
         :type is_big_score_similar: when set to true, then larger score means more similar
         """
         super().__init__(*args, **kwargs)
-        self.normalize_fn = normalize_fn(context=self) if normalize_fn else ModifierScoreFn(context=self)
-        self.score_fn = score_fn(context=self) if score_fn else ModifierScoreFn(context=self)
+        self.normalize_fn = normalize_fn if normalize_fn else ModifierScoreFn(context=self)
+        self.score_fn = score_fn if score_fn else ModifierScoreFn(context=self)
+        self.normalize_fn._context = self
+        self.score_fn._context = self
         self.is_big_score_similar = is_big_score_similar
         self._num_docs = 0
         self._num_chunks = 0
diff --git a/tests/test_score_fn.py b/tests/test_score_fn.py
index adb7697f..936bf244 100644
--- a/tests/test_score_fn.py
+++ b/tests/test_score_fn.py
@@ -40,7 +40,7 @@ def test_combine_score_fn(self):
         q_chunk.embedding.CopyFrom(array2blob(np.array([3, 3, 3])))
 
         for _fn in [WeightedChunkOffsetScoreFn, CoordChunkScoreFn, TFIDFChunkScoreFn, BM25ChunkScoreFn]:
-            indexer = NumpyIndexer(helper_indexer=ListKeyIndexer(), score_fn=_fn)
+            indexer = NumpyIndexer(helper_indexer=ListKeyIndexer(), score_fn=_fn())
             indexer.add(keys=[(0, 1), (1, 2)], vectors=np.array([[1, 1, 1], [2, 2, 2]]), weights=[0.5, 0.8])
             queried_result = indexer.query_and_score(q_chunks=[q_chunk], top_k=2)