The First FOSD-Tacotron-2-based Text-to-Speech Model Dataset for Vietnamese
Tran, Duc Chung · Data Archiving and Networked Services (DANS) · 2020
This is the 1st FPT Open Speech Data (FOSD) and Tacotron-2 -based Text-to-Speech Model Dataset for Vietnamese. It comprises of: - A configuration file in *.json format; - Training and validation text input files (in *.csv format); - A trained model (checkpoint file, after 225,000 steps); - Sample generated audios from the trained model. This dataset is useful for research related to TTS and its applications, text processing and especially TTS output optimization given a set of predefined input texts.