A verse-level aligned speech dataset for Konkani (romanized script), built from the Konkani Bible (KONKABSI, source: bible.com, source_id 1866). Konkani is a low-resource Indic language spoken primarily in Goa, India. This dataset is designed for training and fine-tuning TTS and ASR models.

	Dataset Details

Split Clips Notes

train ~252 95% split

validation ~13 5% split

Book processed: Matthew (MAT) — 265 verses Audio format:… See the full description on the dataset page: https://huggingface.co/datasets/Reubencf/konkani-bible-audio.