This speech dataset consists of 10385 short audio clips of multiple speakers conversing in Polish. A transcription is provided for each clip, also gender of speaker is provided for part of the dataset. Clips have total length of almost 10 hours.
This dataset was created from LUNA dataset of human-human and human-computer dialogues on the topic of public transport. The postprocessing consisted of extracting segments with human speech together with their transcripts. If you are interested in the… See the full description on the dataset page:
https://huggingface.co/datasets/czyzi0/luna-speech-dataset.
