speecht5_tts_jvs_ver1_e20_openjtalk_longer_20230809-031157_tokenizer

speecht5_openjtalk_tokenizer.py

@@ -34,7 +34,6 @@ def _g2p_with_np(text: str, np_lsit: str) -> List[str]:
 
 VOCAB_FILES_NAMES = {
     "vocab_file": "vocab.json",
-    "tokenizer_file": "tokenizer.json",
 }
 
 PRETRAINED_VOCAB_FILES_MAP = {
@@ -119,24 +118,7 @@ class SpeechT5OpenjtalkTokenizer(SpeechT5Tokenizer):
         with open(vocab_path, "w", encoding="utf-8") as f:
             json.dump(self.label2id, f, ensure_ascii=False, indent=2)
 
-        # special_tokens_path = Path(save_directory) / Path(
-        #     f"special_tokens_map{filename_prefix}"
-        # )
-        # with open(special_tokens_path, "w", encoding="utf-8") as f:
-        #     json.dump(
-        #         {
-        #             "bos_token": self.bos_token,
-        #             "eos_token": self.eos_token,
-        #             "unk_token": self.unk_token,
-        #             "pad_token": self.pad_token,
-        #             "mask_token": self.mask_token,
-        #         },
-        #         f,
-        #         ensure_ascii=False,
-        #         indent=2,
-        #     )
-
-        return str(vocab_path), None  # str(special_tokens_path)
+        return (str(vocab_path),)
 
     def _tokenize(self, text: str) -> List[str]:
         return _g2p_with_np(text, self.non_phenome_characters)