TowerBlocks is the dataset used to train TowerInstruct-v0.1, a language model specialized for translation tasks such as machine translation (e.g. general, document, terminology-aware or context-aware translation), automatic post edition, named-entity recognition, gramatical error correction, and paraphrase generation.
Curated by: Unbabel, Instituto Superior Técnico, CentraleSupélec, University of Paris-Saclay;
Language(s) (NLP): English, Portuguese… See the full description on the dataset page:
https://huggingface.co/datasets/Unbabel/TowerBlocks-v0.2.
