Loading...
Loading...
Browse all stories on DeepNewz
VisitNLLB-200 achieves 95% accuracy for 50 languages by Dec 31, 2024?
Yes • 50%
No • 50%
Official Meta announcements or peer-reviewed studies published in reputable journals
Meta's NLLB-200 AI Model Translates 200 Languages, Published in Nature
Jun 5, 2024, 07:51 PM
Meta has developed an artificial intelligence model named No Language Left Behind (NLLB-200) that can translate 200 different languages, including many low-resource and overlooked languages. The model, which has been in development for two years, was recently published in the journal Nature. The NLLB project, highlighted by NaturePortfolio and AIatMeta, aims to provide high-quality translations directly between these 200 languages, even with sparse training data. This breakthrough in neural machine translation is expected to fuel numerous innovations in the field. The work has been recognized as a significant achievement in scaling machine translation technology.
View original story
Hausa • 25%
Zulu • 25%
Basque • 25%
Amharic • 25%
Highly Effective • 33%
Moderately Effective • 34%
Needs Improvement • 33%
India • 25%
Nigeria • 25%
South Africa • 25%
Indonesia • 25%
Llama3-8B • 33%
Llama3-70B • 33%
Other • 34%
Qwen2 • 25%
Llama 3 • 25%
GLM 4 • 25%
Other • 25%
English-Hindi • 25%
English-Spanish • 25%
English-French • 25%
English-Chinese • 25%
No Award • 25%
Turing Award • 25%
ACL Best Paper • 25%
NeurIPS Best Paper • 25%