README.md

metadata

license: cc-by-sa-4.0
language:
  - sl
dataset_info:
  features:
    - name: Question
      dtype: string
    - name: Answer
      dtype: string
    - name: Context
      dtype: string
    - name: URL
      dtype: string
    - name: Source
      dtype: string
  splits:
    - name: train
      num_bytes: 3690156
      num_examples: 10150
  download_size: 1522809
  dataset_size: 3690156
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

Cultural Heritage Q&A Dataset

This dataset contains 10,000 prepared questions and answers for fine-tuning language models on question-answering (Q&A) tasks. The dataset was manually created by INZ and includes additional context and source information for each entry.

Dataset Overview

Each example in the dataset includes:

• Question: The question text.
• Answer: The answer text.
• Context: Additional context that can provide further details relevant to the question and answer.
• Source: The URL to the original source of the content.

The dataset is intended for research and development purposes in the field of natural language processing (NLP), especially for training and evaluating Q&A models.

License

This dataset is distributed under the CC-BY-SA-4.0 license. Please ensure proper attribution if you use or share the dataset.

Authors

In alphabetical order:

• Ana Cvek
• Tamara Logar
• Andrej Pančur
• Jure Skubic
• Mojca Šorn

Funding

The development of the Cultural Heritage Dataset was partially funded by the PoVeJMo project, which aims to develop large language models for the Slovenian language. The project PoVeJMo is cofinanced by:

Citation

If you use this dataset, please cite it as follows:

@misc{INZ_QA_Dataset, title={INZ Cultural Heritage Q&A Dataset}, author={Ana Cvek and Tamara Logar and Andrej Pančur and Jure Skubic and Mojca Šorn}, year={2024}, publisher={INZ}, license={CC-BY-SA-4.0}, url={https://huggingface.co/klokedm/povejmo_CH} }