Update README.md

README.md

@@ -12,8 +12,8 @@ language: hk
 
 Speech-to-speech translation model from fairseq:
 - English-Hokkien
-- Trained with supervised data in TED domain, and weakly supervised data in TED and Audiobook domain. See [here]( https://research.facebook.com/publications/hokkien-direct-speech-to-speech-translation) 
-for training details
+- Trained with [TAT-TTS](https://sites.google.com/speech.ntut.edu.tw/fsw/home/tat-tts-corpus) data with 4 speakers. See [here]( https://research.facebook.com/publications/hokkien-direct-speech-to-speech-translation) 
+for more training details.
 
 ## Usage