1 dataset found

Tags: natural language processing

Filter Results

  • BEIR

    The BEIR dataset is a large-scale zero-shot evaluation dataset for information retrieval models, consisting of 13,000 documents and 1,000 questions.