Language

[magicse]

Hello. Thank you for Your work.
https://colab.research.google.com/drive/1cI9VzlimS51uAw4uCR-OBeSXRPBc4KoK?usp=sharing
I try convert vit model to onnx for my languagem, and all test good with onnxruntime.
But in step with test with sherpa-onnx it say that my language unknown by sherpa.
Could you please add Ukrainian language?
Model

%%shell
sherpa-onnx-offline-tts \
  --vits-model=./model.onnx \
  --vits-lexicon=./lexicon.txt \
  --vits-tokens=./tokens.txt \
  --output-filename=./test.wav \
  "Вітаю, як ваші справи? Що ви тут робите?"
  
  /project/sherpa-onnx/csrc/parse-options.cc:Read:361 sherpa-onnx-offline-tts --vits-model=./model.onnx --vits-lexicon=./lexicon.txt --vits-tokens=./tokens.txt --output-filename=./test.wav 'Вітаю, як ващі справи?. Що ви тут робите?' 

/project/sherpa-onnx/csrc/lexicon.cc:InitLanguage:321 Unknown language: Ukrainian
---------------------------------------------------------------------------
CalledProcessError                        Traceback (most recent call last)
[<ipython-input-50-490d59815704>](https://localhost:8080/#) in <cell line: 1>()
----> 1 get_ipython().run_cell_magic('shell', '', 'sherpa-onnx-offline-tts \\\n  --vits-model=./model.onnx \\\n  --vits-lexicon=./lexicon.txt \\\n  --vits-tokens=./tokens.txt \\\n  --output-filename=./test.wav \\\n  "Вітаю, як ваші справи? Що ви тут робите?"\n')

3 frames
[/usr/local/lib/python3.10/dist-packages/google/colab/_system_commands.py](https://localhost:8080/#) in check_returncode(self)
    135   def check_returncode(self):
    136     if self.returncode:
--> 137       raise subprocess.CalledProcessError(
    138           returncode=self.returncode, cmd=self.args, output=self.output
    139       )

CalledProcessError: Command 'sherpa-onnx-offline-tts \
  --vits-model=./model.onnx \
  --vits-lexicon=./lexicon.txt \
  --vits-tokens=./tokens.txt \
  --output-filename=./test.wav \
  "Вітаю, як ващі справи?. Що ви тут робите?"
' returned non-zero exit status 255.

[csukuangfj]

please wait for #453

[csukuangfj]

@magicse

Please try Ukrainian tts models using our huggingface space.

https://huggingface.co/spaces/k2-fsa/text-to-speech

[magicse]

I tested this https://huggingface.co/spaces/k2-fsa/text-to-speech but it nothing output to player....
Wave duration : 1.856 s
Processing time: 0.086 s
RTF: 0.086/1.856 = 0.046
But after progress player is still disabled.

Also i know this model with voice lada vits-piper-uk_UA-lada-x_low

Could you try convert this model?
https://github.com/coqui-ai/TTS/releases/download/v0.8.0_models/tts_models--uk--mai--vits.zip
Because it has more pleasure voice.

[magicse]

I tested and it work ok.
Wave duration : 1.760 s
Processing time: 0.072 s
RTF: 0.072/1.760 = 0.041

Could you try convert this model instead lada?
https://github.com/coqui-ai/TTS/releases/download/v0.8.0_models/tts_models--uk--mai--vits.zip
Because it has more pleasure voice.

[magicse]

Hi @csukuangfj
could you please add to https://huggingface.co/spaces/k2-fsa/text-to-speech
this model https://github.com/coqui-ai/TTS/releases/download/v0.8.0_models/tts_models--uk--mai--vits.zip ?

[csukuangfj]

Hi @csukuangfj could you please add to https://huggingface.co/spaces/k2-fsa/text-to-speech this model https://github.com/coqui-ai/TTS/releases/download/v0.8.0_models/tts_models--uk--mai--vits.zip ?

If VITS models from coqui-ai/TTS are also using espeak-ng to convert text to phonemes, I will support all of them in the coming week including the model for Ukrainian.

[csukuangfj]

Hi @csukuangfj could you please add to https://huggingface.co/spaces/k2-fsa/text-to-speech this model https://github.com/coqui-ai/TTS/releases/download/v0.8.0_models/tts_models--uk--mai--vits.zip ?

@magicse

I am happy to tell you that I just managed to convert it. Please wait for a moment before I add it to sherpa-onnx.

[csukuangfj]

Fixed by #469