search

owos / afri_augs

Data Augmentation for Generative models
1 stars 5 forks source link

Create finetune and eval script for NLLB #1

Open owos opened 10 months ago

owos commented 10 months ago

The issue requires a finetunig script and evaluation script for NLLB.

Script should take in CLI arguments and should be uses to run a baseline experiment on the Mafand dataset for 3 epoch. Please look at the "A few thousand translation" paper for more details.

Useful repo: https://github.com/masakhane-io/afriqa/blob/main/translation_script/nllb_translate.py https://github.com/TartuNLP/m2m-100-finetune/tree/main

Tundeh commented 10 months ago

I would like to be assigned this task

hemhemoh commented 10 months ago

I would like to be assigned this task

Hi Tunde, I am working on the fine-tuning script for this already, maybe you could work on the eval.

Tundeh commented 10 months ago

I would like to be assigned this task

Hi Tunde, I am working on the fine-tuning script for this already, maybe you could work on the eval.

Okay

Tundeh commented 10 months ago

I will work on the eval script