xiaomi-research / MiLMMT-46-12B-v1.0

Other12B82019Updated Aug 12, 2026

Downloads

820

Likes

19

Last updated

Aug 12, 2026

License

gemma

MiLMMT-46-12B-v1.0 is an LLM-based translation model. Its training builds on Gemma3-12B through four stages: (1) continual pretraining on 143 billion tokens of monolingual and parallel data across 46 languages, yielding MiLMMT-46-12B-Pretrain; (2) supervised finetuning, yieldi…

Tasks & library

  • translation
  • transformers

Tags

  • transformers
  • gemma3
  • image-text-to-text
  • translation

Technical specifications

Parameters12B
Librarytransformers
LanguagesNot available
Licensegemma