Why don't you use stop_tokens

soobinseo / Transformer-TTS

A Pytorch Implementation of "Neural Speech Synthesis with Transformer Network"

MIT License

659 stars 140 forks source link

Open lfchener opened 4 years ago

lfchener commented 4 years ago

In your code, you don't use stop_token to compute loss, why?

soobinseo commented 4 years ago

When I used stop token loss the learning did not work. Therefore, I just only implemented it, didn't use the stop token loss.

Thanks.

bkumardevan07 commented 3 years ago

Can you specify your stopping criteria?