Investigate Hindi datasets for NER and TC about nlp-recipes HOT 7 CLOSED

I also found this repo online:
https://github.com/goru001/nlp-for-hindi
https://github.com/NirantK/hindi2vec

Not sure if this is helpful but worth a look?
@miguelgfierro , @saidbleik

from nlp-recipes.

I have ran the text classification for hindi bbc news dataset in https://github.com/NirantK/hindi2vec. As the dataset is small with total 3468 examples, and very imbalanced with some classes having under 20 examples, I chose only the top 6 classes and achieved result
accuracy 0.73 653
macro avg 0.67 0.66 0.66 653
weighted avg 0.72 0.73 0.72 653

from nlp-recipes.

One option is xnli dataset that has Hindi translations, @hlums had been working work it

from nlp-recipes.

https://github.com/SilentFlame/Named-Entity-Recognition/tree/master/Twitterdata

from nlp-recipes.

hey @mithun-prasad, I found this repo: https://github.com/juand-r/entity-recognition-datasets/blob/master/README.rst#hindi. There are 4 datasets, do you think they will be a good fit for us?

from nlp-recipes.

Here is a parallel dataset: http://www.cfilt.iitb.ac.in/~moses/iitb_en_hi_parallel/ (Hindi/English)
It has the Creative Commons Attribution-NonCommercial 4.0 International License: https://creativecommons.org/licenses/by-nc/4.0/
@saidbleik , will this dataset work?

from nlp-recipes.

hey @daden-ms, are you planning to push your results to the repo?

from nlp-recipes.