README.md

metadata

dataset_info:
  features:
    - name: lang
      dtype: string
    - name: example_id
      dtype: string
    - name: query
      dtype: string
    - name: answer
      dtype: string
  splits:
    - name: train
      num_bytes: 4193271
      num_examples: 40548
  download_size: 2118715
  dataset_size: 4193271
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

mkqa filtered version

For a better dataset description, please visit the official site of the source dataset: LINK

This dataset was prepared by converting mkqa dataset.

I additionaly share the code which I used to convert the original dataset to make everything more clear

mkqa = load_dataset("mkqa", split="train").to_pandas()
needed_langs = ["en", "ar", "de", "es", "vi", "zh_cn"]
rows = []
for i, row in tqdm(mkqa.iterrows(), total=mkqa.shape[0]):
    for lang in needed_langs:
        rows.append([lang, row["example_id"], row["queries"][lang], row["answers"][lang][0]["text"]])
        
filtered_dataset = pd.DataFrame(rows, columns=["lang", "example_id", "query", "answer"])
filtered_dataset.dropna(inplace=True)
filtered_dataset.reset_index(drop=True, inplace=True)

How to download

from datasets import load_dataset
data = load_dataset("dkoterwa/oasst1_filtered_retrieval")