diff --git a/wenet/text/bpe_tokenizer.py b/wenet/text/bpe_tokenizer.py
index 6dffa0989..de1b504a9 100644
--- a/wenet/text/bpe_tokenizer.py
+++ b/wenet/text/bpe_tokenizer.py
@@ -17,11 +17,19 @@ def __init__(
     ) -> None:
         super().__init__(symbol_table, non_lang_syms, split_with_space,
                          connect_symbol, unk)
-        import sentencepiece as spm
-        self.bpe_model = spm.SentencePieceProcessor()
-        self.bpe_model.load(bpe_model)
+        self._model = bpe_model
+        # NOTE(Mddct): multiprocessing.Process() issues
+        #              don't build sp here
+        self.bpe_model = None
+
+    def _build_sp(self):
+        if self.bpe_model is None:
+            import sentencepiece as spm
+            self.bpe_model = spm.SentencePieceProcessor()
+            self.bpe_model.load(self._model)
 
     def text2tokens(self, line: str) -> List[str]:
+        self._build_sp()
         line = line.strip()
         if self.non_lang_syms_pattern is not None:
             parts = self.non_lang_syms_pattern.split(line.upper())
@@ -38,5 +46,6 @@ def text2tokens(self, line: str) -> List[str]:
         return tokens
 
     def tokens2text(self, tokens: List[str]) -> str:
+        self._build_sp()
         text = super().tokens2text(tokens)
         return text.replace("▁", ' ').strip()