Datasets:

Modalities:
Text
Formats:
parquet
ArXiv:
Libraries:
Datasets
pandas
License:
albertvillanova HF staff commited on
Commit
805090d
1 Parent(s): 95696bf

Update dataset card (#6)

Browse files

- Update dataset card (7616737cc2c6715bd6325183863db15c8ac16068)

Files changed (1) hide show
  1. README.md +60 -20
README.md CHANGED
@@ -120,7 +120,7 @@ task_categories:
120
  - translation
121
  task_ids: []
122
  paperswithcode_id: opus-100
123
- pretty_name: Opus100
124
  config_names:
125
  - af-en
126
  - am-en
@@ -3232,7 +3232,7 @@ configs:
3232
  path: ru-zh/test-*
3233
  ---
3234
 
3235
- # Dataset Card for Opus100
3236
 
3237
  ## Table of Contents
3238
  - [Dataset Description](#dataset-description)
@@ -3260,19 +3260,23 @@ configs:
3260
 
3261
  ## Dataset Description
3262
 
3263
- - **Homepage:** [Link](http://opus.nlpl.eu/opus-100.php)
3264
- - **Repository:** [GitHub](https://github.com/EdinburghNLP/opus-100-corpus)
3265
- - **Paper:** [ARXIV](https://arxiv.org/abs/2004.11867)
3266
- - **Leaderboard:**
3267
- - **Point of Contact:**
 
3268
 
3269
  ### Dataset Summary
3270
 
3271
- OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side. The corpus covers 100 languages (including English). Selected the languages based on the volume of parallel data available in OPUS.
 
 
 
3272
 
3273
  ### Supported Tasks and Leaderboards
3274
 
3275
- [More Information Needed]
3276
 
3277
  ### Languages
3278
 
@@ -3284,15 +3288,16 @@ OPUS-100 contains approximately 55M sentence pairs. Of the 99 language pairs, 44
3284
 
3285
  ```
3286
  {
3287
- "ca": "El departament de bombers té el seu propi equip d'investigació.",
3288
- "en": "Well, the fire department has its own investigative unit."
 
 
3289
  }
3290
  ```
3291
 
3292
  ### Data Fields
3293
 
3294
- - `src_tag`: `string` text in source language
3295
- - `tgt_tag`: `string` translation of source language in target language
3296
 
3297
  ### Data Splits
3298
 
@@ -3356,14 +3361,49 @@ The dataset is split into training, development, and test portions. Data was pre
3356
 
3357
  ### Citation Information
3358
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3359
  ```
3360
- @misc{zhang2020improving,
3361
- title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},
3362
- author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},
3363
- year={2020},
3364
- eprint={2004.11867},
3365
- archivePrefix={arXiv},
3366
- primaryClass={cs.CL}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3367
  }
3368
  ```
3369
 
 
120
  - translation
121
  task_ids: []
122
  paperswithcode_id: opus-100
123
+ pretty_name: OPUS-100
124
  config_names:
125
  - af-en
126
  - am-en
 
3232
  path: ru-zh/test-*
3233
  ---
3234
 
3235
+ # Dataset Card for OPUS-100
3236
 
3237
  ## Table of Contents
3238
  - [Dataset Description](#dataset-description)
 
3260
 
3261
  ## Dataset Description
3262
 
3263
+ - **Homepage:** https://opus.nlpl.eu/OPUS-100
3264
+ - **Repository:** https://github.com/EdinburghNLP/opus-100-corpus
3265
+ - **Paper:** https://arxiv.org/abs/2004.11867
3266
+ - **Paper:** https://aclanthology.org/L10-1473/
3267
+ - **Leaderboard:** [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
3268
+ - **Point of Contact:** [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
3269
 
3270
  ### Dataset Summary
3271
 
3272
+ OPUS-100 is an English-centric multilingual corpus covering 100 languages.
3273
+
3274
+ OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side. The corpus covers 100 languages (including English).
3275
+ The languages were selected based on the volume of parallel data available in OPUS.
3276
 
3277
  ### Supported Tasks and Leaderboards
3278
 
3279
+ Translation.
3280
 
3281
  ### Languages
3282
 
 
3288
 
3289
  ```
3290
  {
3291
+ "translation": {
3292
+ "ca": "El departament de bombers el seu propi equip d'investigació.",
3293
+ "en": "Well, the fire department has its own investigative unit."
3294
+ }
3295
  }
3296
  ```
3297
 
3298
  ### Data Fields
3299
 
3300
+ - `translation` (`dict`): Parallel sentences for the pair of languages.
 
3301
 
3302
  ### Data Splits
3303
 
 
3361
 
3362
  ### Citation Information
3363
 
3364
+ If you use this corpus, please cite the paper:
3365
+ ```bibtex
3366
+ @inproceedings{zhang-etal-2020-improving,
3367
+ title = "Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation",
3368
+ author = "Zhang, Biao and
3369
+ Williams, Philip and
3370
+ Titov, Ivan and
3371
+ Sennrich, Rico",
3372
+ editor = "Jurafsky, Dan and
3373
+ Chai, Joyce and
3374
+ Schluter, Natalie and
3375
+ Tetreault, Joel",
3376
+ booktitle = "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics",
3377
+ month = jul,
3378
+ year = "2020",
3379
+ address = "Online",
3380
+ publisher = "Association for Computational Linguistics",
3381
+ url = "https://aclanthology.org/2020.acl-main.148",
3382
+ doi = "10.18653/v1/2020.acl-main.148",
3383
+ pages = "1628--1639",
3384
+ }
3385
  ```
3386
+ and, please, also acknowledge OPUS:
3387
+ ```bibtex
3388
+ @inproceedings{tiedemann-2012-parallel,
3389
+ title = "Parallel Data, Tools and Interfaces in {OPUS}",
3390
+ author = {Tiedemann, J{\"o}rg},
3391
+ editor = "Calzolari, Nicoletta and
3392
+ Choukri, Khalid and
3393
+ Declerck, Thierry and
3394
+ Do{\u{g}}an, Mehmet U{\u{g}}ur and
3395
+ Maegaard, Bente and
3396
+ Mariani, Joseph and
3397
+ Moreno, Asuncion and
3398
+ Odijk, Jan and
3399
+ Piperidis, Stelios",
3400
+ booktitle = "Proceedings of the Eighth International Conference on Language Resources and Evaluation ({LREC}'12)",
3401
+ month = may,
3402
+ year = "2012",
3403
+ address = "Istanbul, Turkey",
3404
+ publisher = "European Language Resources Association (ELRA)",
3405
+ url = "http://www.lrec-conf.org/proceedings/lrec2012/pdf/463_Paper.pdf",
3406
+ pages = "2214--2218",
3407
  }
3408
  ```
3409