Aynursusuz/voice-design-bench-50-dnsmos

这是一个用于文本到语音任务的小型数据集，包含50个训练样本。每个样本包含文本和对应的三种不同音频格式（qwen_audio、echo_audio、omni_audio），以及它们的DNSMOS评分。音频采样率分别为24000Hz和44100Hz。数据集还提供了三种模型的DNSMOS评分统计信息。This is a small-scale dataset for text-to-speech (TTS) tasks, containing 50 training samples. Each sample includes the input text, three corresponding audio files in different formats (qwen_audio, echo_audio, omni_audio), as well as their respective DNSMOS scores. The audio files have sampling rates of 24000 Hz and 44100 Hz, respectively. Additionally, the dataset provides statistical summaries of DNSMOS scores for the three models.