turkish-nlp-suite/

temiz-mC4

Temiz mC4 is the cleaned version of CulturaX corpus' Turkish split. This corpus is a part of large scale Turkish corpus Bella Turca. For more details about Bella Turca, please refer to the publication.

split num instances size num of words

train 76.432.893 168GB 21.06B

Total 76.432.893 168GB 21.06B

This collection includes web text, crawled from internet for mC4 corpus. CulturaX is even refined version of mC4 with quality filtering and… See the full description on the dataset page: https://huggingface.co/datasets/turkish-nlp-suite/temiz-mC4.