Lets rename this forum to STT!

Archives DeepSpeech
#1

I am a litte sad that this forum is so inactive since the team from DeepSpeech switched to Coqui. I belive a general forum about using the Common Voice Dataset in STT networks would be a cool thing. Plus there is already a forum for TTS with the same name scheme.

I would love to discuss how to train e.g. a Vosk model with the Common Voice Dataset, but this was never really possible in this forum. Now that Deepspeech is inactive, it would be great to rename the forum and give it new life.

What do you think about it?

1 Like
About the Using the common voice dataset category
#2

We can do it without renaming. Kaldi has commonvoice recipe here: https://github.com/kaldi-asr/kaldi/tree/master/egs/commonvoice/s5, it should be a good start. Or default mini_librispeech recipe we advise.

1 Like
#3

That is likely going to get messy, this forum has always been about DeepSpeech project. I suggest you ask administrator for a new “Common Voice Dataset usages” forum rather, it’s likely something that @heyhillary can welcome I think.

1 Like
#4

I agree with Alex on this. I think what I can do is make a subforum under Common voice similar to the language groups called “Dataset use cases” . Also @Stefano if you are interested I’m happy to host you for a lightening talk to share you work on the vosk model.

2 Likes
#5

These two ideas sound great to me!

#6

Awesome, I will follow up with you on direct message

#7

Hey everyone, just to update the new page has been created -> https://discourse.mozilla.org/c/voice/using/661

3 Likes