albertvillanova HF staff commited on
Commit
bc7e4fa
1 Parent(s): 1cdcf07

Convert dataset to Parquet

Browse files

Convert dataset to Parquet.

README.md CHANGED
@@ -18,6 +18,66 @@ language:
18
  paperswithcode_id: xnli
19
  pretty_name: Cross-lingual Natural Language Inference
20
  dataset_info:
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
21
  - config_name: ar
22
  features:
23
  - name: premise
@@ -33,16 +93,16 @@ dataset_info:
33
  '2': contradiction
34
  splits:
35
  - name: train
36
- num_bytes: 107399934
37
  num_examples: 392702
38
  - name: test
39
- num_bytes: 1294561
40
  num_examples: 5010
41
  - name: validation
42
- num_bytes: 633009
43
  num_examples: 2490
44
- download_size: 483963712
45
- dataset_size: 109327504
46
  - config_name: bg
47
  features:
48
  - name: premise
@@ -393,66 +453,15 @@ dataset_info:
393
  num_examples: 2490
394
  download_size: 483963712
395
  dataset_size: 73387957
396
- - config_name: all_languages
397
- features:
398
- - name: premise
399
- dtype:
400
- translation:
401
- languages:
402
- - ar
403
- - bg
404
- - de
405
- - el
406
- - en
407
- - es
408
- - fr
409
- - hi
410
- - ru
411
- - sw
412
- - th
413
- - tr
414
- - ur
415
- - vi
416
- - zh
417
- - name: hypothesis
418
- dtype:
419
- translation_variable_languages:
420
- languages:
421
- - ar
422
- - bg
423
- - de
424
- - el
425
- - en
426
- - es
427
- - fr
428
- - hi
429
- - ru
430
- - sw
431
- - th
432
- - tr
433
- - ur
434
- - vi
435
- - zh
436
- num_languages: 15
437
- - name: label
438
- dtype:
439
- class_label:
440
- names:
441
- '0': entailment
442
- '1': neutral
443
- '2': contradiction
444
- splits:
445
- - name: train
446
- num_bytes: 1581474731
447
- num_examples: 392702
448
- - name: test
449
- num_bytes: 19387508
450
- num_examples: 5010
451
- - name: validation
452
- num_bytes: 9566255
453
- num_examples: 2490
454
- download_size: 483963712
455
- dataset_size: 1610428494
456
  ---
457
 
458
  # Dataset Card for "xnli"
18
  paperswithcode_id: xnli
19
  pretty_name: Cross-lingual Natural Language Inference
20
  dataset_info:
21
+ - config_name: all_languages
22
+ features:
23
+ - name: premise
24
+ dtype:
25
+ translation:
26
+ languages:
27
+ - ar
28
+ - bg
29
+ - de
30
+ - el
31
+ - en
32
+ - es
33
+ - fr
34
+ - hi
35
+ - ru
36
+ - sw
37
+ - th
38
+ - tr
39
+ - ur
40
+ - vi
41
+ - zh
42
+ - name: hypothesis
43
+ dtype:
44
+ translation_variable_languages:
45
+ languages:
46
+ - ar
47
+ - bg
48
+ - de
49
+ - el
50
+ - en
51
+ - es
52
+ - fr
53
+ - hi
54
+ - ru
55
+ - sw
56
+ - th
57
+ - tr
58
+ - ur
59
+ - vi
60
+ - zh
61
+ num_languages: 15
62
+ - name: label
63
+ dtype:
64
+ class_label:
65
+ names:
66
+ '0': entailment
67
+ '1': neutral
68
+ '2': contradiction
69
+ splits:
70
+ - name: train
71
+ num_bytes: 1581474731
72
+ num_examples: 392702
73
+ - name: test
74
+ num_bytes: 19387508
75
+ num_examples: 5010
76
+ - name: validation
77
+ num_bytes: 9566255
78
+ num_examples: 2490
79
+ download_size: 483963712
80
+ dataset_size: 1610428494
81
  - config_name: ar
82
  features:
83
  - name: premise
93
  '2': contradiction
94
  splits:
95
  - name: train
96
+ num_bytes: 107399614
97
  num_examples: 392702
98
  - name: test
99
+ num_bytes: 1294553
100
  num_examples: 5010
101
  - name: validation
102
+ num_bytes: 633001
103
  num_examples: 2490
104
+ download_size: 59215902
105
+ dataset_size: 109327168
106
  - config_name: bg
107
  features:
108
  - name: premise
453
  num_examples: 2490
454
  download_size: 483963712
455
  dataset_size: 73387957
456
+ configs:
457
+ - config_name: ar
458
+ data_files:
459
+ - split: train
460
+ path: ar/train-*
461
+ - split: test
462
+ path: ar/test-*
463
+ - split: validation
464
+ path: ar/validation-*
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
465
  ---
466
 
467
  # Dataset Card for "xnli"
ar/test-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:89521231aa5f8404c46fcc5d421a9453819ca48bb99590680fa31fb8c82cf8bd
3
+ size 391980
ar/train-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5a27850bd0e20411f7c08e7c2247413c0050669090ef23cb5263d138be937e89
3
+ size 58630165
ar/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8df098db4682a6a97dc4a08be518b60e58112f0e32df2d4c4c933e34db577cd3
3
+ size 193757
dataset_infos.json CHANGED
@@ -1 +1,1202 @@
1
- {"ar": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "ar", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 107399934, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 1294561, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 633009, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 109327504, "size_in_bytes": 593291216}, "bg": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "bg", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 125973545, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 1573042, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 774069, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 128320656, "size_in_bytes": 612284368}, "de": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "de", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 84684460, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 996496, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 494612, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 86175568, "size_in_bytes": 570139280}, "el": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "el", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 139753678, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 1704793, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 841234, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 142299705, "size_in_bytes": 626263417}, "en": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "en", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 74444346, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 875142, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 433471, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 75752959, "size_in_bytes": 559716671}, "es": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "es", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 81383604, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 969821, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 478430, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 82831855, "size_in_bytes": 566795567}, "fr": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "fr", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 85809099, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 1029247, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 510112, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 87348458, "size_in_bytes": 571312170}, "hi": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "hi", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 170594284, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 2073081, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 1023923, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 173691288, "size_in_bytes": 657655000}, "ru": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "ru", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 129859935, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 1603474, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 786450, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 132249859, "size_in_bytes": 616213571}, "sw": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "sw", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 69286045, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 871659, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 429858, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 70587562, "size_in_bytes": 554551274}, "th": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "th", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 176063212, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 2147023, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 1061168, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 179271403, "size_in_bytes": 663235115}, "tr": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "tr", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 71637460, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 934942, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 459316, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 73031718, "size_in_bytes": 556995430}, "ur": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "ur", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 96441806, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 1416249, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 699960, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 98558015, "size_in_bytes": 582521727}, "vi": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "vi", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 101417750, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 1190225, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 590688, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 103198663, "size_in_bytes": 587162375}, "zh": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "zh", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 72225161, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 777937, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 384859, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 73387957, "size_in_bytes": 557351669}, "all_languages": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"premise": {"languages": ["ar", "bg", "de", "el", "en", "es", "fr", "hi", "ru", "sw", "th", "tr", "ur", "vi", "zh"], "id": null, "_type": "Translation"}, "hypothesis": {"languages": ["ar", "bg", "de", "el", "en", "es", "fr", "hi", "ru", "sw", "th", "tr", "ur", "vi", "zh"], "num_languages": 15, "id": null, "_type": "TranslationVariableLanguages"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "xnli", "config_name": "all_languages", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 1581474731, "num_examples": 392702, "dataset_name": "xnli"}, "test": {"name": "test", "num_bytes": 19387508, "num_examples": 5010, "dataset_name": "xnli"}, "validation": {"name": "validation", "num_bytes": 9566255, "num_examples": 2490, "dataset_name": "xnli"}}, "download_checksums": {"https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {"num_bytes": 466098360, "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"}, "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {"num_bytes": 17865352, "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"}}, "download_size": 483963712, "post_processing_size": null, "dataset_size": 1610428494, "size_in_bytes": 2094392206}}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "ar": {
3
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
4
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
5
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
6
+ "license": "",
7
+ "features": {
8
+ "premise": {
9
+ "dtype": "string",
10
+ "_type": "Value"
11
+ },
12
+ "hypothesis": {
13
+ "dtype": "string",
14
+ "_type": "Value"
15
+ },
16
+ "label": {
17
+ "names": [
18
+ "entailment",
19
+ "neutral",
20
+ "contradiction"
21
+ ],
22
+ "_type": "ClassLabel"
23
+ }
24
+ },
25
+ "builder_name": "xnli",
26
+ "dataset_name": "xnli",
27
+ "config_name": "ar",
28
+ "version": {
29
+ "version_str": "1.1.0",
30
+ "description": "",
31
+ "major": 1,
32
+ "minor": 1,
33
+ "patch": 0
34
+ },
35
+ "splits": {
36
+ "train": {
37
+ "name": "train",
38
+ "num_bytes": 107399614,
39
+ "num_examples": 392702,
40
+ "dataset_name": null
41
+ },
42
+ "test": {
43
+ "name": "test",
44
+ "num_bytes": 1294553,
45
+ "num_examples": 5010,
46
+ "dataset_name": null
47
+ },
48
+ "validation": {
49
+ "name": "validation",
50
+ "num_bytes": 633001,
51
+ "num_examples": 2490,
52
+ "dataset_name": null
53
+ }
54
+ },
55
+ "download_size": 59215902,
56
+ "dataset_size": 109327168,
57
+ "size_in_bytes": 168543070
58
+ },
59
+ "bg": {
60
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
61
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
62
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
63
+ "license": "",
64
+ "features": {
65
+ "premise": {
66
+ "dtype": "string",
67
+ "id": null,
68
+ "_type": "Value"
69
+ },
70
+ "hypothesis": {
71
+ "dtype": "string",
72
+ "id": null,
73
+ "_type": "Value"
74
+ },
75
+ "label": {
76
+ "num_classes": 3,
77
+ "names": [
78
+ "entailment",
79
+ "neutral",
80
+ "contradiction"
81
+ ],
82
+ "names_file": null,
83
+ "id": null,
84
+ "_type": "ClassLabel"
85
+ }
86
+ },
87
+ "post_processed": null,
88
+ "supervised_keys": null,
89
+ "builder_name": "xnli",
90
+ "config_name": "bg",
91
+ "version": {
92
+ "version_str": "1.1.0",
93
+ "description": "",
94
+ "major": 1,
95
+ "minor": 1,
96
+ "patch": 0
97
+ },
98
+ "splits": {
99
+ "train": {
100
+ "name": "train",
101
+ "num_bytes": 125973545,
102
+ "num_examples": 392702,
103
+ "dataset_name": "xnli"
104
+ },
105
+ "test": {
106
+ "name": "test",
107
+ "num_bytes": 1573042,
108
+ "num_examples": 5010,
109
+ "dataset_name": "xnli"
110
+ },
111
+ "validation": {
112
+ "name": "validation",
113
+ "num_bytes": 774069,
114
+ "num_examples": 2490,
115
+ "dataset_name": "xnli"
116
+ }
117
+ },
118
+ "download_checksums": {
119
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
120
+ "num_bytes": 466098360,
121
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
122
+ },
123
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
124
+ "num_bytes": 17865352,
125
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
126
+ }
127
+ },
128
+ "download_size": 483963712,
129
+ "post_processing_size": null,
130
+ "dataset_size": 128320656,
131
+ "size_in_bytes": 612284368
132
+ },
133
+ "de": {
134
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
135
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
136
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
137
+ "license": "",
138
+ "features": {
139
+ "premise": {
140
+ "dtype": "string",
141
+ "id": null,
142
+ "_type": "Value"
143
+ },
144
+ "hypothesis": {
145
+ "dtype": "string",
146
+ "id": null,
147
+ "_type": "Value"
148
+ },
149
+ "label": {
150
+ "num_classes": 3,
151
+ "names": [
152
+ "entailment",
153
+ "neutral",
154
+ "contradiction"
155
+ ],
156
+ "names_file": null,
157
+ "id": null,
158
+ "_type": "ClassLabel"
159
+ }
160
+ },
161
+ "post_processed": null,
162
+ "supervised_keys": null,
163
+ "builder_name": "xnli",
164
+ "config_name": "de",
165
+ "version": {
166
+ "version_str": "1.1.0",
167
+ "description": "",
168
+ "major": 1,
169
+ "minor": 1,
170
+ "patch": 0
171
+ },
172
+ "splits": {
173
+ "train": {
174
+ "name": "train",
175
+ "num_bytes": 84684460,
176
+ "num_examples": 392702,
177
+ "dataset_name": "xnli"
178
+ },
179
+ "test": {
180
+ "name": "test",
181
+ "num_bytes": 996496,
182
+ "num_examples": 5010,
183
+ "dataset_name": "xnli"
184
+ },
185
+ "validation": {
186
+ "name": "validation",
187
+ "num_bytes": 494612,
188
+ "num_examples": 2490,
189
+ "dataset_name": "xnli"
190
+ }
191
+ },
192
+ "download_checksums": {
193
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
194
+ "num_bytes": 466098360,
195
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
196
+ },
197
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
198
+ "num_bytes": 17865352,
199
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
200
+ }
201
+ },
202
+ "download_size": 483963712,
203
+ "post_processing_size": null,
204
+ "dataset_size": 86175568,
205
+ "size_in_bytes": 570139280
206
+ },
207
+ "el": {
208
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
209
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
210
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
211
+ "license": "",
212
+ "features": {
213
+ "premise": {
214
+ "dtype": "string",
215
+ "id": null,
216
+ "_type": "Value"
217
+ },
218
+ "hypothesis": {
219
+ "dtype": "string",
220
+ "id": null,
221
+ "_type": "Value"
222
+ },
223
+ "label": {
224
+ "num_classes": 3,
225
+ "names": [
226
+ "entailment",
227
+ "neutral",
228
+ "contradiction"
229
+ ],
230
+ "names_file": null,
231
+ "id": null,
232
+ "_type": "ClassLabel"
233
+ }
234
+ },
235
+ "post_processed": null,
236
+ "supervised_keys": null,
237
+ "builder_name": "xnli",
238
+ "config_name": "el",
239
+ "version": {
240
+ "version_str": "1.1.0",
241
+ "description": "",
242
+ "major": 1,
243
+ "minor": 1,
244
+ "patch": 0
245
+ },
246
+ "splits": {
247
+ "train": {
248
+ "name": "train",
249
+ "num_bytes": 139753678,
250
+ "num_examples": 392702,
251
+ "dataset_name": "xnli"
252
+ },
253
+ "test": {
254
+ "name": "test",
255
+ "num_bytes": 1704793,
256
+ "num_examples": 5010,
257
+ "dataset_name": "xnli"
258
+ },
259
+ "validation": {
260
+ "name": "validation",
261
+ "num_bytes": 841234,
262
+ "num_examples": 2490,
263
+ "dataset_name": "xnli"
264
+ }
265
+ },
266
+ "download_checksums": {
267
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
268
+ "num_bytes": 466098360,
269
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
270
+ },
271
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
272
+ "num_bytes": 17865352,
273
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
274
+ }
275
+ },
276
+ "download_size": 483963712,
277
+ "post_processing_size": null,
278
+ "dataset_size": 142299705,
279
+ "size_in_bytes": 626263417
280
+ },
281
+ "en": {
282
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
283
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
284
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
285
+ "license": "",
286
+ "features": {
287
+ "premise": {
288
+ "dtype": "string",
289
+ "id": null,
290
+ "_type": "Value"
291
+ },
292
+ "hypothesis": {
293
+ "dtype": "string",
294
+ "id": null,
295
+ "_type": "Value"
296
+ },
297
+ "label": {
298
+ "num_classes": 3,
299
+ "names": [
300
+ "entailment",
301
+ "neutral",
302
+ "contradiction"
303
+ ],
304
+ "names_file": null,
305
+ "id": null,
306
+ "_type": "ClassLabel"
307
+ }
308
+ },
309
+ "post_processed": null,
310
+ "supervised_keys": null,
311
+ "builder_name": "xnli",
312
+ "config_name": "en",
313
+ "version": {
314
+ "version_str": "1.1.0",
315
+ "description": "",
316
+ "major": 1,
317
+ "minor": 1,
318
+ "patch": 0
319
+ },
320
+ "splits": {
321
+ "train": {
322
+ "name": "train",
323
+ "num_bytes": 74444346,
324
+ "num_examples": 392702,
325
+ "dataset_name": "xnli"
326
+ },
327
+ "test": {
328
+ "name": "test",
329
+ "num_bytes": 875142,
330
+ "num_examples": 5010,
331
+ "dataset_name": "xnli"
332
+ },
333
+ "validation": {
334
+ "name": "validation",
335
+ "num_bytes": 433471,
336
+ "num_examples": 2490,
337
+ "dataset_name": "xnli"
338
+ }
339
+ },
340
+ "download_checksums": {
341
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
342
+ "num_bytes": 466098360,
343
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
344
+ },
345
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
346
+ "num_bytes": 17865352,
347
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
348
+ }
349
+ },
350
+ "download_size": 483963712,
351
+ "post_processing_size": null,
352
+ "dataset_size": 75752959,
353
+ "size_in_bytes": 559716671
354
+ },
355
+ "es": {
356
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
357
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
358
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
359
+ "license": "",
360
+ "features": {
361
+ "premise": {
362
+ "dtype": "string",
363
+ "id": null,
364
+ "_type": "Value"
365
+ },
366
+ "hypothesis": {
367
+ "dtype": "string",
368
+ "id": null,
369
+ "_type": "Value"
370
+ },
371
+ "label": {
372
+ "num_classes": 3,
373
+ "names": [
374
+ "entailment",
375
+ "neutral",
376
+ "contradiction"
377
+ ],
378
+ "names_file": null,
379
+ "id": null,
380
+ "_type": "ClassLabel"
381
+ }
382
+ },
383
+ "post_processed": null,
384
+ "supervised_keys": null,
385
+ "builder_name": "xnli",
386
+ "config_name": "es",
387
+ "version": {
388
+ "version_str": "1.1.0",
389
+ "description": "",
390
+ "major": 1,
391
+ "minor": 1,
392
+ "patch": 0
393
+ },
394
+ "splits": {
395
+ "train": {
396
+ "name": "train",
397
+ "num_bytes": 81383604,
398
+ "num_examples": 392702,
399
+ "dataset_name": "xnli"
400
+ },
401
+ "test": {
402
+ "name": "test",
403
+ "num_bytes": 969821,
404
+ "num_examples": 5010,
405
+ "dataset_name": "xnli"
406
+ },
407
+ "validation": {
408
+ "name": "validation",
409
+ "num_bytes": 478430,
410
+ "num_examples": 2490,
411
+ "dataset_name": "xnli"
412
+ }
413
+ },
414
+ "download_checksums": {
415
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
416
+ "num_bytes": 466098360,
417
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
418
+ },
419
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
420
+ "num_bytes": 17865352,
421
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
422
+ }
423
+ },
424
+ "download_size": 483963712,
425
+ "post_processing_size": null,
426
+ "dataset_size": 82831855,
427
+ "size_in_bytes": 566795567
428
+ },
429
+ "fr": {
430
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
431
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
432
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
433
+ "license": "",
434
+ "features": {
435
+ "premise": {
436
+ "dtype": "string",
437
+ "id": null,
438
+ "_type": "Value"
439
+ },
440
+ "hypothesis": {
441
+ "dtype": "string",
442
+ "id": null,
443
+ "_type": "Value"
444
+ },
445
+ "label": {
446
+ "num_classes": 3,
447
+ "names": [
448
+ "entailment",
449
+ "neutral",
450
+ "contradiction"
451
+ ],
452
+ "names_file": null,
453
+ "id": null,
454
+ "_type": "ClassLabel"
455
+ }
456
+ },
457
+ "post_processed": null,
458
+ "supervised_keys": null,
459
+ "builder_name": "xnli",
460
+ "config_name": "fr",
461
+ "version": {
462
+ "version_str": "1.1.0",
463
+ "description": "",
464
+ "major": 1,
465
+ "minor": 1,
466
+ "patch": 0
467
+ },
468
+ "splits": {
469
+ "train": {
470
+ "name": "train",
471
+ "num_bytes": 85809099,
472
+ "num_examples": 392702,
473
+ "dataset_name": "xnli"
474
+ },
475
+ "test": {
476
+ "name": "test",
477
+ "num_bytes": 1029247,
478
+ "num_examples": 5010,
479
+ "dataset_name": "xnli"
480
+ },
481
+ "validation": {
482
+ "name": "validation",
483
+ "num_bytes": 510112,
484
+ "num_examples": 2490,
485
+ "dataset_name": "xnli"
486
+ }
487
+ },
488
+ "download_checksums": {
489
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
490
+ "num_bytes": 466098360,
491
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
492
+ },
493
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
494
+ "num_bytes": 17865352,
495
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
496
+ }
497
+ },
498
+ "download_size": 483963712,
499
+ "post_processing_size": null,
500
+ "dataset_size": 87348458,
501
+ "size_in_bytes": 571312170
502
+ },
503
+ "hi": {
504
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
505
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
506
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
507
+ "license": "",
508
+ "features": {
509
+ "premise": {
510
+ "dtype": "string",
511
+ "id": null,
512
+ "_type": "Value"
513
+ },
514
+ "hypothesis": {
515
+ "dtype": "string",
516
+ "id": null,
517
+ "_type": "Value"
518
+ },
519
+ "label": {
520
+ "num_classes": 3,
521
+ "names": [
522
+ "entailment",
523
+ "neutral",
524
+ "contradiction"
525
+ ],
526
+ "names_file": null,
527
+ "id": null,
528
+ "_type": "ClassLabel"
529
+ }
530
+ },
531
+ "post_processed": null,
532
+ "supervised_keys": null,
533
+ "builder_name": "xnli",
534
+ "config_name": "hi",
535
+ "version": {
536
+ "version_str": "1.1.0",
537
+ "description": "",
538
+ "major": 1,
539
+ "minor": 1,
540
+ "patch": 0
541
+ },
542
+ "splits": {
543
+ "train": {
544
+ "name": "train",
545
+ "num_bytes": 170594284,
546
+ "num_examples": 392702,
547
+ "dataset_name": "xnli"
548
+ },
549
+ "test": {
550
+ "name": "test",
551
+ "num_bytes": 2073081,
552
+ "num_examples": 5010,
553
+ "dataset_name": "xnli"
554
+ },
555
+ "validation": {
556
+ "name": "validation",
557
+ "num_bytes": 1023923,
558
+ "num_examples": 2490,
559
+ "dataset_name": "xnli"
560
+ }
561
+ },
562
+ "download_checksums": {
563
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
564
+ "num_bytes": 466098360,
565
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
566
+ },
567
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
568
+ "num_bytes": 17865352,
569
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
570
+ }
571
+ },
572
+ "download_size": 483963712,
573
+ "post_processing_size": null,
574
+ "dataset_size": 173691288,
575
+ "size_in_bytes": 657655000
576
+ },
577
+ "ru": {
578
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
579
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
580
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
581
+ "license": "",
582
+ "features": {
583
+ "premise": {
584
+ "dtype": "string",
585
+ "id": null,
586
+ "_type": "Value"
587
+ },
588
+ "hypothesis": {
589
+ "dtype": "string",
590
+ "id": null,
591
+ "_type": "Value"
592
+ },
593
+ "label": {
594
+ "num_classes": 3,
595
+ "names": [
596
+ "entailment",
597
+ "neutral",
598
+ "contradiction"
599
+ ],
600
+ "names_file": null,
601
+ "id": null,
602
+ "_type": "ClassLabel"
603
+ }
604
+ },
605
+ "post_processed": null,
606
+ "supervised_keys": null,
607
+ "builder_name": "xnli",
608
+ "config_name": "ru",
609
+ "version": {
610
+ "version_str": "1.1.0",
611
+ "description": "",
612
+ "major": 1,
613
+ "minor": 1,
614
+ "patch": 0
615
+ },
616
+ "splits": {
617
+ "train": {
618
+ "name": "train",
619
+ "num_bytes": 129859935,
620
+ "num_examples": 392702,
621
+ "dataset_name": "xnli"
622
+ },
623
+ "test": {
624
+ "name": "test",
625
+ "num_bytes": 1603474,
626
+ "num_examples": 5010,
627
+ "dataset_name": "xnli"
628
+ },
629
+ "validation": {
630
+ "name": "validation",
631
+ "num_bytes": 786450,
632
+ "num_examples": 2490,
633
+ "dataset_name": "xnli"
634
+ }
635
+ },
636
+ "download_checksums": {
637
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
638
+ "num_bytes": 466098360,
639
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
640
+ },
641
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
642
+ "num_bytes": 17865352,
643
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
644
+ }
645
+ },
646
+ "download_size": 483963712,
647
+ "post_processing_size": null,
648
+ "dataset_size": 132249859,
649
+ "size_in_bytes": 616213571
650
+ },
651
+ "sw": {
652
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
653
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
654
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
655
+ "license": "",
656
+ "features": {
657
+ "premise": {
658
+ "dtype": "string",
659
+ "id": null,
660
+ "_type": "Value"
661
+ },
662
+ "hypothesis": {
663
+ "dtype": "string",
664
+ "id": null,
665
+ "_type": "Value"
666
+ },
667
+ "label": {
668
+ "num_classes": 3,
669
+ "names": [
670
+ "entailment",
671
+ "neutral",
672
+ "contradiction"
673
+ ],
674
+ "names_file": null,
675
+ "id": null,
676
+ "_type": "ClassLabel"
677
+ }
678
+ },
679
+ "post_processed": null,
680
+ "supervised_keys": null,
681
+ "builder_name": "xnli",
682
+ "config_name": "sw",
683
+ "version": {
684
+ "version_str": "1.1.0",
685
+ "description": "",
686
+ "major": 1,
687
+ "minor": 1,
688
+ "patch": 0
689
+ },
690
+ "splits": {
691
+ "train": {
692
+ "name": "train",
693
+ "num_bytes": 69286045,
694
+ "num_examples": 392702,
695
+ "dataset_name": "xnli"
696
+ },
697
+ "test": {
698
+ "name": "test",
699
+ "num_bytes": 871659,
700
+ "num_examples": 5010,
701
+ "dataset_name": "xnli"
702
+ },
703
+ "validation": {
704
+ "name": "validation",
705
+ "num_bytes": 429858,
706
+ "num_examples": 2490,
707
+ "dataset_name": "xnli"
708
+ }
709
+ },
710
+ "download_checksums": {
711
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
712
+ "num_bytes": 466098360,
713
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
714
+ },
715
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
716
+ "num_bytes": 17865352,
717
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
718
+ }
719
+ },
720
+ "download_size": 483963712,
721
+ "post_processing_size": null,
722
+ "dataset_size": 70587562,
723
+ "size_in_bytes": 554551274
724
+ },
725
+ "th": {
726
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
727
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
728
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
729
+ "license": "",
730
+ "features": {
731
+ "premise": {
732
+ "dtype": "string",
733
+ "id": null,
734
+ "_type": "Value"
735
+ },
736
+ "hypothesis": {
737
+ "dtype": "string",
738
+ "id": null,
739
+ "_type": "Value"
740
+ },
741
+ "label": {
742
+ "num_classes": 3,
743
+ "names": [
744
+ "entailment",
745
+ "neutral",
746
+ "contradiction"
747
+ ],
748
+ "names_file": null,
749
+ "id": null,
750
+ "_type": "ClassLabel"
751
+ }
752
+ },
753
+ "post_processed": null,
754
+ "supervised_keys": null,
755
+ "builder_name": "xnli",
756
+ "config_name": "th",
757
+ "version": {
758
+ "version_str": "1.1.0",
759
+ "description": "",
760
+ "major": 1,
761
+ "minor": 1,
762
+ "patch": 0
763
+ },
764
+ "splits": {
765
+ "train": {
766
+ "name": "train",
767
+ "num_bytes": 176063212,
768
+ "num_examples": 392702,
769
+ "dataset_name": "xnli"
770
+ },
771
+ "test": {
772
+ "name": "test",
773
+ "num_bytes": 2147023,
774
+ "num_examples": 5010,
775
+ "dataset_name": "xnli"
776
+ },
777
+ "validation": {
778
+ "name": "validation",
779
+ "num_bytes": 1061168,
780
+ "num_examples": 2490,
781
+ "dataset_name": "xnli"
782
+ }
783
+ },
784
+ "download_checksums": {
785
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
786
+ "num_bytes": 466098360,
787
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
788
+ },
789
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
790
+ "num_bytes": 17865352,
791
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
792
+ }
793
+ },
794
+ "download_size": 483963712,
795
+ "post_processing_size": null,
796
+ "dataset_size": 179271403,
797
+ "size_in_bytes": 663235115
798
+ },
799
+ "tr": {
800
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
801
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
802
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
803
+ "license": "",
804
+ "features": {
805
+ "premise": {
806
+ "dtype": "string",
807
+ "id": null,
808
+ "_type": "Value"
809
+ },
810
+ "hypothesis": {
811
+ "dtype": "string",
812
+ "id": null,
813
+ "_type": "Value"
814
+ },
815
+ "label": {
816
+ "num_classes": 3,
817
+ "names": [
818
+ "entailment",
819
+ "neutral",
820
+ "contradiction"
821
+ ],
822
+ "names_file": null,
823
+ "id": null,
824
+ "_type": "ClassLabel"
825
+ }
826
+ },
827
+ "post_processed": null,
828
+ "supervised_keys": null,
829
+ "builder_name": "xnli",
830
+ "config_name": "tr",
831
+ "version": {
832
+ "version_str": "1.1.0",
833
+ "description": "",
834
+ "major": 1,
835
+ "minor": 1,
836
+ "patch": 0
837
+ },
838
+ "splits": {
839
+ "train": {
840
+ "name": "train",
841
+ "num_bytes": 71637460,
842
+ "num_examples": 392702,
843
+ "dataset_name": "xnli"
844
+ },
845
+ "test": {
846
+ "name": "test",
847
+ "num_bytes": 934942,
848
+ "num_examples": 5010,
849
+ "dataset_name": "xnli"
850
+ },
851
+ "validation": {
852
+ "name": "validation",
853
+ "num_bytes": 459316,
854
+ "num_examples": 2490,
855
+ "dataset_name": "xnli"
856
+ }
857
+ },
858
+ "download_checksums": {
859
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
860
+ "num_bytes": 466098360,
861
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
862
+ },
863
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
864
+ "num_bytes": 17865352,
865
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
866
+ }
867
+ },
868
+ "download_size": 483963712,
869
+ "post_processing_size": null,
870
+ "dataset_size": 73031718,
871
+ "size_in_bytes": 556995430
872
+ },
873
+ "ur": {
874
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
875
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
876
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
877
+ "license": "",
878
+ "features": {
879
+ "premise": {
880
+ "dtype": "string",
881
+ "id": null,
882
+ "_type": "Value"
883
+ },
884
+ "hypothesis": {
885
+ "dtype": "string",
886
+ "id": null,
887
+ "_type": "Value"
888
+ },
889
+ "label": {
890
+ "num_classes": 3,
891
+ "names": [
892
+ "entailment",
893
+ "neutral",
894
+ "contradiction"
895
+ ],
896
+ "names_file": null,
897
+ "id": null,
898
+ "_type": "ClassLabel"
899
+ }
900
+ },
901
+ "post_processed": null,
902
+ "supervised_keys": null,
903
+ "builder_name": "xnli",
904
+ "config_name": "ur",
905
+ "version": {
906
+ "version_str": "1.1.0",
907
+ "description": "",
908
+ "major": 1,
909
+ "minor": 1,
910
+ "patch": 0
911
+ },
912
+ "splits": {
913
+ "train": {
914
+ "name": "train",
915
+ "num_bytes": 96441806,
916
+ "num_examples": 392702,
917
+ "dataset_name": "xnli"
918
+ },
919
+ "test": {
920
+ "name": "test",
921
+ "num_bytes": 1416249,
922
+ "num_examples": 5010,
923
+ "dataset_name": "xnli"
924
+ },
925
+ "validation": {
926
+ "name": "validation",
927
+ "num_bytes": 699960,
928
+ "num_examples": 2490,
929
+ "dataset_name": "xnli"
930
+ }
931
+ },
932
+ "download_checksums": {
933
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
934
+ "num_bytes": 466098360,
935
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
936
+ },
937
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
938
+ "num_bytes": 17865352,
939
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
940
+ }
941
+ },
942
+ "download_size": 483963712,
943
+ "post_processing_size": null,
944
+ "dataset_size": 98558015,
945
+ "size_in_bytes": 582521727
946
+ },
947
+ "vi": {
948
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
949
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
950
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
951
+ "license": "",
952
+ "features": {
953
+ "premise": {
954
+ "dtype": "string",
955
+ "id": null,
956
+ "_type": "Value"
957
+ },
958
+ "hypothesis": {
959
+ "dtype": "string",
960
+ "id": null,
961
+ "_type": "Value"
962
+ },
963
+ "label": {
964
+ "num_classes": 3,
965
+ "names": [
966
+ "entailment",
967
+ "neutral",
968
+ "contradiction"
969
+ ],
970
+ "names_file": null,
971
+ "id": null,
972
+ "_type": "ClassLabel"
973
+ }
974
+ },
975
+ "post_processed": null,
976
+ "supervised_keys": null,
977
+ "builder_name": "xnli",
978
+ "config_name": "vi",
979
+ "version": {
980
+ "version_str": "1.1.0",
981
+ "description": "",
982
+ "major": 1,
983
+ "minor": 1,
984
+ "patch": 0
985
+ },
986
+ "splits": {
987
+ "train": {
988
+ "name": "train",
989
+ "num_bytes": 101417750,
990
+ "num_examples": 392702,
991
+ "dataset_name": "xnli"
992
+ },
993
+ "test": {
994
+ "name": "test",
995
+ "num_bytes": 1190225,
996
+ "num_examples": 5010,
997
+ "dataset_name": "xnli"
998
+ },
999
+ "validation": {
1000
+ "name": "validation",
1001
+ "num_bytes": 590688,
1002
+ "num_examples": 2490,
1003
+ "dataset_name": "xnli"
1004
+ }
1005
+ },
1006
+ "download_checksums": {
1007
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
1008
+ "num_bytes": 466098360,
1009
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
1010
+ },
1011
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
1012
+ "num_bytes": 17865352,
1013
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
1014
+ }
1015
+ },
1016
+ "download_size": 483963712,
1017
+ "post_processing_size": null,
1018
+ "dataset_size": 103198663,
1019
+ "size_in_bytes": 587162375
1020
+ },
1021
+ "zh": {
1022
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
1023
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
1024
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
1025
+ "license": "",
1026
+ "features": {
1027
+ "premise": {
1028
+ "dtype": "string",
1029
+ "id": null,
1030
+ "_type": "Value"
1031
+ },
1032
+ "hypothesis": {
1033
+ "dtype": "string",
1034
+ "id": null,
1035
+ "_type": "Value"
1036
+ },
1037
+ "label": {
1038
+ "num_classes": 3,
1039
+ "names": [
1040
+ "entailment",
1041
+ "neutral",
1042
+ "contradiction"
1043
+ ],
1044
+ "names_file": null,
1045
+ "id": null,
1046
+ "_type": "ClassLabel"
1047
+ }
1048
+ },
1049
+ "post_processed": null,
1050
+ "supervised_keys": null,
1051
+ "builder_name": "xnli",
1052
+ "config_name": "zh",
1053
+ "version": {
1054
+ "version_str": "1.1.0",
1055
+ "description": "",
1056
+ "major": 1,
1057
+ "minor": 1,
1058
+ "patch": 0
1059
+ },
1060
+ "splits": {
1061
+ "train": {
1062
+ "name": "train",
1063
+ "num_bytes": 72225161,
1064
+ "num_examples": 392702,
1065
+ "dataset_name": "xnli"
1066
+ },
1067
+ "test": {
1068
+ "name": "test",
1069
+ "num_bytes": 777937,
1070
+ "num_examples": 5010,
1071
+ "dataset_name": "xnli"
1072
+ },
1073
+ "validation": {
1074
+ "name": "validation",
1075
+ "num_bytes": 384859,
1076
+ "num_examples": 2490,
1077
+ "dataset_name": "xnli"
1078
+ }
1079
+ },
1080
+ "download_checksums": {
1081
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
1082
+ "num_bytes": 466098360,
1083
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
1084
+ },
1085
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
1086
+ "num_bytes": 17865352,
1087
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
1088
+ }
1089
+ },
1090
+ "download_size": 483963712,
1091
+ "post_processing_size": null,
1092
+ "dataset_size": 73387957,
1093
+ "size_in_bytes": 557351669
1094
+ },
1095
+ "all_languages": {
1096
+ "description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n",
1097
+ "citation": "@InProceedings{conneau2018xnli,\n author = {Conneau, Alexis\n and Rinott, Ruty\n and Lample, Guillaume\n and Williams, Adina\n and Bowman, Samuel R.\n and Schwenk, Holger\n and Stoyanov, Veselin},\n title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n in Natural Language Processing},\n year = {2018},\n publisher = {Association for Computational Linguistics},\n location = {Brussels, Belgium},\n}",
1098
+ "homepage": "https://www.nyu.edu/projects/bowman/xnli/",
1099
+ "license": "",
1100
+ "features": {
1101
+ "premise": {
1102
+ "languages": [
1103
+ "ar",
1104
+ "bg",
1105
+ "de",
1106
+ "el",
1107
+ "en",
1108
+ "es",
1109
+ "fr",
1110
+ "hi",
1111
+ "ru",
1112
+ "sw",
1113
+ "th",
1114
+ "tr",
1115
+ "ur",
1116
+ "vi",
1117
+ "zh"
1118
+ ],
1119
+ "id": null,
1120
+ "_type": "Translation"
1121
+ },
1122
+ "hypothesis": {
1123
+ "languages": [
1124
+ "ar",
1125
+ "bg",
1126
+ "de",
1127
+ "el",
1128
+ "en",
1129
+ "es",
1130
+ "fr",
1131
+ "hi",
1132
+ "ru",
1133
+ "sw",
1134
+ "th",
1135
+ "tr",
1136
+ "ur",
1137
+ "vi",
1138
+ "zh"
1139
+ ],
1140
+ "num_languages": 15,
1141
+ "id": null,
1142
+ "_type": "TranslationVariableLanguages"
1143
+ },
1144
+ "label": {
1145
+ "num_classes": 3,
1146
+ "names": [
1147
+ "entailment",
1148
+ "neutral",
1149
+ "contradiction"
1150
+ ],
1151
+ "names_file": null,
1152
+ "id": null,
1153
+ "_type": "ClassLabel"
1154
+ }
1155
+ },
1156
+ "post_processed": null,
1157
+ "supervised_keys": null,
1158
+ "builder_name": "xnli",
1159
+ "config_name": "all_languages",
1160
+ "version": {
1161
+ "version_str": "1.1.0",
1162
+ "description": "",
1163
+ "major": 1,
1164
+ "minor": 1,
1165
+ "patch": 0
1166
+ },
1167
+ "splits": {
1168
+ "train": {
1169
+ "name": "train",
1170
+ "num_bytes": 1581474731,
1171
+ "num_examples": 392702,
1172
+ "dataset_name": "xnli"
1173
+ },
1174
+ "test": {
1175
+ "name": "test",
1176
+ "num_bytes": 19387508,
1177
+ "num_examples": 5010,
1178
+ "dataset_name": "xnli"
1179
+ },
1180
+ "validation": {
1181
+ "name": "validation",
1182
+ "num_bytes": 9566255,
1183
+ "num_examples": 2490,
1184
+ "dataset_name": "xnli"
1185
+ }
1186
+ },
1187
+ "download_checksums": {
1188
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-MT-1.0.zip": {
1189
+ "num_bytes": 466098360,
1190
+ "checksum": "f732517ba2fb1d550e9f3c2aabaef6017c91ee2dcec90e878f138764d224db05"
1191
+ },
1192
+ "https://dl.fbaipublicfiles.com/XNLI/XNLI-1.0.zip": {
1193
+ "num_bytes": 17865352,
1194
+ "checksum": "4ba1d5e1afdb7161f0f23c66dc787802ccfa8a25a3ddd3b165a35e50df346ab1"
1195
+ }
1196
+ },
1197
+ "download_size": 483963712,
1198
+ "post_processing_size": null,
1199
+ "dataset_size": 1610428494,
1200
+ "size_in_bytes": 2094392206
1201
+ }
1202
+ }