ado

kit if you're going to update the space by adding muyu, just use this version ok??

Sounds better than the OV2 version.

But i think you should add more laughs to the dataset next time or use a bigger dataset.

i thought that the laughs might ruin the model

As you can see, you can hear how the voice breaks when it tries to laugh on a sample audio.

So, the fact that "laughs might ruin the model" isn't right at all.

next is probably nasa

but this time i'll try to train it locally

might update the dataset and retrain it