This dataset merges the human-labeled Oromo ASR split from
google/WaxalNLP with the autolabeled Oromo split from
israel/waxal-autolabled.
For TTS use, the leading [ORM] language tag has been removed from
autolabeled transcriptions. Rows include both text and transcription
with the same cleaned value.
Target repo: b1n1yam/waxal-orm-tts-merged