Building African-language AI is easier than finding the data
Vambo AI’s MORENA highlights Africa’s AI data gap, as developers struggle to find enough high-quality African-language text to train models.
Via TechCabal
When Vambo AI, a Johannesburg-based artificial intelligence (AI) model builder, set out to build MORENA, a 1.5-billion-parameter model covering 12 African languages, it had access to a supercomputer through a United Nations Development Programme (UNDP)-backed programme. It had the graphics processing units (GPUs), specialised chips that power AI computing, but struggled to find enough real African-language text.
“Finding that much real data was close to impossible, so a lot of it was synthetic,” Isheanesu Misi, Vambo’s co-founder and chief technology…
Courtesy
This story was originally published by TechCabal. All rights belong to the original publisher.
Read the original on techcabal.com ↗