Downloads · 30 days
12
16% of all-time downloads
JosueG/adja-nmt-nllb-600m-forward-r10ks4k-seed42
adja-nmt-nllb-600m-forward-r10ks4k-seed42 is a translation model from JosueG. Use it when you need text moved from one language to another. It is set up for transformers. The card lists the license as cc-by-nc-4.0.
This is the thesis-final French-to-Adja machine translation checkpoint for the May 2026 CS thesis release. It is one of the two NLLB models used to make the French-Adja MT work inspectable and reproducible.
Downloads · 30 days
12
16% of all-time downloads
All-time downloads
76
Public
Parameters
615M
2.5 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors2.5 GB · 99%
From the Hugging Face model README
This is the thesis-final French-to-Adja machine translation checkpoint for the May 2026 CS thesis release. It is one of the two NLLB models used to make the French-Adja MT work inspectable and reproducible.
This model is the forward MT artifact: French source text to Adja target text. It supports the thesis experiments on grammar-guided dataset construction, low-resource NMT fine-tuning, and the speech-to-text-to-speech pipeline.
facebook/nllb-200-distilled-600Mfra_Latn -> aj_LatnJosueG/french-adja-parallel-corpusFrejusGdm/cs-thesis-may-2026The seed-42 NLLB-600M full/module runs are the strongest MT baseline in the release. The public GitHub result tables report about 21-22 BLEU and 29 chrF on the held-out MT evaluation for the full-corpus NLLB setting.
See:
results/mt/summary/all_results.csvresults/mt/README_thesis_qualitative_decode.mdIf you use this model, cite:
Josue Godeme. 2026. CS Thesis May 2026: French-Adja MT and Adja Speech Experiments. https://github.com/FrejusGdm/cs-thesis-may-2026