huu-ontocord commited on
Commit
dc71d24
1 Parent(s): f82fef1

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -2
README.md CHANGED
@@ -369,8 +369,7 @@ Therefore, when using, please only utilize the two columns text and url.
369
  ### Process for Creating CulturaY
370
  Firstly, to create CulturaY, we began with the HPLT dataset (version 1.1). This is also a notable difference between X and Y. While X was generated from cleaning data from Common Crawl (mC4, Oscar), Y was generated from cleaning raw data from the Internet Archive (HPLT). While Common Crawl is quite popular, data from the Internet Archive is less known and exploited, even though the data from both sources are similar. HPLT or CulturaY could be considered the first publicly released datasets originating from the Internet Archive. Using both CulturaX and CulturaY simultaneously will help your model have a more diverse source of data.
371
 
372
- Our pipeline is built based on Bloom's data cleaning pipeline: evaluating each document in the dataset according to criteria such as document length, perplexity, bad words ratio, etc., and removing documents that do not perform well in any of these criteria. Note that we perform data cleaning separately for each language, so the following explanations will apply to each language in the HPLT dataset rather than the entire data in HPLT.
373
-
374
  See our [Blog](https://www.ontocord.ai/blog/cultura-y) for more details.
375
 
376
  ### Citation
 
369
  ### Process for Creating CulturaY
370
  Firstly, to create CulturaY, we began with the HPLT dataset (version 1.1). This is also a notable difference between X and Y. While X was generated from cleaning data from Common Crawl (mC4, Oscar), Y was generated from cleaning raw data from the Internet Archive (HPLT). While Common Crawl is quite popular, data from the Internet Archive is less known and exploited, even though the data from both sources are similar. HPLT or CulturaY could be considered the first publicly released datasets originating from the Internet Archive. Using both CulturaX and CulturaY simultaneously will help your model have a more diverse source of data.
371
 
372
+ Our pipeline is built based on Bloom's data cleaning pipeline: evaluating each document in the dataset according to criteria such as document length, perplexity, bad words ratio, etc., and removing documents that do not perform well in any of these criteria.
 
373
  See our [Blog](https://www.ontocord.ai/blog/cultura-y) for more details.
374
 
375
  ### Citation