synczech50 / README.md
simecek's picture
Update README.md
b469a47 verified
metadata
dataset_info:
  features:
    - name: question
      dtype: string
    - name: optionA
      dtype: string
    - name: optionB
      dtype: string
    - name: optionC
      dtype: string
    - name: optionD
      dtype: string
    - name: solution
      dtype: string
  splits:
    - name: train
      num_bytes: 5813
      num_examples: 50
  download_size: 7834
  dataset_size: 5813
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: apache-2.0
task_categories:
  - question-answering
language:
  - cs
size_categories:
  - n<1K

Synthetic dataset of 50 Czech A/B/C/D questions for quick evaluation how the LMM understands Czech and Czech specific knowledge. Draft generated with the help of GPT4 and later manualy validated and edited.

from datasets import load_dataset