mesolitica/Malaysian-Voice-Conversion-Speech-Enhancement

该数据集用于训练马来西亚语音转换模型，采用了Postfilter技术进行数据预处理，包括检查重复的三元组和使用Silero-VAD验证语音活动。This is a Malaysian language dataset for voice conversion and speech enhancement. The dataset is processed using a postfilter on the [Malaysian-Emilia-annotated](https://huggingface.co/datasets/mesolitica/Malaysian-Emilia-annotated) dataset to train voice conversion models. The postfiltering steps include checking for repetitive trigrams and verifying voice activity using Silero-VAD.