Cross-Lingual-Voice-Cloning
Cross-Lingual-Voice-Cloning copied to clipboard
Problem with the input mel length
Hi there,
I'm having a Runtime Error: CUDA error: device-side assert triggered when trying to train your model in LJspeech dataset. Is it because a problem caused by the excess of the mel length of the data over the default decoder or something else?
Many thanks.
@TinBarbie I have seen this error first time. Some internet browsing shows this article. I would suggest to check number of speaker, number of languages speaker id etc. are correctly provided as specified here. Also make sure that your hparams
are as specified here.