[Docs]: Ask for Dataset used in pretrained model.

[signofthefour]

Documentation Reference

VALLE Documents

Feedback on documentation

Hi, first of all, thank you (Amphion team) so much for making this useful repo.

I am writing to ask about the 6k hours of LibriLight that Amphion team used for the pretrained model provided in the docs above.
Did it randomly be collected from LibriLight or some specific portions was used? I'd appreciate that if you could clarify this part more.

Additional context

It's would be useful for academic usage since we can directly cite Amphion if we know the correct dataset for fair comparision.