Fula Pulaar <-> English Resource (Sentence Pairs)

[nikisix]

Hi I would like to contribute a Pulaar translator model, but need pointed to the the sentence pairs. Can anyone help me out?

[juliakreutzer]

Hi @nikisix ! It looks like JW300 which we used as source for other languages does not include Pulaar. On the OPUS website you can look for other corpora: https://opus.nlpl.eu/ -- It lists CCAligned, Wikimedia, Ubuntu, QED for Fula, but I'm not sure if it's Pulaar. The CCAligned corpus was previously found (https://arxiv.org/abs/2103.12028) to contain mostly noise for Fula, so I would not recommend using it. Perhaps Wikimedia, Ubuntu or QED? These might be quite domain-specific though.

[nikisix]

Haven't used those last sources you mention before. I did notice JW300 has code 'fub' for pular defined, but no supporting data files unfortunately.

[flutter-painter]

Hi @nikisix,

NLLB contains 400k lines : https://opus.nlpl.eu/NLLB/en&ff/v1/NLLB
I started a repo to list resources : https://github.com/flutter-painter/awesome_fula_nl_resources
it also includes religious text and common phrases

If you want to contribute to language pairs specifically on fula, may I suggest you take a look at : https://cawoylel.com/
they are looking for contributors (github and hugging face)
You might want to take a look at the speech compatible translator : https://firtanam.cawoylel.com/