felixmr1/librivox-tts-swedish

LibriVox TTS Swedish (5-minute chunks)是一个瑞典语的长文本转语音（TTS）和自动语音识别（ASR）数据集，数据来源于LibriVox的有声读物。音频被按照说话者顺序拼接成约5分钟的片段（16 kHz 单声道），适用于长上下文TTS/ASR训练。数据集包含音频、文本、语言、说话者ID、性别、DNSMOS评分和持续时间等信息。LibriVox TTS Swedish (5-minute chunks) is a Swedish long-form text-to-speech (TTS) and automatic speech recognition (ASR) dataset derived from LibriVox audiobooks. The audio is concatenated in source order into ~5-minute chunks (16 kHz mono), suitable for long-context TTS/ASR training. The dataset includes fields such as audio, text, language, speaker ID, gender, DNSMOS score, and duration.