Merge pull request #742 from hertz-pj/japenese

Support Japenese TTS, and fix some bug.

Author: dathudeptrai

tensorflow_tts/configs/fastspeech.py

@@ -96,7 +96,7 @@ def __init__(
             self.vocab_size = len(bk_symbols)
         elif dataset == "libritts":
             self.vocab_size = len(lbri_symbols)
-        elif dataset == "jsut_symbols":
+        elif dataset == "jsut":
             self.vocab_size = len(jsut_symbols)
         else:
             raise ValueError("No such dataset: {}".format(dataset))