diff --git a/keras_nlp/tokenizers/sentence_piece_tokenizer.py b/keras_nlp/tokenizers/sentence_piece_tokenizer.py
index e8037fc331..e1ab5aeff6 100644
--- a/keras_nlp/tokenizers/sentence_piece_tokenizer.py
+++ b/keras_nlp/tokenizers/sentence_piece_tokenizer.py
@@ -174,7 +174,7 @@ def set_proto(self, proto):
         )
         # Keras cannot serialize a bytestring, so we base64 encode the model
         # byte array as a string for saving.
-        self.proto = base64.b64encode(proto_bytes).decode("ascii")
+        self.proto = proto
 
     def vocabulary_size(self) -> int:
         """Get the size of the tokenizer vocabulary."""