XRXRX/X-Voice-Dataset-Train

X-Voice Training Dataset Overview The X-Voice training dataset is a large-scale multilingual speech corpus curated for high-performance speech models. It provides a robust foundation for cross-lingual phonetic and prosodic modeling. Also the train set of X-Voice Model. Core Statistics Total Speech Duration: 420K hours 30 languages European: bg (Bulgarian), cs (Czech), da (Danish), de (German), el (Greek), en (English), es (Spanish), et (Estonian)… See the full description on the dataset page: https://huggingface.co/datasets/XRXRX/X-Voice-Dataset-Train.