HunSum-2-abstractive is a Hungarian dataset containing news articles along with lead, titles and metadata.
You can evaluate an embedding model on this dataset using the following code:
import mteb
task = mteb.get_tasks(["HunSum2AbstractiveRetrieval"])
evaluator =… See the full description on the dataset page:
https://huggingface.co/datasets/mteb/HunSum2AbstractiveRetrieval.