🎙️ Yoruba–English Code-Switching TTS
A voice-cloning text-to-speech model for Yoruba, including natural Yoruba–English code-switching — full fine-tuned from VoxCPM2 on ~1,039 hours pooled from DSN African Voices, NaijaVoices, YECS, and WAXAL.
⚠️ Experimental — trained for only ~1 epoch. Quality varies by sentence; rarer words/proper nouns are more likely to be mispronounced.
Try an example
Part of the YECS-Downstream speech research effort by LyngualLabs · Model card