[Bug] HELP HELP HELP

[paisekamao]

Describe the bug

I M trying to use this repo for urdu language i have found some pretrained module on hugging face
but i m unable to use
i dont have any prior knowledge of python i m not familiar with tts
can some one pls help???????

To Reproduce

from TTS.api import TTS

Initialize TTS with a VITS model

model_name = "facebook/mms-tts-urd-script_devanagari"
tts = TTS(model_name=model_name)

Define your text input

text = "Kya kehna chahte ho?"

Generate and save speech

tts.tts_to_file(text=text, file_path="output.wav")

print("Audio saved at: output.wav")

Expected behavior

No response

Logs

Exception has occurred: ValueError
not enough values to unpack (expected 4, got 2)
  File "E:\PROJECTS\Quotes_urdu\t2.py", line 27, in <module>
    tts = TTS(model_name=model_name)
ValueError: not enough values to unpack (expected 4, got 2)

Environment

using windows 10
python 3.10.6
TTS                          0.22.0

Additional context

No response

[eginhard]

You can load this model in Coqui with:

tts = TTS("tts_models/urd-script_devanagari/fairseq/vits")

_arabic or _latin would also work instead of _devanagari depending on your input text.

There is also a bug that you might encounter when loading fairseq models, which has been fixed in our Coqui fork, which you can install with pip install coqui-tts. This repository is not updated anymore.

[paisekamao]

can u please define in detail?
as i tried to run

from TTS.api import TTS

tts = TTS("tts_models/urd-script_devanagari/fairseq/vits")

Define your text input

text = "Kya kehna chahte ho?"

Generate and save speech

tts.tts_to_file(text=text, file_path="output.wav")

print("Audio saved at: output.wav")

and got error as
Exception has occurred: UnboundLocalError
local variable 'dataset' referenced before assignment
File "E:\PROJECTS\Quotes_urdu\t2.py", line 25, in
tts = TTS("tts_models/urd-script_devanagari/fairseq/vits")
UnboundLocalError: local variable 'dataset' referenced before assignment

[Mkkrishnasai]

issue fixed in the PR #3994

[rose07]

https://tts.byylook.com/ai/text-to-speech