UniAudio

UniAudio copied to clipboard

SVS task recipe will be provided soon?

baipeng1

great work! I notice that it seems only audio generation tasks are supported, but some of tasks like `ASR` is defined in code, have you tried task for text-predict-task,thks!

HaiFengZeng

Nice work! Could you please also share your recipe of TSE?

wjyfelicity

inference TTS

1

Hi i have succeded to debug a few errors from the inference code. Can you please add the DecoderOnlyModel in model.py and what is phone.1.pt file that is required by...

pawanhv

several bugs

3

running TTS egs 1. text_tokenizer.py https://github.com/yangdongchao/UniAudio/blob/b4da009653486828b4d71a6efe1772b1403ce324/UniAudio/tools/tokenizer/phone/text_tokenizer.py#L12 no TextTokenizer ``` class Text2PhoneTokenizer(AbsTokenizer): def __init__(self, langdir='UniAudio/checkpoints/lang_nosp'): super(TextTokenizer, self).__init__() ``` 2. offline_tokenization.py ``` 5 Traceback (most recent call last): 6 File "data_scripts/offline_tokenization.py", line...

Jackiexiao

Can you please add a license?

kachiO

Amazing work! Could you please let me know more detailed training information such as the training time? Thanks.

huifu99

example of tts(zero shot) on libritts?

2

Hi, there are some examples of tts(zero shot) on libritts?

deyituo

Hello, dear author!I have a question to ask you for advice. During my preparing TTS experimental data (LibriLight for tts), I noticed that the transcribed results only consist of text...

Alidaling

The difference between AudioTokenizer and EncodecTokenizer?

3

I find 2 tokenizer models for audio, AudioTokenizer and EncodecTokenizer. In egs, tts, vc, and se all use tokenizer "audio". I guess these models are all based on SoundStream. What's...

chenxinglili