Hugging Face's logo

Datasets:
everflip-app
/
flashcards

Tasks:
Translation
Text Classification
Modalities:
Tabular
Text
Formats:
csv
Size:
10K - 100K
Tags:
language-learning
flashcards
spaced-repetition
vocabulary
jlpt
hsk
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
flashcards / README.md
EverFlip's picture
EverFlip
Initial release: 24,164 cards, 81 languages (CC-BY-4.0)
0f76c38 verified
preview code
|
Raw
History Blame Contribute Delete
1.71 kB
metadata 
license: cc-by-4.0
pretty_name: EverFlip Multilingual Flashcard Corpus
size_categories:
  - 10K<n<100K
task_categories:
  - translation
  - text-classification
tags:
  - language-learning
  - flashcards
  - spaced-repetition
  - vocabulary
  - jlpt
  - hsk
  - topik
  - cefr
source_datasets:
  - original
configs:
  - config_name: cards
    data_files: cards.csv
  - config_name: decks
    data_files: decks.csv
  - config_name: languages
    data_files: languages.csv

EverFlip Multilingual Flashcard Corpus

24,164 flashcards across 81 languages, organised into 1,573 themed decks and exam ladders (JLPT, HSK, TOPIK, DELE / DELF / Goethe / CILS / CAPLE).

Each card is a factual lexical correspondence — a word or phrase in the target language, its reading (for non-Latin scripts), and its English meaning.

Files

  • cards.csv — every card (24,164 rows): language_code, language, deck_slug, deck_name, front, meaning_en, back_raw, translation_raw. Use meaning_en for the normalized English meaning (it resolves the field across Latin and non-Latin scripts).
  • decks.csv — deck catalogue (1,573 rows).
  • languages.csv — language catalogue (81 rows).
  • datapackage.json — frictionlessdata.io Tabular Data Package descriptor.

License

CC-BY-4.0 — free to share, adapt, and train on, with attribution to EverFlip (https://everflip.app). Only the factual card data is released; the app, its FSRS scheduling, and any editorial prose are not included.

Source & citation