Hugging Face's logo

MultiSynt
community
Activity Feed

AI & ML interests

None defined yet.

Recent Activity

Villekom  updated a model 6 days ago
MultiSynt/nemotron-cc-danish-tower9b
Zihao-Li  authored a paper 8 days ago
Test-Time Scaling of Reasoning Models for Machine Translation
Zihao-Li  authored a paper 2 months ago
EMMA-500: Enhancing Massively Multilingual Adaptation of Large Language Models
View all activity

Organization Card
Community About org cards

MultiSynt is a collaborative initiative between OpenEuroLLM and EuroLLM focused on developing high-quality multilingual synthetic datasets for language model pretraining. By combining expertise from both organizations, MultiSynt aims to advance the creation of synthetic training data that supports diverse European languages to enable more inclusive AI development across languages.

Collections 2

Nemotron-cc Translations
A ~100B token sample from nemotron-cc translated into various languages. We use three different translation models: OPUS-MT, Tower+ 9B and Tower+ 72B.
Nemotron-cc Translations
A ~100B token sample from nemotron-cc translated into various languages. We use three different translation models: OPUS-MT, Tower+ 9B and Tower+ 72B.