Brotis-embed-spanish-v0.1Author: Convence
A high-quality Spanish text embedding dataset containing 23,410 unique
triplets (anchor, positive, negative) across 30 technical domains.
Each sample is a triplet designed for contrastive / triplet-loss training of
text embedding models:
Field
Description
anchor
A query or question in a technical domain
positive
A semantically relevant, detailed answer
negative
An unrelated passage from a different… See the full description on the dataset page:
https://huggingface.co/datasets/Convence/brotis-embed-spanish-v0.1.
