Datasets:

ai4bharat
/

sangraha

Tasks:

Text Generation

Modalities:

Formats:

Languages:

Size:

ArXiv:

Tags:

language-modeling

Libraries:

License:

Dataset card Data Studio Files Files and versions Community

safikhan commited on 5 days ago

Commit

8b813c3

·

verified ·

1 Parent(s): 016955b

Update README.md

Files changed (1) hide show

README.md +0 -4

README.md CHANGED Viewed

@@ -265,10 +265,6 @@ size_categories:
 Sangraha is the largest high-quality, cleaned Indic language pretraining data containing 251B tokens summed up over 22 languages, extracted from curated sources, existing multilingual corpora and large scale translations.
-**Coming Soon**:
-- Sangraha Synthetic - Translated and Romanised English Wikimedia data.
-- Sangraha Verified - Hindi YouTube transcribed data.
 **More information**:


265
266	Sangraha is the largest high-quality, cleaned Indic language pretraining data containing 251B tokens summed up over 22 languages, extracted from curated sources, existing multilingual corpora and large scale translations.
267




268
269	More information:
270