* fix batch_size, test=tts

* update readme and fixed ngpu bug

* update png in readme

* update readme and FT web ttsText

* update finetune, test=tts

[s2t] support bitransformer decoder

use cmd for voice clone , finetune and ernie-sat

* Solve execute the following code with return  wav:
iob = io.BytesIO(wav)
wave.open(iob, 'rb')
will throw an "unknown format: 3"  exception

* update yaml and readme, test=tts

* Replaced the fixed path with the dynamic path in MFA, test=doc

* add finetune demos, test=doc