Instructions to use tartuNLP/septilang with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Fairseq
How to use tartuNLP/septilang with Fairseq:
from fairseq.checkpoint_utils import load_model_ensemble_and_task_from_hf_hub models, cfg, task = load_model_ensemble_and_task_from_hf_hub( "tartuNLP/septilang" ) - Notebooks
- Google Colab
- Kaggle
A Modular Translation Model for 7 Languages
This model supports translation in all directions between the following languages: et, en, de, ru, fi, lt, lv.
The model uses a modular architecture, where each language has its own encoder and decoder that is used for all translation direction combinations. The model can be used with our custom version of FairSeq and with our translation API components (API and NMT workers). Additionally, it is fully compatible with the MTee platform and its NMT workers.
| Files: | |
|---|---|
| Fairseq translation model | modular_model.pt |
| SentecePiece models | sp-model.{lang}.model |
| translation model vocabularies | dict.{lang}.txt |