This dataset contains Urdu audio files along with their transcriptions. It is designed to be used for training and evaluating automatic speech recognition (ASR) models. The audio files are processed and segmented based on silence to facilitate training and fine-tuning of speech models.