SPEAK-PP/openslr-sinhala-spelling-correction-prediction-reference-60000

OpenSLR Sinhala Spelling Correction – Prediction / Reference This dataset contains Sinhala sentence pairs intended for training and evaluating spelling-correction models on ASR output. Column Description dyslexic_sentence Noisy / dyslexic sentence (model input – ASR hypothesis) clean_sentence Clean / correct sentence (ground truth) Splits Split Source file Rows (approx.) train openslr-60000.csv ~60 000 eval openslr-7000.csv ~7 000… See the full description on the dataset page: https://huggingface.co/datasets/SPEAK-PP/openslr-sinhala-spelling-correction-prediction-reference-60000.