data/README.md

metadata

language:
  - en
  - da
  - sv
  - vi
  - hu
  - fa
  - tr
  - cs
  - ar
  - el
  - id
  - nl
  - pl
  - pt
  - it
  - ja
  - fr
  - es
  - de
  - zh
  - ru
task_categories:
  - question-answering
language_bcp47:
  - pt-BR
tags:
  - arithmetic
  - addition
  - multilingual
  - synthetic
  - tokenization
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/*/train.chunk.*.jsonl
      - split: val
        path: data/*/val.chunk.*.jsonl
    default: true
  - config_name: eng_Latn
    data_files:
      - split: train
        path: data/eng_Latn/train.chunk.*.jsonl
      - split: val
        path: data/eng_Latn/val.chunk.*.jsonl
  - config_name: dan_Latn
    data_files:
      - split: train
        path: data/dan_Latn/train.chunk.*.jsonl
      - split: val
        path: data/dan_Latn/val.chunk.*.jsonl
  - config_name: swe_Latn
    data_files:
      - split: train
        path: data/swe_Latn/train.chunk.*.jsonl
      - split: val
        path: data/swe_Latn/val.chunk.*.jsonl
  - config_name: vie_Latn
    data_files:
      - split: train
        path: data/vie_Latn/train.chunk.*.jsonl
      - split: val
        path: data/vie_Latn/val.chunk.*.jsonl
  - config_name: hun_Latn
    data_files:
      - split: train
        path: data/hun_Latn/train.chunk.*.jsonl
      - split: val
        path: data/hun_Latn/val.chunk.*.jsonl
  - config_name: fas_Arab
    data_files:
      - split: train
        path: data/fas_Arab/train.chunk.*.jsonl
      - split: val
        path: data/fas_Arab/val.chunk.*.jsonl
  - config_name: tur_Latn
    data_files:
      - split: train
        path: data/tur_Latn/train.chunk.*.jsonl
      - split: val
        path: data/tur_Latn/val.chunk.*.jsonl
  - config_name: ces_Latn
    data_files:
      - split: train
        path: data/ces_Latn/train.chunk.*.jsonl
      - split: val
        path: data/ces_Latn/val.chunk.*.jsonl
  - config_name: arb_Arab
    data_files:
      - split: train
        path: data/arb_Arab/train.chunk.*.jsonl
      - split: val
        path: data/arb_Arab/val.chunk.*.jsonl
  - config_name: ell_Grek
    data_files:
      - split: train
        path: data/ell_Grek/train.chunk.*.jsonl
      - split: val
        path: data/ell_Grek/val.chunk.*.jsonl
  - config_name: ind_Latn
    data_files:
      - split: train
        path: data/ind_Latn/train.chunk.*.jsonl
      - split: val
        path: data/ind_Latn/val.chunk.*.jsonl
  - config_name: nld_Latn
    data_files:
      - split: train
        path: data/nld_Latn/train.chunk.*.jsonl
      - split: val
        path: data/nld_Latn/val.chunk.*.jsonl
  - config_name: pol_Latn
    data_files:
      - split: train
        path: data/pol_Latn/train.chunk.*.jsonl
      - split: val
        path: data/pol_Latn/val.chunk.*.jsonl
  - config_name: por_Latn
    data_files:
      - split: train
        path: data/por_Latn/train.chunk.*.jsonl
      - split: val
        path: data/por_Latn/val.chunk.*.jsonl
  - config_name: ita_Latn
    data_files:
      - split: train
        path: data/ita_Latn/train.chunk.*.jsonl
      - split: val
        path: data/ita_Latn/val.chunk.*.jsonl
  - config_name: jpn_Jpan
    data_files:
      - split: train
        path: data/jpn_Jpan/train.chunk.*.jsonl
      - split: val
        path: data/jpn_Jpan/val.chunk.*.jsonl
  - config_name: fra_Latn
    data_files:
      - split: train
        path: data/fra_Latn/train.chunk.*.jsonl
      - split: val
        path: data/fra_Latn/val.chunk.*.jsonl
  - config_name: spa_Latn
    data_files:
      - split: train
        path: data/spa_Latn/train.chunk.*.jsonl
      - split: val
        path: data/spa_Latn/val.chunk.*.jsonl
  - config_name: deu_Latn
    data_files:
      - split: train
        path: data/deu_Latn/train.chunk.*.jsonl
      - split: val
        path: data/deu_Latn/val.chunk.*.jsonl
  - config_name: cmn_Hani
    data_files:
      - split: train
        path: data/cmn_Hani/train.chunk.*.jsonl
      - split: val
        path: data/cmn_Hani/val.chunk.*.jsonl
  - config_name: rus_Cyrl
    data_files:
      - split: train
        path: data/rus_Cyrl/train.chunk.*.jsonl
      - split: val
        path: data/rus_Cyrl/val.chunk.*.jsonl
  - config_name: digit
    data_files:
      - split: train
        path: data/digit/train.chunk.*.jsonl
      - split: val
        path: data/digit/val.chunk.*.jsonl

Multilingual Addition Dataset

Synthetic dataset of addition problems of the form a+b=answer, where a and b are written-form representations of integers in 21 languages, plus a 22nd split using raw digit strings.

Task format

Each sample contains:

field	type	description
a_str	str	written-form (or digit) representation of a
a_digit	int	integer value of a
b_str	str	written-form (or digit) representation of b
b_digit	int	integer value of b
answer	str	written-form (or digit) of a + b
answer_digit	int	integer value of a + b
text	str	"{a_str}+{b_str}={answer}" (completion target)
question	str	"{a_str}+{b_str}=" (prompt)
lang	str	language tag, e.g. eng_Latn, or digit

Numbers range from 0 to 999 for both a and b (answers up to 1998).

Languages

lang	train	val
eng_Latn	900,000	100,000
dan_Latn	900,000	100,000
swe_Latn	900,000	100,000
vie_Latn	900,000	100,000
hun_Latn	900,000	100,000
fas_Arab	900,000	100,000
tur_Latn	900,000	100,000
ces_Latn	900,000	100,000
arb_Arab	900,000	100,000
ell_Grek	900,000	100,000
ind_Latn	900,000	100,000
nld_Latn	900,000	100,000
pol_Latn	900,000	100,000
por_Latn	900,000	100,000
ita_Latn	900,000	100,000
jpn_Jpan	900,000	100,000
fra_Latn	900,000	100,000
spa_Latn	900,000	100,000
deu_Latn	900,000	100,000
cmn_Hani	900,000	100,000
rus_Cyrl	900,000	100,000
digit	900,000	100,000

Generation

python create_multilingual_addition_data.py \
    --hf_repo_id flexitok/multilingual-addition \
    --publish_to_hf \
    --a_min 0 --a_max 999 \
    --seed 42 --train_ratio 0.9