Hugging Face's logo

Datasets:
nala-cub
/
americas_nli

Tasks:
Text Classification
Modalities:
Text
Formats:
parquet
Sub-tasks:
natural-language-inference
Languages:
Aymara
Bribri
Asháninka
Size:
10K - 100K
ArXiv:
Libraries:
Datasets
pandas
License:
Dataset card Viewer Files Community
5
albertvillanova HF staff commited on
Commit
0b4bf17
•
1 Parent(s): 24e96ba

Delete legacy dataset_infos.json

Browse files
Files changed (1) hide show
  1. dataset_infos.json +0 -567
dataset_infos.json DELETED
@@ -1,567 +0,0 @@
1
- {
2
- "aym": {
3
- "description": "AmericasNLI is an extension of XNLI (Conneau et al., 2018) \u2013 a natural language inference (NLI) dataset covering 15 high-resource languages \u2013 to 10 low-resource indigenous languages spoken in the Americas: Ashaninka, Aymara, Bribri, Guarani, Nahuatl, Otomi, Quechua, Raramuri, Shipibo-Konibo, and Wixarika. As with MNLI, the goal is to predict textual entailment (does sentence A imply/contradict/neither sentence B) and is a classification task (given two sentences, predict one of three labels).\n",
4
- "citation": "\n@article{DBLP:journals/corr/abs-2104-08726,\n author = {Abteen Ebrahimi and\n Manuel Mager and\n Arturo Oncevay and\n Vishrav Chaudhary and\n Luis Chiruzzo and\n Angela Fan and\n John Ortega and\n Ricardo Ramos and\n Annette Rios and\n Ivan Vladimir and\n Gustavo A. Gim{'{e}}nez{-}Lugo and\n Elisabeth Mager and\n Graham Neubig and\n Alexis Palmer and\n Rolando A. Coto Solano and\n Ngoc Thang Vu and\n Katharina Kann},\n title = {AmericasNLI: Evaluating Zero-shot Natural Language Understanding of\n Pretrained Multilingual Models in Truly Low-resource Languages},\n journal = {CoRR},\n volume = {abs/2104.08726},\n year = {2021},\n url = {https://arxiv.org/abs/2104.08726},\n eprinttype = {arXiv},\n eprint = {2104.08726},\n timestamp = {Mon, 26 Apr 2021 17:25:10 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-2104-08726.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
5
- "homepage": "https://github.com/nala-cub/AmericasNLI",
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": "americas_nli",
26
- "dataset_name": "americas_nli",
27
- "config_name": "aym",
28
- "version": {
29
- "version_str": "1.0.0",
30
- "description": "",
31
- "major": 1,
32
- "minor": 0,
33
- "patch": 0
34
- },
35
- "splits": {
36
- "validation": {
37
- "name": "validation",
38
- "num_bytes": 117530,
39
- "num_examples": 743,
40
- "dataset_name": null
41
- },
42
- "test": {
43
- "name": "test",
44
- "num_bytes": 115251,
45
- "num_examples": 750,
46
- "dataset_name": null
47
- }
48
- },
49
- "download_size": 87882,
50
- "dataset_size": 232781,
51
- "size_in_bytes": 320663
52
- },
53
- "bzd": {
54
- "description": "AmericasNLI is an extension of XNLI (Conneau et al., 2018) \u2013 a natural language inference (NLI) dataset covering 15 high-resource languages \u2013 to 10 low-resource indigenous languages spoken in the Americas: Ashaninka, Aymara, Bribri, Guarani, Nahuatl, Otomi, Quechua, Raramuri, Shipibo-Konibo, and Wixarika. As with MNLI, the goal is to predict textual entailment (does sentence A imply/contradict/neither sentence B) and is a classification task (given two sentences, predict one of three labels).\n",
55
- "citation": "\n@article{DBLP:journals/corr/abs-2104-08726,\n author = {Abteen Ebrahimi and\n Manuel Mager and\n Arturo Oncevay and\n Vishrav Chaudhary and\n Luis Chiruzzo and\n Angela Fan and\n John Ortega and\n Ricardo Ramos and\n Annette Rios and\n Ivan Vladimir and\n Gustavo A. Gim{'{e}}nez{-}Lugo and\n Elisabeth Mager and\n Graham Neubig and\n Alexis Palmer and\n Rolando A. Coto Solano and\n Ngoc Thang Vu and\n Katharina Kann},\n title = {AmericasNLI: Evaluating Zero-shot Natural Language Understanding of\n Pretrained Multilingual Models in Truly Low-resource Languages},\n journal = {CoRR},\n volume = {abs/2104.08726},\n year = {2021},\n url = {https://arxiv.org/abs/2104.08726},\n eprinttype = {arXiv},\n eprint = {2104.08726},\n timestamp = {Mon, 26 Apr 2021 17:25:10 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-2104-08726.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
56
- "homepage": "https://github.com/nala-cub/AmericasNLI",
57
- "license": "",
58
- "features": {
59
- "premise": {
60
- "dtype": "string",
61
- "_type": "Value"
62
- },
63
- "hypothesis": {
64
- "dtype": "string",
65
- "_type": "Value"
66
- },
67
- "label": {
68
- "names": [
69
- "entailment",
70
- "neutral",
71
- "contradiction"
72
- ],
73
- "_type": "ClassLabel"
74
- }
75
- },
76
- "builder_name": "americas_nli",
77
- "dataset_name": "americas_nli",
78
- "config_name": "bzd",
79
- "version": {
80
- "version_str": "1.0.0",
81
- "description": "",
82
- "major": 1,
83
- "minor": 0,
84
- "patch": 0
85
- },
86
- "splits": {
87
- "validation": {
88
- "name": "validation",
89
- "num_bytes": 143354,
90
- "num_examples": 743,
91
- "dataset_name": null
92
- },
93
- "test": {
94
- "name": "test",
95
- "num_bytes": 127676,
96
- "num_examples": 750,
97
- "dataset_name": null
98
- }
99
- },
100
- "download_size": 91039,
101
- "dataset_size": 271030,
102
- "size_in_bytes": 362069
103
- },
104
- "cni": {
105
- "description": "AmericasNLI is an extension of XNLI (Conneau et al., 2018) \u2013 a natural language inference (NLI) dataset covering 15 high-resource languages \u2013 to 10 low-resource indigenous languages spoken in the Americas: Ashaninka, Aymara, Bribri, Guarani, Nahuatl, Otomi, Quechua, Raramuri, Shipibo-Konibo, and Wixarika. As with MNLI, the goal is to predict textual entailment (does sentence A imply/contradict/neither sentence B) and is a classification task (given two sentences, predict one of three labels).\n",
106
- "citation": "\n@article{DBLP:journals/corr/abs-2104-08726,\n author = {Abteen Ebrahimi and\n Manuel Mager and\n Arturo Oncevay and\n Vishrav Chaudhary and\n Luis Chiruzzo and\n Angela Fan and\n John Ortega and\n Ricardo Ramos and\n Annette Rios and\n Ivan Vladimir and\n Gustavo A. Gim{'{e}}nez{-}Lugo and\n Elisabeth Mager and\n Graham Neubig and\n Alexis Palmer and\n Rolando A. Coto Solano and\n Ngoc Thang Vu and\n Katharina Kann},\n title = {AmericasNLI: Evaluating Zero-shot Natural Language Understanding of\n Pretrained Multilingual Models in Truly Low-resource Languages},\n journal = {CoRR},\n volume = {abs/2104.08726},\n year = {2021},\n url = {https://arxiv.org/abs/2104.08726},\n eprinttype = {arXiv},\n eprint = {2104.08726},\n timestamp = {Mon, 26 Apr 2021 17:25:10 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-2104-08726.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
107
- "homepage": "https://github.com/nala-cub/AmericasNLI",
108
- "license": "",
109
- "features": {
110
- "premise": {
111
- "dtype": "string",
112
- "_type": "Value"
113
- },
114
- "hypothesis": {
115
- "dtype": "string",
116
- "_type": "Value"
117
- },
118
- "label": {
119
- "names": [
120
- "entailment",
121
- "neutral",
122
- "contradiction"
123
- ],
124
- "_type": "ClassLabel"
125
- }
126
- },
127
- "builder_name": "americas_nli",
128
- "dataset_name": "americas_nli",
129
- "config_name": "cni",
130
- "version": {
131
- "version_str": "1.0.0",
132
- "description": "",
133
- "major": 1,
134
- "minor": 0,
135
- "patch": 0
136
- },
137
- "splits": {
138
- "validation": {
139
- "name": "validation",
140
- "num_bytes": 113256,
141
- "num_examples": 658,
142
- "dataset_name": null
143
- },
144
- "test": {
145
- "name": "test",
146
- "num_bytes": 116284,
147
- "num_examples": 750,
148
- "dataset_name": null
149
- }
150
- },
151
- "download_size": 78899,
152
- "dataset_size": 229540,
153
- "size_in_bytes": 308439
154
- },
155
- "gn": {
156
- "description": "AmericasNLI is an extension of XNLI (Conneau et al., 2018) \u2013 a natural language inference (NLI) dataset covering 15 high-resource languages \u2013 to 10 low-resource indigenous languages spoken in the Americas: Ashaninka, Aymara, Bribri, Guarani, Nahuatl, Otomi, Quechua, Raramuri, Shipibo-Konibo, and Wixarika. As with MNLI, the goal is to predict textual entailment (does sentence A imply/contradict/neither sentence B) and is a classification task (given two sentences, predict one of three labels).\n",
157
- "citation": "\n@article{DBLP:journals/corr/abs-2104-08726,\n author = {Abteen Ebrahimi and\n Manuel Mager and\n Arturo Oncevay and\n Vishrav Chaudhary and\n Luis Chiruzzo and\n Angela Fan and\n John Ortega and\n Ricardo Ramos and\n Annette Rios and\n Ivan Vladimir and\n Gustavo A. Gim{'{e}}nez{-}Lugo and\n Elisabeth Mager and\n Graham Neubig and\n Alexis Palmer and\n Rolando A. Coto Solano and\n Ngoc Thang Vu and\n Katharina Kann},\n title = {AmericasNLI: Evaluating Zero-shot Natural Language Understanding of\n Pretrained Multilingual Models in Truly Low-resource Languages},\n journal = {CoRR},\n volume = {abs/2104.08726},\n year = {2021},\n url = {https://arxiv.org/abs/2104.08726},\n eprinttype = {arXiv},\n eprint = {2104.08726},\n timestamp = {Mon, 26 Apr 2021 17:25:10 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-2104-08726.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
158
- "homepage": "https://github.com/nala-cub/AmericasNLI",
159
- "license": "",
160
- "features": {
161
- "premise": {
162
- "dtype": "string",
163
- "_type": "Value"
164
- },
165
- "hypothesis": {
166
- "dtype": "string",
167
- "_type": "Value"
168
- },
169
- "label": {
170
- "names": [
171
- "entailment",
172
- "neutral",
173
- "contradiction"
174
- ],
175
- "_type": "ClassLabel"
176
- }
177
- },
178
- "builder_name": "americas_nli",
179
- "dataset_name": "americas_nli",
180
- "config_name": "gn",
181
- "version": {
182
- "version_str": "1.0.0",
183
- "description": "",
184
- "major": 1,
185
- "minor": 0,
186
- "patch": 0
187
- },
188
- "splits": {
189
- "validation": {
190
- "name": "validation",
191
- "num_bytes": 115135,
192
- "num_examples": 743,
193
- "dataset_name": null
194
- },
195
- "test": {
196
- "name": "test",
197
- "num_bytes": 101948,
198
- "num_examples": 750,
199
- "dataset_name": null
200
- }
201
- },
202
- "download_size": 80429,
203
- "dataset_size": 217083,
204
- "size_in_bytes": 297512
205
- },
206
- "hch": {
207
- "description": "AmericasNLI is an extension of XNLI (Conneau et al., 2018) \u2013 a natural language inference (NLI) dataset covering 15 high-resource languages \u2013 to 10 low-resource indigenous languages spoken in the Americas: Ashaninka, Aymara, Bribri, Guarani, Nahuatl, Otomi, Quechua, Raramuri, Shipibo-Konibo, and Wixarika. As with MNLI, the goal is to predict textual entailment (does sentence A imply/contradict/neither sentence B) and is a classification task (given two sentences, predict one of three labels).\n",
208
- "citation": "\n@article{DBLP:journals/corr/abs-2104-08726,\n author = {Abteen Ebrahimi and\n Manuel Mager and\n Arturo Oncevay and\n Vishrav Chaudhary and\n Luis Chiruzzo and\n Angela Fan and\n John Ortega and\n Ricardo Ramos and\n Annette Rios and\n Ivan Vladimir and\n Gustavo A. Gim{'{e}}nez{-}Lugo and\n Elisabeth Mager and\n Graham Neubig and\n Alexis Palmer and\n Rolando A. Coto Solano and\n Ngoc Thang Vu and\n Katharina Kann},\n title = {AmericasNLI: Evaluating Zero-shot Natural Language Understanding of\n Pretrained Multilingual Models in Truly Low-resource Languages},\n journal = {CoRR},\n volume = {abs/2104.08726},\n year = {2021},\n url = {https://arxiv.org/abs/2104.08726},\n eprinttype = {arXiv},\n eprint = {2104.08726},\n timestamp = {Mon, 26 Apr 2021 17:25:10 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-2104-08726.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
209
- "homepage": "https://github.com/nala-cub/AmericasNLI",
210
- "license": "",
211
- "features": {
212
- "premise": {
213
- "dtype": "string",
214
- "_type": "Value"
215
- },
216
- "hypothesis": {
217
- "dtype": "string",
218
- "_type": "Value"
219
- },
220
- "label": {
221
- "names": [
222
- "entailment",
223
- "neutral",
224
- "contradiction"
225
- ],
226
- "_type": "ClassLabel"
227
- }
228
- },
229
- "builder_name": "americas_nli",
230
- "dataset_name": "americas_nli",
231
- "config_name": "hch",
232
- "version": {
233
- "version_str": "1.0.0",
234
- "description": "",
235
- "major": 1,
236
- "minor": 0,
237
- "patch": 0
238
- },
239
- "splits": {
240
- "validation": {
241
- "name": "validation",
242
- "num_bytes": 127966,
243
- "num_examples": 743,
244
- "dataset_name": null
245
- },
246
- "test": {
247
- "name": "test",
248
- "num_bytes": 120857,
249
- "num_examples": 750,
250
- "dataset_name": null
251
- }
252
- },
253
- "download_size": 90748,
254
- "dataset_size": 248823,
255
- "size_in_bytes": 339571
256
- },
257
- "nah": {
258
- "description": "AmericasNLI is an extension of XNLI (Conneau et al., 2018) \u2013 a natural language inference (NLI) dataset covering 15 high-resource languages \u2013 to 10 low-resource indigenous languages spoken in the Americas: Ashaninka, Aymara, Bribri, Guarani, Nahuatl, Otomi, Quechua, Raramuri, Shipibo-Konibo, and Wixarika. As with MNLI, the goal is to predict textual entailment (does sentence A imply/contradict/neither sentence B) and is a classification task (given two sentences, predict one of three labels).\n",
259
- "citation": "\n@article{DBLP:journals/corr/abs-2104-08726,\n author = {Abteen Ebrahimi and\n Manuel Mager and\n Arturo Oncevay and\n Vishrav Chaudhary and\n Luis Chiruzzo and\n Angela Fan and\n John Ortega and\n Ricardo Ramos and\n Annette Rios and\n Ivan Vladimir and\n Gustavo A. Gim{'{e}}nez{-}Lugo and\n Elisabeth Mager and\n Graham Neubig and\n Alexis Palmer and\n Rolando A. Coto Solano and\n Ngoc Thang Vu and\n Katharina Kann},\n title = {AmericasNLI: Evaluating Zero-shot Natural Language Understanding of\n Pretrained Multilingual Models in Truly Low-resource Languages},\n journal = {CoRR},\n volume = {abs/2104.08726},\n year = {2021},\n url = {https://arxiv.org/abs/2104.08726},\n eprinttype = {arXiv},\n eprint = {2104.08726},\n timestamp = {Mon, 26 Apr 2021 17:25:10 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-2104-08726.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
260
- "homepage": "https://github.com/nala-cub/AmericasNLI",
261
- "license": "",
262
- "features": {
263
- "premise": {
264
- "dtype": "string",
265
- "_type": "Value"
266
- },
267
- "hypothesis": {
268
- "dtype": "string",
269
- "_type": "Value"
270
- },
271
- "label": {
272
- "names": [
273
- "entailment",
274
- "neutral",
275
- "contradiction"
276
- ],
277
- "_type": "ClassLabel"
278
- }
279
- },
280
- "builder_name": "americas_nli",
281
- "dataset_name": "americas_nli",
282
- "config_name": "nah",
283
- "version": {
284
- "version_str": "1.0.0",
285
- "description": "",
286
- "major": 1,
287
- "minor": 0,
288
- "patch": 0
289
- },
290
- "splits": {
291
- "validation": {
292
- "name": "validation",
293
- "num_bytes": 50741,
294
- "num_examples": 376,
295
- "dataset_name": null
296
- },
297
- "test": {
298
- "name": "test",
299
- "num_bytes": 102953,
300
- "num_examples": 738,
301
- "dataset_name": null
302
- }
303
- },
304
- "download_size": 56953,
305
- "dataset_size": 153694,
306
- "size_in_bytes": 210647
307
- },
308
- "oto": {
309
- "description": "AmericasNLI is an extension of XNLI (Conneau et al., 2018) \u2013 a natural language inference (NLI) dataset covering 15 high-resource languages \u2013 to 10 low-resource indigenous languages spoken in the Americas: Ashaninka, Aymara, Bribri, Guarani, Nahuatl, Otomi, Quechua, Raramuri, Shipibo-Konibo, and Wixarika. As with MNLI, the goal is to predict textual entailment (does sentence A imply/contradict/neither sentence B) and is a classification task (given two sentences, predict one of three labels).\n",
310
- "citation": "\n@article{DBLP:journals/corr/abs-2104-08726,\n author = {Abteen Ebrahimi and\n Manuel Mager and\n Arturo Oncevay and\n Vishrav Chaudhary and\n Luis Chiruzzo and\n Angela Fan and\n John Ortega and\n Ricardo Ramos and\n Annette Rios and\n Ivan Vladimir and\n Gustavo A. Gim{'{e}}nez{-}Lugo and\n Elisabeth Mager and\n Graham Neubig and\n Alexis Palmer and\n Rolando A. Coto Solano and\n Ngoc Thang Vu and\n Katharina Kann},\n title = {AmericasNLI: Evaluating Zero-shot Natural Language Understanding of\n Pretrained Multilingual Models in Truly Low-resource Languages},\n journal = {CoRR},\n volume = {abs/2104.08726},\n year = {2021},\n url = {https://arxiv.org/abs/2104.08726},\n eprinttype = {arXiv},\n eprint = {2104.08726},\n timestamp = {Mon, 26 Apr 2021 17:25:10 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-2104-08726.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
311
- "homepage": "https://github.com/nala-cub/AmericasNLI",
312
- "license": "",
313
- "features": {
314
- "premise": {
315
- "dtype": "string",
316
- "_type": "Value"
317
- },
318
- "hypothesis": {
319
- "dtype": "string",
320
- "_type": "Value"
321
- },
322
- "label": {
323
- "names": [
324
- "entailment",
325
- "neutral",
326
- "contradiction"
327
- ],
328
- "_type": "ClassLabel"
329
- }
330
- },
331
- "builder_name": "americas_nli",
332
- "dataset_name": "americas_nli",
333
- "config_name": "oto",
334
- "version": {
335
- "version_str": "1.0.0",
336
- "description": "",
337
- "major": 1,
338
- "minor": 0,
339
- "patch": 0
340
- },
341
- "splits": {
342
- "validation": {
343
- "name": "validation",
344
- "num_bytes": 27010,
345
- "num_examples": 222,
346
- "dataset_name": null
347
- },
348
- "test": {
349
- "name": "test",
350
- "num_bytes": 119650,
351
- "num_examples": 748,
352
- "dataset_name": null
353
- }
354
- },
355
- "download_size": 57849,
356
- "dataset_size": 146660,
357
- "size_in_bytes": 204509
358
- },
359
- "quy": {
360
- "description": "AmericasNLI is an extension of XNLI (Conneau et al., 2018) \u2013 a natural language inference (NLI) dataset covering 15 high-resource languages \u2013 to 10 low-resource indigenous languages spoken in the Americas: Ashaninka, Aymara, Bribri, Guarani, Nahuatl, Otomi, Quechua, Raramuri, Shipibo-Konibo, and Wixarika. As with MNLI, the goal is to predict textual entailment (does sentence A imply/contradict/neither sentence B) and is a classification task (given two sentences, predict one of three labels).\n",
361
- "citation": "\n@article{DBLP:journals/corr/abs-2104-08726,\n author = {Abteen Ebrahimi and\n Manuel Mager and\n Arturo Oncevay and\n Vishrav Chaudhary and\n Luis Chiruzzo and\n Angela Fan and\n John Ortega and\n Ricardo Ramos and\n Annette Rios and\n Ivan Vladimir and\n Gustavo A. Gim{'{e}}nez{-}Lugo and\n Elisabeth Mager and\n Graham Neubig and\n Alexis Palmer and\n Rolando A. Coto Solano and\n Ngoc Thang Vu and\n Katharina Kann},\n title = {AmericasNLI: Evaluating Zero-shot Natural Language Understanding of\n Pretrained Multilingual Models in Truly Low-resource Languages},\n journal = {CoRR},\n volume = {abs/2104.08726},\n year = {2021},\n url = {https://arxiv.org/abs/2104.08726},\n eprinttype = {arXiv},\n eprint = {2104.08726},\n timestamp = {Mon, 26 Apr 2021 17:25:10 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-2104-08726.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
362
- "homepage": "https://github.com/nala-cub/AmericasNLI",
363
- "license": "",
364
- "features": {
365
- "premise": {
366
- "dtype": "string",
367
- "_type": "Value"
368
- },
369
- "hypothesis": {
370
- "dtype": "string",
371
- "_type": "Value"
372
- },
373
- "label": {
374
- "names": [
375
- "entailment",
376
- "neutral",
377
- "contradiction"
378
- ],
379
- "_type": "ClassLabel"
380
- }
381
- },
382
- "builder_name": "americas_nli",
383
- "dataset_name": "americas_nli",
384
- "config_name": "quy",
385
- "version": {
386
- "version_str": "1.0.0",
387
- "description": "",
388
- "major": 1,
389
- "minor": 0,
390
- "patch": 0
391
- },
392
- "splits": {
393
- "validation": {
394
- "name": "validation",
395
- "num_bytes": 125636,
396
- "num_examples": 743,
397
- "dataset_name": null
398
- },
399
- "test": {
400
- "name": "test",
401
- "num_bytes": 112750,
402
- "num_examples": 750,
403
- "dataset_name": null
404
- }
405
- },
406
- "download_size": 85673,
407
- "dataset_size": 238386,
408
- "size_in_bytes": 324059
409
- },
410
- "shp": {
411
- "description": "AmericasNLI is an extension of XNLI (Conneau et al., 2018) \u2013 a natural language inference (NLI) dataset covering 15 high-resource languages \u2013 to 10 low-resource indigenous languages spoken in the Americas: Ashaninka, Aymara, Bribri, Guarani, Nahuatl, Otomi, Quechua, Raramuri, Shipibo-Konibo, and Wixarika. As with MNLI, the goal is to predict textual entailment (does sentence A imply/contradict/neither sentence B) and is a classification task (given two sentences, predict one of three labels).\n",
412
- "citation": "\n@article{DBLP:journals/corr/abs-2104-08726,\n author = {Abteen Ebrahimi and\n Manuel Mager and\n Arturo Oncevay and\n Vishrav Chaudhary and\n Luis Chiruzzo and\n Angela Fan and\n John Ortega and\n Ricardo Ramos and\n Annette Rios and\n Ivan Vladimir and\n Gustavo A. Gim{'{e}}nez{-}Lugo and\n Elisabeth Mager and\n Graham Neubig and\n Alexis Palmer and\n Rolando A. Coto Solano and\n Ngoc Thang Vu and\n Katharina Kann},\n title = {AmericasNLI: Evaluating Zero-shot Natural Language Understanding of\n Pretrained Multilingual Models in Truly Low-resource Languages},\n journal = {CoRR},\n volume = {abs/2104.08726},\n year = {2021},\n url = {https://arxiv.org/abs/2104.08726},\n eprinttype = {arXiv},\n eprint = {2104.08726},\n timestamp = {Mon, 26 Apr 2021 17:25:10 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-2104-08726.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
413
- "homepage": "https://github.com/nala-cub/AmericasNLI",
414
- "license": "",
415
- "features": {
416
- "premise": {
417
- "dtype": "string",
418
- "_type": "Value"
419
- },
420
- "hypothesis": {
421
- "dtype": "string",
422
- "_type": "Value"
423
- },
424
- "label": {
425
- "names": [
426
- "entailment",
427
- "neutral",
428
- "contradiction"
429
- ],
430
- "_type": "ClassLabel"
431
- }
432
- },
433
- "builder_name": "americas_nli",
434
- "dataset_name": "americas_nli",
435
- "config_name": "shp",
436
- "version": {
437
- "version_str": "1.0.0",
438
- "description": "",
439
- "major": 1,
440
- "minor": 0,
441
- "patch": 0
442
- },
443
- "splits": {
444
- "validation": {
445
- "name": "validation",
446
- "num_bytes": 124500,
447
- "num_examples": 743,
448
- "dataset_name": null
449
- },
450
- "test": {
451
- "name": "test",
452
- "num_bytes": 118934,
453
- "num_examples": 750,
454
- "dataset_name": null
455
- }
456
- },
457
- "download_size": 85544,
458
- "dataset_size": 243434,
459
- "size_in_bytes": 328978
460
- },
461
- "tar": {
462
- "description": "AmericasNLI is an extension of XNLI (Conneau et al., 2018) \u2013 a natural language inference (NLI) dataset covering 15 high-resource languages \u2013 to 10 low-resource indigenous languages spoken in the Americas: Ashaninka, Aymara, Bribri, Guarani, Nahuatl, Otomi, Quechua, Raramuri, Shipibo-Konibo, and Wixarika. As with MNLI, the goal is to predict textual entailment (does sentence A imply/contradict/neither sentence B) and is a classification task (given two sentences, predict one of three labels).\n",
463
- "citation": "\n@article{DBLP:journals/corr/abs-2104-08726,\n author = {Abteen Ebrahimi and\n Manuel Mager and\n Arturo Oncevay and\n Vishrav Chaudhary and\n Luis Chiruzzo and\n Angela Fan and\n John Ortega and\n Ricardo Ramos and\n Annette Rios and\n Ivan Vladimir and\n Gustavo A. Gim{'{e}}nez{-}Lugo and\n Elisabeth Mager and\n Graham Neubig and\n Alexis Palmer and\n Rolando A. Coto Solano and\n Ngoc Thang Vu and\n Katharina Kann},\n title = {AmericasNLI: Evaluating Zero-shot Natural Language Understanding of\n Pretrained Multilingual Models in Truly Low-resource Languages},\n journal = {CoRR},\n volume = {abs/2104.08726},\n year = {2021},\n url = {https://arxiv.org/abs/2104.08726},\n eprinttype = {arXiv},\n eprint = {2104.08726},\n timestamp = {Mon, 26 Apr 2021 17:25:10 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-2104-08726.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
464
- "homepage": "https://github.com/nala-cub/AmericasNLI",
465
- "license": "",
466
- "features": {
467
- "premise": {
468
- "dtype": "string",
469
- "_type": "Value"
470
- },
471
- "hypothesis": {
472
- "dtype": "string",
473
- "_type": "Value"
474
- },
475
- "label": {
476
- "names": [
477
- "entailment",
478
- "neutral",
479
- "contradiction"
480
- ],
481
- "_type": "ClassLabel"
482
- }
483
- },
484
- "builder_name": "americas_nli",
485
- "dataset_name": "americas_nli",
486
- "config_name": "tar",
487
- "version": {
488
- "version_str": "1.0.0",
489
- "description": "",
490
- "major": 1,
491
- "minor": 0,
492
- "patch": 0
493
- },
494
- "splits": {
495
- "validation": {
496
- "name": "validation",
497
- "num_bytes": 139496,
498
- "num_examples": 743,
499
- "dataset_name": null
500
- },
501
- "test": {
502
- "name": "test",
503
- "num_bytes": 122624,
504
- "num_examples": 750,
505
- "dataset_name": null
506
- }
507
- },
508
- "download_size": 89683,
509
- "dataset_size": 262120,
510
- "size_in_bytes": 351803
511
- },
512
- "all_languages": {
513
- "description": "AmericasNLI is an extension of XNLI (Conneau et al., 2018) \u2013 a natural language inference (NLI) dataset covering 15 high-resource languages \u2013 to 10 low-resource indigenous languages spoken in the Americas: Ashaninka, Aymara, Bribri, Guarani, Nahuatl, Otomi, Quechua, Raramuri, Shipibo-Konibo, and Wixarika. As with MNLI, the goal is to predict textual entailment (does sentence A imply/contradict/neither sentence B) and is a classification task (given two sentences, predict one of three labels).\n",
514
- "citation": "\n@article{DBLP:journals/corr/abs-2104-08726,\n author = {Abteen Ebrahimi and\n Manuel Mager and\n Arturo Oncevay and\n Vishrav Chaudhary and\n Luis Chiruzzo and\n Angela Fan and\n John Ortega and\n Ricardo Ramos and\n Annette Rios and\n Ivan Vladimir and\n Gustavo A. Gim{'{e}}nez{-}Lugo and\n Elisabeth Mager and\n Graham Neubig and\n Alexis Palmer and\n Rolando A. Coto Solano and\n Ngoc Thang Vu and\n Katharina Kann},\n title = {AmericasNLI: Evaluating Zero-shot Natural Language Understanding of\n Pretrained Multilingual Models in Truly Low-resource Languages},\n journal = {CoRR},\n volume = {abs/2104.08726},\n year = {2021},\n url = {https://arxiv.org/abs/2104.08726},\n eprinttype = {arXiv},\n eprint = {2104.08726},\n timestamp = {Mon, 26 Apr 2021 17:25:10 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-2104-08726.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
515
- "homepage": "https://github.com/nala-cub/AmericasNLI",
516
- "license": "",
517
- "features": {
518
- "language": {
519
- "dtype": "string",
520
- "_type": "Value"
521
- },
522
- "premise": {
523
- "dtype": "string",
524
- "_type": "Value"
525
- },
526
- "hypothesis": {
527
- "dtype": "string",
528
- "_type": "Value"
529
- },
530
- "label": {
531
- "names": [
532
- "entailment",
533
- "neutral",
534
- "contradiction"
535
- ],
536
- "_type": "ClassLabel"
537
- }
538
- },
539
- "builder_name": "americas_nli",
540
- "dataset_name": "americas_nli",
541
- "config_name": "all_languages",
542
- "version": {
543
- "version_str": "1.0.0",
544
- "description": "",
545
- "major": 1,
546
- "minor": 0,
547
- "patch": 0
548
- },
549
- "splits": {
550
- "validation": {
551
- "name": "validation",
552
- "num_bytes": 1129080,
553
- "num_examples": 6457,
554
- "dataset_name": null
555
- },
556
- "test": {
557
- "name": "test",
558
- "num_bytes": 1210579,
559
- "num_examples": 7486,
560
- "dataset_name": null
561
- }
562
- },
563
- "download_size": 791239,
564
- "dataset_size": 2339659,
565
- "size_in_bytes": 3130898
566
- }
567
- }