Dataset Card for Segmented Moroccan Darija Wiki Dataset
Dataset Summary
This dataset provides short Moroccan Darija (Moroccan Arabic) speech segments derived from the atlasia/Moroccan-Darija-Wiki-Audio-Dataset.It is a cleaned and segmented version of the parent dataset, text-cleaned with Gemini 2.5-flash and processed using a fine-tuned Whisper model for Darija (to be open-sourced soon).
Each audio is split into segments of up to 30 seconds to make it suitable for… See the full description on the dataset page: https://huggingface.co/datasets/anaszil/Segmented-Moroccan-Darija-Wiki-Audio-Dataset.
