You can deploy this model and start earning money today!

dkoterwa/
kor-sts-cosine-embedding-loss

Views

No views yet

Korean Semantic Textual Similarity (KorSTS) Dataset (with converted scores)

For a better dataset description, please visit this GitHub repository prepared by the authors of the article: LINK

This dataset was prepared by converting this dataset Specifically, I've changed the values of scores. If the score was equal or above 2.5, then it was assigned with label "1" (similar). In other cases, label "-1" was assigned (not similar). I've made these changes in order to apply… See the full description on the dataset page: https://huggingface.co/datasets/dkoterwa/kor-sts-cosine-embedding-loss.