Introduced in our EMNLP 2025 paper
Voice of a Continent, the
Simba Series represents the current state-of-the-art for African speech AI.
1from transformers import VitsModel, AutoTokenizer
2import torch
3
4model_name="Simba-TTS-afr" ## Simba-TTS-twi-asanti, Simba-TTS-twi-akuapem, Simba-TTS-lin, Simba-TTS-sot, Simba-TTS-tsn, Simba-TTS-xho
5
6model = VitsModel.from_pretrained(model_name)
7tokenizer = AutoTokenizer.from_pretrained(model_name)
8
9text = "Ons noem hierdie deeltjies sub-atomiese deeltjies" #example of Afrikaans (afr) language
10inputs = tokenizer(text, return_tensors="pt")
11
12with torch.no_grad():
13 output = model(**inputs).waveform
14
1scipy.io.wavfile.write("outputfile.wav", rate=model.config.sampling_rate, data=output.float().numpy())
2
If you use the Simba models or SimbaBench benchmark for your scientific publication, or if you find the resources in this website useful, please cite our paper.
1
2@inproceedings{elmadany-etal-2025-voice,
3 title = "Voice of a Continent: Mapping {A}frica{'}s Speech Technology Frontier",
4 author = "Elmadany, AbdelRahim A. and
5 Kwon, Sang Yun and
6 Toyin, Hawau Olamide and
7 Alcoba Inciarte, Alcides and
8 Aldarmaki, Hanan and
9 Abdul-Mageed, Muhammad",
10 editor = "Christodoulopoulos, Christos and
11 Chakraborty, Tanmoy and
12 Rose, Carolyn and
13 Peng, Violet",
14 booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing",
15 month = nov,
16 year = "2025",
17 address = "Suzhou, China",
18 publisher = "Association for Computational Linguistics",
19 url = "https://aclanthology.org/2025.emnlp-main.559/",
20 doi = "10.18653/v1/2025.emnlp-main.559",
21 pages = "11039--11061",
22 ISBN = "979-8-89176-332-6",
23}
24
