This is a processed version of the Fahd1199/darija-tts dataset,
formatted specifically for finetuning VITS/MMS TTS models.
The dataset contains:
1740 training samples
436 validation samples
Each sample has:
audio: Path to a 16kHz WAV file
text: Darija (Moroccan Arabic) transcription
speaker_id: Speaker identifier (single speaker)
How to Use