Create dataset loader for MyWSL2023

[SamuelCahyawijaya]

Dataloader name: mywsl2023/mywsl2023.py DataCatalogue: http://seacrowd.github.io/seacrowd-catalogue/card.html?mywsl2023

Dataset	mywsl2023
Description	This dataset contains pictures of hand gestures corresponding to ten commonly-used Malaysian Sign Language (XML) words. Gestures are performed by five university students who belong to different ethnic groups and are proficient in XML. Each gesture class contains 350 instances.
Subsets	-
Languages	xml
Tasks	Language Modeling
License	Creative Commons Attribution 4.0 (cc-by-4.0)
Homepage	https://data.mendeley.com/datasets/zvk55p7ktd/1
HF URL	-
Paper URL	https://www.sciencedirect.com/science/article/pii/S2352340923004560

[Alex-HaochenLi]

self-assign

[Alex-HaochenLi]

Hello @holylovenia @SamuelCahyawijaya @sabilmakbar,

After checking the data, I find that this dataset is used for sign languages, which means that data are all pictures. I don't think it could support language modeling task. Please have a check.

[sabilmakbar]

Apologies for laterep, we're on it rn as we found some other datasets having similar issues w/ this one.

[holylovenia]

@Alex-HaochenLi Sorry for the mistake. I've fixed the datasheet to have Sign Language Recognition instead of Language Modeling. Probably we have to add a new task in the constants.py to cater to this dataloader, though.

What do you think?

cc: @sabilmakbar @SamuelCahyawijaya

[holylovenia]

Hi @Alex-HaochenLi, @sabilmakbar has kindly added the SIGN_LANGUAGE_RECOGNITION task so you can proceed with the dataloader implementation.

[Enliven26]

self-assign

[github-actions[bot]]

Hi @, may I know if you are still working on this issue? Please let @holylovenia @SamuelCahyawijaya @sabilmakbar know if you need any help.

[Enliven26]

yes