Tibetan dialect speech synthesis dataset

It comprises 3 large-size folders (ad, kb, wz), each large folder comprises 4 middle-size folders (e.g. ad1, ad2, ad3, and ad4 under ad), and each multiple-size folder comprises multiple small folders (e.g. ad1-1, ad1-2... ad1-10 under ad1), in which *.wav is voice data, and *.txt is text data. In total, the dataset includes: (1) 4684 *.wav audio files, with a total length of more than 10 hours and a data volume of 8.01GB, including 2.45GB for Amdo, 2.13GB for Kangba and 3.34GB for Weizang; (2) 65 *.txt text files, with data volume of 1.631MB, and a total of 4684 sentences.