metadata
language:
- ar
task_categories:
- audio-text-to-text
tags:
- arabic
- asr
- correction
- whisper
license: apache-2.0
Arabic ASR Post-Correction Dataset
Dataset for correcting common errors in Whisper ASR outputs for Arabic.
Features
- Input: Noisy ASR transcriptions
- Target: Corrected Modern Standard Arabic (MSA)