speecht5_tts-wolof
Wolof speech synthesis, MIT
Official site consultedAI: models
Last verified: 1 October 2026
Wolof text-to-speech model of 100 million parameters, fine-tuned from Microsoft's SpeechT5 on the galsenai/wolof_tts dataset (40,000 texts), with a Wolof tokenizer, a HiFi-GAN vocoder and speaker embeddings. MIT license.
Services
- Wolof text-to-speech
Access requirements
Open weights on Hugging Face, MIT license, usable with Transformers.
Sources
- speecht5_tts-wolof (new tab), verified on 1 October 2026Official site consulted