default train id speaker_id text transcription language gender audio
This dataset merges the human-labeled Oromo ASR split from google/WaxalNLP with the autolabeled Orom