does IndicCorpus and OSCAR corpus the same ?

AI4Bharat / indicnlp_corpus

Description Describes the IndicNLP corpus and associated datasets

150 stars 24 forks source link

does IndicCorpus and OSCAR corpus the same ? #11

Open StephennFernandes opened 2 years ago

StephennFernandes commented 2 years ago

Hey there, Does IndicCorpus and OSCAR corpus come from the same source. ie: CommonCrawl ? i have been thinking to combining OSCAR + IndicCorpus to get a better and bigger corpus(with deduplication). Just wanted to confirm if the IndicCorpus and OSCAR are the same corpus at source or not.

anoopkunchukuttan commented 2 years ago

The version of IndicCorpus does not contain Oscar. However, the newer version that you can find here contains OSCAR as a subset - https://indicnlp.ai4bharat.org/corpora/

StephennFernandes commented 2 years ago

@anoopkunchukuttan how do i find the previous version of the corpus (that doesn't contain Oscar) ? btw, the Oscar Corpus is generated from common crawl corpus. when you said that the IndicCorpus does not contain Oscar. does it mean the IndicCorpus does not contain content from common crawl ?

StephennFernandes commented 2 years ago

@anoopkunchukuttan Hello Sir, just a follow up on the previous question.

does the corpus contain wikipedia content ?
is there a way i could get the previous version of Indic Corpus that doesn't contain oscar corpus ?
is there a way i could get the corpus in unshuffled format ?

as i would be adding content from oscar corpus separately. also additionally is there a way i could get the corpus in unshuffled format.