Sanity check

[JRMeyer]

>>> ravish0007
[August 5, 2019, 3:39am]

Hey there,

can I just use few sentences in the dataset and try to overfit the model
for sanity check?

My goal is actually to train a model for different language.

Thank you

[This is an archived TTS discussion thread from discourse.mozilla.org/t/sanity-check]

[JRMeyer]

>>> nmstoker
[August 7, 2019, 12:37pm]

Why not give it a go?

I'm pretty sure it won't work. You need hours of audio before it even
gets remotely recognisable, so what you're suggesting sounds naively
optimistic (but I'm open to being proved wrong!)

[Archived Post]

[JRMeyer]

>>> carlfm01
[August 14, 2019, 3:22am]

{.avatar

> can I just use few sentences in the dataset and try to overfit the
> model for sanity check?

Why not to use the notebooks? The notebooks helped me to spot couple of
corrupted files and bad transcriptions.

[Archived Post]

[JRMeyer]

>>> ravish0007
[August 15, 2019, 5:46pm]

Hey there,

Why does my model doesn't fit, after I train it in google colab, not
completely though.

I have a question, if I train a model from google colab, how do I get it
to work ?, I got a model mismatch when working with it.

Thank you

[Archived Post]