semidark/

hokuspokus-qwen3-tts-hybrid

This dataset aims to provide a professional-grade German TTS training corpus. It combines high-quality human narration with state-of-the-art synthetic augmentation to provide a legally safe alternative for training models like kokoro. Work in Progress

	Dataset Details




	
	Dataset Sources

Source Audio: HUI Audio Corpus / LibriVox Voice Seed: "Hokuspokus" (Professional Narrator)

	Uses




	
	Direct Use… See the full description on the dataset page: https://huggingface.co/datasets/semidark/hokuspokus-qwen3-tts-hybrid.