metadata
license: cc-by-sa-4.0
language:
- ar
- en
- de
- ru
- tr
- ab
- af
- sq
- am
- az
- bm
- eu
- be
- bn
- br
- bg
- ca
- zh
- hr
- cs
- da
- nl
- et
- fi
- fr
- ceb
- bur
- myv
- gl
- ka
- el
- gu
- ht
- he
- hi
- hu
- is
- id
- ga
- it
- ja
- kk
- kpv
- ko
- ky
- lv
- lt
- mk
- ml
- gv
- mr
- mdf
- 'no'
- oc
- ps
- fa
- pl
- pt
- ro
- sa
- gd
- se
- sd
- si
- sk
- sl
- es
- sr
- tl
- ta
- tt
- uk
- ur
- ug
- uz
- vep
- vi
- cy
- wo
- sah
- ltg
- lij
- hsb
size_categories:
- 1M<n<10M
MorphScore
MorphScore is a tokenizer evaluation framework, which evaluates the extent to which a tokenizer segments words along morpheme boundaries. This repository contains the datasets used to calculate MorphScore. In total, we have datasetes for 86 languages, but only 70 languages have at least 100 items after filtering.
All datasets are derived from existing Universal Dependencies treebanks. In the table below, we link the source dataset for each language.
See the new preprint for details and see the GitHub repository for details on using MorphScore.
Languages
| Language | ISO 639-3 | ISO 15924 | Num. Words | Num. Unique Words | Data Source(s) |
|---|---|---|---|---|---|
| Abkhaz | abk | cyrl | 4518 | 3222 | UD_Abkhaz-AbNC v2.16 |
| Afrikaans | afr | latn | 7011 | 2093 | UD_Afrikaans-AfriBooms v2.16 |
| Albanian | sqi (tos) | latn | 897 | 702 | UD_Albanian-STAF v2.15 |
| Amharic | amh | ethi | 39 | 9 | UD_Amharic-ATT v2.11 |
| Armenian | hye | armn | 18906 | 8994 | UD_Armenian-ArmTDP v2.13 (except test = v2.10) |
| Azerbaijani | aze | latn | 368 | 251 | UD_Azerbaijani-TueCL v2.16 |
| Bambara | bam | latn | 6849 | 665 | UD_Bambara-CRB v2.13 |
| Basque | eus | latn | 50901 | 15932 | UD_Basque-BDT v2.14 |
| Belarusian | bel | cyrl | 104096 | 31129 | UD_Belarusian-HSE v2.16 |
| Bengali | ben | beng | 102 | 67 | UD_Bengali-BRU v2.11 |
| Bhojpuri | bho | deva | 1371 | 360 | UD_Bhojpuri-BHTB v2.16 |
| Breton | bre | latn | 3054 | 1021 | UD_Breton-KEB v2.12 |
| Bulgarian | bul | cyrl | 44412 | 15468 | UD_Bulgarian-BTB 2.16 |
| Buriat/Buryat | bur | cyrl | 3646 | 2322 | UD_Buryat-BDT v2.11 |
| Cantonese/Yue Chinese | yue | hant | 42 | 13 | UD_Cantonese-HK v2.12 |
| Catalan | cat | latn | 15495 | 3385 | UD_Catalan-AnCora v2.15 |
| Cebuano | ceb | latn | 264 | 136 | UD_Cebuano-GJA v2.15 |
| Mandarin Chinese | zho/cmn | hans | 915 | 140 | UD_Chinese-GSDSimp v2.13 |
| Croatian | hrv | latn | 75681 | 23363 | UD_Croatian-SET v2.11 (except train = v2.13) |
| Czech | ces | latn | 200387 | 44025 | UD_Czech-CAC v2.16 |
| Danish | dan | latn | 25662 | 8256 | UD_Danish-DDT v2.14 |
| Dutch | nld | latn | 42900 | 12512 | UD_Dutch-Alpino v2.14 |
| English | eng | latn | 31039 | 5844 | UD_English-EWT v2.16 |
| Erzya | myv | cyrl | 8147 | 4826 | UD_Erzya-JR v2.16 |
| Estonian | est | latn | 199597 | 62170 | UD_Estonian-EDT v2.16 |
| Finnish | fin | latn | 99308 | 41095 | UD_Finnish-TDT v2.16 |
| French | fra | latn | 103887 | 13508 | UD_French-GSD v2.16 |
| Galician | glg | latn | 29484 | 7113 | UD_Galician-CTG v2.12 |
| Georgian | kat | geor | 20148 | 8622 | UD_Georgian-GLC v2.16 |
| German | deu | latn | 949279 | 49239 | UD_German-HDT v2.16 |
| Greek | ell | grek | 21428 | 6986 | UD_Greek-GDT v2.16 |
| Gujarati | guj | gujr | 120 | 81 | UD_Gujarati-GujTB v2.14 |
| Haitian Creole | hat | latn | 950 | 37 | UD_Haitian_Creole-Adolphe v2.16 |
| Hebrew | heb | hebr | 43102 | 9768 | UD_Hebrew-HTB v2.15 |
| Hindi | hin | deva | 83865 | 4749 | UD_Hindi-HDTB v2.15 |
| Hungarian | hun | latn | 12349 | 8103 | UD_Hungarian-Szeged v2.11 |
| Icelandic | isl | latn | 329620 | 40686 | UD_Icelandic-IcePaHC v2.16 |
| Indonesian | ind | latn | 16844 | 3400 | UD_Indonesian-GSD v2.16 |
| Irish | gle | latn | 31911 | 7474 | UD_Irish-IDT v2.15 |
| Italian | ita | latn | 73790 | 10981 | UD_Italian-ISDT v2.16 |
| Japanese | jpn | jpan | 16898 | 5427 | UD_Japanese-GSDLUW v2.15 |
| Kazakh | kaz | cyrl | 4200 | 2738 | UD_Kazakh-KTB v2.11 |
| Komi-Zyrian | kpv | cyrl | 2670 | 2017 | UD_Komi_Zyrian-Lattice v2.16 |
| Korean | kor | hang | 239862 | 86349 | UD_Korean-Kaist v2.11 |
| Kirghiz/Kyrgyz | kir | cyrl | 11752 | 4659 | UD_Kyrgyz-KTMU v2.15 (train), v2.16 (test) |
| Latvian | lav | latn | 141550 | 40081 | UD_Latvian-LVTB v2.16 |
| Lithuanian | lit | latn | 32378 | 13140 | UD_Lithuanian-ALKSNIS v2.16, except test (v2.15) |
| Macedonian | mkd | cyrl | 285 | 241 | UD_Macedonian-MTB v2.13 |
| Malayalam | mal | mlym | 867 | 710 | UD_Malayalam-UFAL v2.12 |
| Manx | glv | latn | 2299 | 873 | UD_Manx-Cadhan v2.15 |
| Marathi | mar | deva | 1610 | 663 | UD_Marathi-UFAL v2.11 |
| Moksha | mdf | cyrl | 1733 | 1401 | UD_Moksha-JR v2.16 |
| Northern Sami | sme | latn | 10360 | 5024 | UD_North_Sami-Giella v2.15 (train), v2.12 (test) |
| Norwegian | nob | latn | 87394 | 15716 | UD_Norwegian-Bokmaal v2.16 |
| Occitan | oci | latn | 6751 | 2561 | UD_Occitan-TTB v2.16 |
| Pashto | pus | arab | 779 | 351 | UD_Pashto-Sikaram v2.16 |
| Persian | fas | arab | 102374 | 14847 | UD_Persian-PerDT v2.16, except dev (v2.12) |
| Polish | pol | latn | 131836 | 42545 | UD_Polish-PDB v2.13 |
| Portuguese | por | latn | 69430 | 13074 | UD_Portuguese-CINTIL v2.13 |
| Romanian | ron | latn | 72968 | 19108 | UD_Romanian-RRT v2.13 |
| Russian | rus | cyrl | 590060 | 105749 | UD_Russian-SynTagRus v2.16 |
| Sanskrit | san | deva | 144583 | 32671 | UD_Sanskrit-Vedic v2.14 |
| Scottish Gaelic | gla | latn | 27181 | 3495 | UD_Scottish_Gaelic-ARCOSG v2.16 |
| Serbian | srp | latn | 37324 | 12278 | UD_Serbian-SET v2.11, except train (v2.13) |
| Sindhi | snd | arab | 3584 | 1132 | UD_Sindhi-Isra v2.16 |
| Sinhala | sin | sinh | 354 | 284 | UD_Sinhala-STB v2.12 |
| Slovak | slk | latn | 35545 | 16169 | UD_Slovak-SNK v2.16 |
| Slovenian | slv | latn | 93048 | 32134 | UD_Slovenian-SSJ v2.16 |
| Spanish | spa | latn | 134449 | 17341 | UD_Spanish-AnCora v2.15 |
| Swedish | swe | latn | 29514 | 8135 | UD_Swedish-LinES v2.16 |
| Tagalog | tgl | latn | 212 | 78 | UD_Tagalog-Ugnayan v2.6 |
| Tamil | tam | taml | 4499 | 2178 | UD_Tamil-TTB v2.16 (except dev = v2.13) |
| Tatar | tat | cyrl | 899 | 708 | UD_Tatar-NMCTT v2.11 |
| Turkish | tur | latn | 71984 | 32587 | UD_Turkish-Kenet train (v2.13), dev (v2.10), test (v2.9) |
| Ukrainian | ukr | cyrl | 44288 | 21584 | UD_Ukrainian-IU v2.11 |
| Urdu | urd | arab | 30825 | 2450 | UD_Urdu-UDTB v2.14 |
| Uighur/Uyghur | uig | arab | 10427 | 3969 | UD_Uyghur-UDT v2.13 |
| Uzbek | uzb | latn | 2466 | 1934 | UD_Uzbek-UT v2.16 |
| Veps | vep | latn | 598 | 399 | UD_Veps-VWT v2.16 |
| Vietnamese | vie | latn | 32 | 29 | UD_Vietnamese-VTB v2.11 (except dev = v2.12) |
| Welsh | cym | latn | 11781 | 2860 | UD_Welsh-CCG v2.16 |
| Wolof | wol | latn | 9211 | 1734 | UD_Wolof-WTB v2.13 |
| Yakut | sah | cyrl | 479 | 330 | UD_Yakut-YKTDT v2.12 |
| Latgalian | ltg | latn | 69 | 63 | UD_Latgalian-Cairo v2.16 |
| Ligurian | lij | latn | 1318 | 637 | UD_Ligurian-GLT train (v2.9), test (v2.16) |
| Upper Sorbian | hsb | latn | 3797 | 2498 | UD_Upper_Sorbian-UFAL train (v2.3), test (v2.8) |
How to Cite
@inproceedings{arnett2025alignment,
author = {Arnett, Catherine and Hudspeth, Marisa and O'Connor, Brendan},
title = {{Evaluating Morphological Alignment of Tokenizers in 70 Languages}},
year = {2025},
booktitle={Proceedings of the ICML 2025 Tokenization Workshop (TokShop)},
url = {https://arxiv.org/abs/2507.06378}
}
@inproceedings{arnett-bergen-2025-language,
title = "Why do language models perform worse for morphologically complex languages?",
author = "Arnett, Catherine and
Bergen, Benjamin",
editor = "Rambow, Owen and
Wanner, Leo and
Apidianaki, Marianna and
Al-Khalifa, Hend and
Eugenio, Barbara Di and
Schockaert, Steven",
booktitle = "Proceedings of the 31st International Conference on Computational Linguistics",
month = jan,
year = "2025",
address = "Abu Dhabi, UAE",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.coling-main.441/",
pages = "6607--6623"
}