Update files from the datasets library (from 1.16.0)
Browse filesRelease notes: https://github.com/huggingface/datasets/releases/tag/1.16.0
- README.md +7 -1
- dataset_infos.json +1 -1
- dummy/en-mr/{2020.11.9 → 2021.7.22}/dummy_data.zip +0 -0
- dummy/eo-nl/{2020.11.9 → 2021.7.22}/dummy_data.zip +0 -0
- dummy/es-gl/{2020.11.9 → 2021.7.22}/dummy_data.zip +0 -0
- dummy/es-pt/{2020.11.9 → 2021.7.22}/dummy_data.zip +0 -0
- dummy/fr-ru/{2020.11.9 → 2021.7.22}/dummy_data.zip +0 -0
- tatoeba.py +8 -6
README.md
CHANGED
@@ -376,8 +376,9 @@ task_categories:
|
|
376 |
task_ids:
|
377 |
- machine-translation
|
378 |
paperswithcode_id: tatoeba
|
|
|
379 |
---
|
380 |
-
# Dataset Card
|
381 |
|
382 |
## Table of Contents
|
383 |
- [Dataset Description](#dataset-description)
|
@@ -413,6 +414,7 @@ paperswithcode_id: tatoeba
|
|
413 |
|
414 |
### Dataset Summary
|
415 |
|
|
|
416 |
|
417 |
To load a language pair which isn't part of the config, all you need to do is specify the language code as pairs.
|
418 |
You can find the valid pairs in Homepage section of Dataset Description: http://opus.nlpl.eu/Tatoeba.php
|
@@ -420,6 +422,10 @@ E.g.
|
|
420 |
|
421 |
`dataset = load_dataset("tatoeba", lang1="en", lang2="he")`
|
422 |
|
|
|
|
|
|
|
|
|
423 |
|
424 |
### Supported Tasks and Leaderboards
|
425 |
|
|
|
376 |
task_ids:
|
377 |
- machine-translation
|
378 |
paperswithcode_id: tatoeba
|
379 |
+
pretty_name: Tatoeba
|
380 |
---
|
381 |
+
# Dataset Card for Tatoeba
|
382 |
|
383 |
## Table of Contents
|
384 |
- [Dataset Description](#dataset-description)
|
|
|
414 |
|
415 |
### Dataset Summary
|
416 |
|
417 |
+
Tatoeba is a collection of sentences and translations.
|
418 |
|
419 |
To load a language pair which isn't part of the config, all you need to do is specify the language code as pairs.
|
420 |
You can find the valid pairs in Homepage section of Dataset Description: http://opus.nlpl.eu/Tatoeba.php
|
|
|
422 |
|
423 |
`dataset = load_dataset("tatoeba", lang1="en", lang2="he")`
|
424 |
|
425 |
+
The default date is v2021-07-22, but you can also change the date with
|
426 |
+
|
427 |
+
`dataset = load_dataset("tatoeba", lang1="en", lang2="he", date="v2020-11-09")`
|
428 |
+
|
429 |
|
430 |
### Supported Tasks and Leaderboards
|
431 |
|
dataset_infos.json
CHANGED
@@ -1 +1 @@
|
|
1 |
-
{"en-mr": {"description": "This is a collection of translated sentences from Tatoeba\n359 languages, 3,403 bitexts\ntotal number of files: 750\ntotal number of tokens: 65.54M\ntotal number of sentence fragments: 8.96M\n", "citation": "@InProceedings{TIEDEMANN12.463,\n author = {J\
|
|
|
1 |
+
{"en-mr": {"description": "This is a collection of translated sentences from Tatoeba\n359 languages, 3,403 bitexts\ntotal number of files: 750\ntotal number of tokens: 65.54M\ntotal number of sentence fragments: 8.96M\n", "citation": "@InProceedings{TIEDEMANN12.463,\n author = {J{\"o}rg}rg Tiedemann},\n title = {Parallel Data, Tools and Interfaces in OPUS},\n booktitle = {Proceedings of the Eight International Conference on Language Resources and Evaluation (LREC'12)},\n year = {2012},\n month = {may},\n date = {23-25},\n address = {Istanbul, Turkey},\n editor = {Nicoletta Calzolari (Conference Chair) and Khalid Choukri and Thierry Declerck and Mehmet Ugur Dogan and Bente Maegaard and Joseph Mariani and Jan Odijk and Stelios Piperidis},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {978-2-9517408-7-7},\n language = {english}\n }\n", "homepage": "http://opus.nlpl.eu/Tatoeba.php", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "translation": {"languages": ["en", "mr"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "tatoeba", "config_name": "en-mr", "version": {"version_str": "2021.7.22", "description": null, "major": 2021, "minor": 7, "patch": 22}, "splits": {"train": {"name": "train", "num_bytes": 6190484, "num_examples": 53462, "dataset_name": "tatoeba"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-Tatoeba/v2021-07-22/moses/en-mr.txt.zip": {"num_bytes": 1436200, "checksum": "1c8fd9b0cb767ea74f65f2dc82e64c0a2ca77056274ecebf24738ed26c7a13fd"}}, "download_size": 1436200, "post_processing_size": null, "dataset_size": 6190484, "size_in_bytes": 7626684}, "eo-nl": {"description": "This is a collection of translated sentences from Tatoeba\n359 languages, 3,403 bitexts\ntotal number of files: 750\ntotal number of tokens: 65.54M\ntotal number of sentence fragments: 8.96M\n", "citation": "@InProceedings{TIEDEMANN12.463,\n author = {J{\"o}rg}rg Tiedemann},\n title = {Parallel Data, Tools and Interfaces in OPUS},\n booktitle = {Proceedings of the Eight International Conference on Language Resources and Evaluation (LREC'12)},\n year = {2012},\n month = {may},\n date = {23-25},\n address = {Istanbul, Turkey},\n editor = {Nicoletta Calzolari (Conference Chair) and Khalid Choukri and Thierry Declerck and Mehmet Ugur Dogan and Bente Maegaard and Joseph Mariani and Jan Odijk and Stelios Piperidis},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {978-2-9517408-7-7},\n language = {english}\n }\n", "homepage": "http://opus.nlpl.eu/Tatoeba.php", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "translation": {"languages": ["eo", "nl"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "tatoeba", "config_name": "eo-nl", "version": {"version_str": "2021.7.22", "description": null, "major": 2021, "minor": 7, "patch": 22}, "splits": {"train": {"name": "train", "num_bytes": 8150048, "num_examples": 93650, "dataset_name": "tatoeba"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-Tatoeba/v2021-07-22/moses/eo-nl.txt.zip": {"num_bytes": 3020382, "checksum": "5a34a3dd219178d644a2ccb0aa2013e10288e593415ce102304d90797a12a01f"}}, "download_size": 3020382, "post_processing_size": null, "dataset_size": 8150048, "size_in_bytes": 11170430}, "es-pt": {"description": "This is a collection of translated sentences from Tatoeba\n359 languages, 3,403 bitexts\ntotal number of files: 750\ntotal number of tokens: 65.54M\ntotal number of sentence fragments: 8.96M\n", "citation": "@InProceedings{TIEDEMANN12.463,\n author = {J{\"o}rg}rg Tiedemann},\n title = {Parallel Data, Tools and Interfaces in OPUS},\n booktitle = {Proceedings of the Eight International Conference on Language Resources and Evaluation (LREC'12)},\n year = {2012},\n month = {may},\n date = {23-25},\n address = {Istanbul, Turkey},\n editor = {Nicoletta Calzolari (Conference Chair) and Khalid Choukri and Thierry Declerck and Mehmet Ugur Dogan and Bente Maegaard and Joseph Mariani and Jan Odijk and Stelios Piperidis},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {978-2-9517408-7-7},\n language = {english}\n }\n", "homepage": "http://opus.nlpl.eu/Tatoeba.php", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "translation": {"languages": ["es", "pt"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "tatoeba", "config_name": "es-pt", "version": {"version_str": "2021.7.22", "description": null, "major": 2021, "minor": 7, "patch": 22}, "splits": {"train": {"name": "train", "num_bytes": 6180464, "num_examples": 67782, "dataset_name": "tatoeba"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-Tatoeba/v2021-07-22/moses/es-pt.txt.zip": {"num_bytes": 2340361, "checksum": "6299c6aee1c469a0b7b4c41cde861d40b62d72e2bbc0d4d86ab17e77972de352"}}, "download_size": 2340361, "post_processing_size": null, "dataset_size": 6180464, "size_in_bytes": 8520825}, "fr-ru": {"description": "This is a collection of translated sentences from Tatoeba\n359 languages, 3,403 bitexts\ntotal number of files: 750\ntotal number of tokens: 65.54M\ntotal number of sentence fragments: 8.96M\n", "citation": "@InProceedings{TIEDEMANN12.463,\n author = {J{\"o}rg}rg Tiedemann},\n title = {Parallel Data, Tools and Interfaces in OPUS},\n booktitle = {Proceedings of the Eight International Conference on Language Resources and Evaluation (LREC'12)},\n year = {2012},\n month = {may},\n date = {23-25},\n address = {Istanbul, Turkey},\n editor = {Nicoletta Calzolari (Conference Chair) and Khalid Choukri and Thierry Declerck and Mehmet Ugur Dogan and Bente Maegaard and Joseph Mariani and Jan Odijk and Stelios Piperidis},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {978-2-9517408-7-7},\n language = {english}\n }\n", "homepage": "http://opus.nlpl.eu/Tatoeba.php", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "translation": {"languages": ["fr", "ru"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "tatoeba", "config_name": "fr-ru", "version": {"version_str": "2021.7.22", "description": null, "major": 2021, "minor": 7, "patch": 22}, "splits": {"train": {"name": "train", "num_bytes": 19775390, "num_examples": 195161, "dataset_name": "tatoeba"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-Tatoeba/v2021-07-22/moses/fr-ru.txt.zip": {"num_bytes": 5509784, "checksum": "9dac8d336158be9473be2fc0e7e6c771cdeeb9154133eb59cfbd5e310ee49ebc"}}, "download_size": 5509784, "post_processing_size": null, "dataset_size": 19775390, "size_in_bytes": 25285174}, "es-gl": {"description": "This is a collection of translated sentences from Tatoeba\n359 languages, 3,403 bitexts\ntotal number of files: 750\ntotal number of tokens: 65.54M\ntotal number of sentence fragments: 8.96M\n", "citation": "@InProceedings{TIEDEMANN12.463,\n author = {J{\"o}rg}rg Tiedemann},\n title = {Parallel Data, Tools and Interfaces in OPUS},\n booktitle = {Proceedings of the Eight International Conference on Language Resources and Evaluation (LREC'12)},\n year = {2012},\n month = {may},\n date = {23-25},\n address = {Istanbul, Turkey},\n editor = {Nicoletta Calzolari (Conference Chair) and Khalid Choukri and Thierry Declerck and Mehmet Ugur Dogan and Bente Maegaard and Joseph Mariani and Jan Odijk and Stelios Piperidis},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {978-2-9517408-7-7},\n language = {english}\n }\n", "homepage": "http://opus.nlpl.eu/Tatoeba.php", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "translation": {"languages": ["es", "gl"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "tatoeba", "config_name": "es-gl", "version": {"version_str": "2021.7.22", "description": null, "major": 2021, "minor": 7, "patch": 22}, "splits": {"train": {"name": "train", "num_bytes": 287683, "num_examples": 3135, "dataset_name": "tatoeba"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-Tatoeba/v2021-07-22/moses/es-gl.txt.zip": {"num_bytes": 128506, "checksum": "b2b6138c1a8bf940af428e5d8bdacf7d0f0ded5a2628438b9019595d8e9ecbf3"}}, "download_size": 128506, "post_processing_size": null, "dataset_size": 287683, "size_in_bytes": 416189}}
|
dummy/en-mr/{2020.11.9 → 2021.7.22}/dummy_data.zip
RENAMED
File without changes
|
dummy/eo-nl/{2020.11.9 → 2021.7.22}/dummy_data.zip
RENAMED
File without changes
|
dummy/es-gl/{2020.11.9 → 2021.7.22}/dummy_data.zip
RENAMED
File without changes
|
dummy/es-pt/{2020.11.9 → 2021.7.22}/dummy_data.zip
RENAMED
File without changes
|
dummy/fr-ru/{2020.11.9 → 2021.7.22}/dummy_data.zip
RENAMED
File without changes
|
tatoeba.py
CHANGED
@@ -43,9 +43,10 @@ _CITATION = """\
|
|
43 |
}
|
44 |
"""
|
45 |
|
46 |
-
_VERSION = "
|
|
|
47 |
_BASE_NAME = "Tatoeba.{}.{}"
|
48 |
-
_BASE_URL = "https://object.pouta.csc.fi/OPUS-Tatoeba/
|
49 |
|
50 |
# Please note that only few pairs are shown here. You can use config to generate data for all language pairs
|
51 |
_LANGUAGE_PAIRS = [
|
@@ -58,7 +59,7 @@ _LANGUAGE_PAIRS = [
|
|
58 |
|
59 |
|
60 |
class TatoebaConfig(datasets.BuilderConfig):
|
61 |
-
def __init__(self, *args, lang1=None, lang2=None, **kwargs):
|
62 |
super().__init__(
|
63 |
*args,
|
64 |
name=f"{lang1}-{lang2}",
|
@@ -66,6 +67,7 @@ class TatoebaConfig(datasets.BuilderConfig):
|
|
66 |
)
|
67 |
self.lang1 = lang1
|
68 |
self.lang2 = lang2
|
|
|
69 |
|
70 |
|
71 |
class Tatoeba(datasets.GeneratorBasedBuilder):
|
@@ -95,10 +97,10 @@ class Tatoeba(datasets.GeneratorBasedBuilder):
|
|
95 |
)
|
96 |
|
97 |
def _split_generators(self, dl_manager):
|
98 |
-
def _base_url(lang1, lang2):
|
99 |
-
return _BASE_URL.format(lang1, lang2)
|
100 |
|
101 |
-
download_url = _base_url(self.config.lang1, self.config.lang2)
|
102 |
path = dl_manager.download_and_extract(download_url)
|
103 |
return [
|
104 |
datasets.SplitGenerator(
|
|
|
43 |
}
|
44 |
"""
|
45 |
|
46 |
+
_VERSION = "2021.7.22"
|
47 |
+
_DATE = "v" + "-".join(s.zfill(2) for s in _VERSION.split("."))
|
48 |
_BASE_NAME = "Tatoeba.{}.{}"
|
49 |
+
_BASE_URL = "https://object.pouta.csc.fi/OPUS-Tatoeba/{}/moses/{}-{}.txt.zip"
|
50 |
|
51 |
# Please note that only few pairs are shown here. You can use config to generate data for all language pairs
|
52 |
_LANGUAGE_PAIRS = [
|
|
|
59 |
|
60 |
|
61 |
class TatoebaConfig(datasets.BuilderConfig):
|
62 |
+
def __init__(self, *args, lang1=None, lang2=None, date=_DATE, **kwargs):
|
63 |
super().__init__(
|
64 |
*args,
|
65 |
name=f"{lang1}-{lang2}",
|
|
|
67 |
)
|
68 |
self.lang1 = lang1
|
69 |
self.lang2 = lang2
|
70 |
+
self.date = date
|
71 |
|
72 |
|
73 |
class Tatoeba(datasets.GeneratorBasedBuilder):
|
|
|
97 |
)
|
98 |
|
99 |
def _split_generators(self, dl_manager):
|
100 |
+
def _base_url(lang1, lang2, date):
|
101 |
+
return _BASE_URL.format(date, lang1, lang2)
|
102 |
|
103 |
+
download_url = _base_url(self.config.lang1, self.config.lang2, self.config.date)
|
104 |
path = dl_manager.download_and_extract(download_url)
|
105 |
return [
|
106 |
datasets.SplitGenerator(
|