albertvillanova HF staff commited on
Commit
51adfef
1 Parent(s): 8c2924a

Convert dataset to Parquet (#3)

Browse files

- Convert dataset to Parquet (64d6c4344e064f92b1b59d65e0b139566427f820)
- Add xquad.de data files (b3610657ec129fbcf4a0de4f42c5d9ccdcbecb23)
- Add xquad.zh data files (e2e6e707b8e3e3e0cfa31552670946e35fb03a9f)
- Add xquad.vi data files (446539ea43ee7a3b3425f5fc4d4cfe869e3c4679)
- Add xquad.en data files (22ede0325610e959c4e2888f66ada97c78a3b9da)
- Add xquad.es data files (4236455a0538ec13a378732ba03b135d4097ce56)
- Add xquad.hi data files (3b1f0efbbf94278b0321ffd4c4449c264348fe6f)
- Add xquad.el data files (b165a89ec81992f9292690d73410050b1cec01df)
- Add xquad.th data files (61b3ea7d3507831e0f90816bda646de7302f2bc8)
- Add xquad.tr data files (10eb38614c82d81466b0caeeb276bd127f63c245)
- Add xquad.ru data files (7bb3de4b722b64ada6bdbe595bd1feb7eb114ac1)
- Add xquad.ro data files (285410c0ec8dbe7e9bec1ad2daf5f4111ecf84eb)
- Delete loading script (ef4d39b821d7b5b63a1f3af408f422810fee8ba6)
- Delete legacy dataset_infos.json (cfe40ec6f10666ae121463139e212d647a7866b4)

README.md CHANGED
@@ -1,5 +1,4 @@
1
  ---
2
- pretty_name: XQuAD
3
  annotations_creators:
4
  - expert-generated
5
  language_creators:
@@ -30,6 +29,7 @@ task_categories:
30
  task_ids:
31
  - extractive-qa
32
  paperswithcode_id: xquad
 
33
  dataset_info:
34
  - config_name: xquad.ar
35
  features:
@@ -47,10 +47,10 @@ dataset_info:
47
  dtype: int32
48
  splits:
49
  - name: validation
50
- num_bytes: 1722799
51
  num_examples: 1190
52
- download_size: 13962158
53
- dataset_size: 1722799
54
  - config_name: xquad.de
55
  features:
56
  - name: id
@@ -67,11 +67,11 @@ dataset_info:
67
  dtype: int32
68
  splits:
69
  - name: validation
70
- num_bytes: 1283301
71
  num_examples: 1190
72
- download_size: 13962158
73
- dataset_size: 1283301
74
- - config_name: xquad.zh
75
  features:
76
  - name: id
77
  dtype: string
@@ -87,11 +87,11 @@ dataset_info:
87
  dtype: int32
88
  splits:
89
  - name: validation
90
- num_bytes: 984241
91
  num_examples: 1190
92
- download_size: 13962158
93
- dataset_size: 984241
94
- - config_name: xquad.vi
95
  features:
96
  - name: id
97
  dtype: string
@@ -107,11 +107,11 @@ dataset_info:
107
  dtype: int32
108
  splits:
109
  - name: validation
110
- num_bytes: 1477239
111
  num_examples: 1190
112
- download_size: 13962158
113
- dataset_size: 1477239
114
- - config_name: xquad.en
115
  features:
116
  - name: id
117
  dtype: string
@@ -127,11 +127,11 @@ dataset_info:
127
  dtype: int32
128
  splits:
129
  - name: validation
130
- num_bytes: 1116123
131
  num_examples: 1190
132
- download_size: 13962158
133
- dataset_size: 1116123
134
- - config_name: xquad.es
135
  features:
136
  - name: id
137
  dtype: string
@@ -147,11 +147,11 @@ dataset_info:
147
  dtype: int32
148
  splits:
149
  - name: validation
150
- num_bytes: 1273499
151
  num_examples: 1190
152
- download_size: 13962158
153
- dataset_size: 1273499
154
- - config_name: xquad.hi
155
  features:
156
  - name: id
157
  dtype: string
@@ -167,11 +167,11 @@ dataset_info:
167
  dtype: int32
168
  splits:
169
  - name: validation
170
- num_bytes: 2682975
171
  num_examples: 1190
172
- download_size: 13962158
173
- dataset_size: 2682975
174
- - config_name: xquad.el
175
  features:
176
  - name: id
177
  dtype: string
@@ -187,10 +187,10 @@ dataset_info:
187
  dtype: int32
188
  splits:
189
  - name: validation
190
- num_bytes: 2206690
191
  num_examples: 1190
192
- download_size: 13962158
193
- dataset_size: 2206690
194
  - config_name: xquad.th
195
  features:
196
  - name: id
@@ -207,10 +207,10 @@ dataset_info:
207
  dtype: int32
208
  splits:
209
  - name: validation
210
- num_bytes: 2854959
211
  num_examples: 1190
212
- download_size: 13962158
213
- dataset_size: 2854959
214
  - config_name: xquad.tr
215
  features:
216
  - name: id
@@ -227,11 +227,11 @@ dataset_info:
227
  dtype: int32
228
  splits:
229
  - name: validation
230
- num_bytes: 1210763
231
  num_examples: 1190
232
- download_size: 13962158
233
- dataset_size: 1210763
234
- - config_name: xquad.ru
235
  features:
236
  - name: id
237
  dtype: string
@@ -247,11 +247,11 @@ dataset_info:
247
  dtype: int32
248
  splits:
249
  - name: validation
250
- num_bytes: 2136990
251
  num_examples: 1190
252
- download_size: 13962158
253
- dataset_size: 2136990
254
- - config_name: xquad.ro
255
  features:
256
  - name: id
257
  dtype: string
@@ -267,10 +267,59 @@ dataset_info:
267
  dtype: int32
268
  splits:
269
  - name: validation
270
- num_bytes: 1299450
271
  num_examples: 1190
272
- download_size: 13962158
273
- dataset_size: 1299450
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
274
  ---
275
 
276
  # Dataset Card for "xquad"
1
  ---
 
2
  annotations_creators:
3
  - expert-generated
4
  language_creators:
29
  task_ids:
30
  - extractive-qa
31
  paperswithcode_id: xquad
32
+ pretty_name: XQuAD
33
  dataset_info:
34
  - config_name: xquad.ar
35
  features:
47
  dtype: int32
48
  splits:
49
  - name: validation
50
+ num_bytes: 1722775
51
  num_examples: 1190
52
+ download_size: 263002
53
+ dataset_size: 1722775
54
  - config_name: xquad.de
55
  features:
56
  - name: id
67
  dtype: int32
68
  splits:
69
  - name: validation
70
+ num_bytes: 1283277
71
  num_examples: 1190
72
+ download_size: 241957
73
+ dataset_size: 1283277
74
+ - config_name: xquad.el
75
  features:
76
  - name: id
77
  dtype: string
87
  dtype: int32
88
  splits:
89
  - name: validation
90
+ num_bytes: 2206666
91
  num_examples: 1190
92
+ download_size: 324379
93
+ dataset_size: 2206666
94
+ - config_name: xquad.en
95
  features:
96
  - name: id
97
  dtype: string
107
  dtype: int32
108
  splits:
109
  - name: validation
110
+ num_bytes: 1116099
111
  num_examples: 1190
112
+ download_size: 212372
113
+ dataset_size: 1116099
114
+ - config_name: xquad.es
115
  features:
116
  - name: id
117
  dtype: string
127
  dtype: int32
128
  splits:
129
  - name: validation
130
+ num_bytes: 1273475
131
  num_examples: 1190
132
+ download_size: 236874
133
+ dataset_size: 1273475
134
+ - config_name: xquad.hi
135
  features:
136
  - name: id
137
  dtype: string
147
  dtype: int32
148
  splits:
149
  - name: validation
150
+ num_bytes: 2682951
151
  num_examples: 1190
152
+ download_size: 322083
153
+ dataset_size: 2682951
154
+ - config_name: xquad.ro
155
  features:
156
  - name: id
157
  dtype: string
167
  dtype: int32
168
  splits:
169
  - name: validation
170
+ num_bytes: 1299426
171
  num_examples: 1190
172
+ download_size: 244467
173
+ dataset_size: 1299426
174
+ - config_name: xquad.ru
175
  features:
176
  - name: id
177
  dtype: string
187
  dtype: int32
188
  splits:
189
  - name: validation
190
+ num_bytes: 2136966
191
  num_examples: 1190
192
+ download_size: 321728
193
+ dataset_size: 2136966
194
  - config_name: xquad.th
195
  features:
196
  - name: id
207
  dtype: int32
208
  splits:
209
  - name: validation
210
+ num_bytes: 2854935
211
  num_examples: 1190
212
+ download_size: 337307
213
+ dataset_size: 2854935
214
  - config_name: xquad.tr
215
  features:
216
  - name: id
227
  dtype: int32
228
  splits:
229
  - name: validation
230
+ num_bytes: 1210739
231
  num_examples: 1190
232
+ download_size: 228364
233
+ dataset_size: 1210739
234
+ - config_name: xquad.vi
235
  features:
236
  - name: id
237
  dtype: string
247
  dtype: int32
248
  splits:
249
  - name: validation
250
+ num_bytes: 1477215
251
  num_examples: 1190
252
+ download_size: 237644
253
+ dataset_size: 1477215
254
+ - config_name: xquad.zh
255
  features:
256
  - name: id
257
  dtype: string
267
  dtype: int32
268
  splits:
269
  - name: validation
270
+ num_bytes: 984217
271
  num_examples: 1190
272
+ download_size: 205768
273
+ dataset_size: 984217
274
+ configs:
275
+ - config_name: xquad.ar
276
+ data_files:
277
+ - split: validation
278
+ path: xquad.ar/validation-*
279
+ - config_name: xquad.de
280
+ data_files:
281
+ - split: validation
282
+ path: xquad.de/validation-*
283
+ - config_name: xquad.el
284
+ data_files:
285
+ - split: validation
286
+ path: xquad.el/validation-*
287
+ - config_name: xquad.en
288
+ data_files:
289
+ - split: validation
290
+ path: xquad.en/validation-*
291
+ - config_name: xquad.es
292
+ data_files:
293
+ - split: validation
294
+ path: xquad.es/validation-*
295
+ - config_name: xquad.hi
296
+ data_files:
297
+ - split: validation
298
+ path: xquad.hi/validation-*
299
+ - config_name: xquad.ro
300
+ data_files:
301
+ - split: validation
302
+ path: xquad.ro/validation-*
303
+ - config_name: xquad.ru
304
+ data_files:
305
+ - split: validation
306
+ path: xquad.ru/validation-*
307
+ - config_name: xquad.th
308
+ data_files:
309
+ - split: validation
310
+ path: xquad.th/validation-*
311
+ - config_name: xquad.tr
312
+ data_files:
313
+ - split: validation
314
+ path: xquad.tr/validation-*
315
+ - config_name: xquad.vi
316
+ data_files:
317
+ - split: validation
318
+ path: xquad.vi/validation-*
319
+ - config_name: xquad.zh
320
+ data_files:
321
+ - split: validation
322
+ path: xquad.zh/validation-*
323
  ---
324
 
325
  # Dataset Card for "xquad"
dataset_infos.json DELETED
@@ -1 +0,0 @@
1
- {"xquad.ar": {"description": "XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering\nperformance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set\nof SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten languages: Spanish, German,\nGreek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel\nacross 12 languages.\n", "citation": "@article{Artetxe:etal:2019,\n author = {Mikel Artetxe and Sebastian Ruder and Dani Yogatama},\n title = {On the cross-lingual transferability of monolingual representations},\n journal = {CoRR},\n volume = {abs/1910.11856},\n year = {2019},\n archivePrefix = {arXiv},\n eprint = {1910.11856}\n}\n", "homepage": "https://github.com/deepmind/xquad", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "context": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "answers": {"feature": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "answer_start": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "task_templates": [{"task": "question-answering-extractive", "question_column": "question", "context_column": "context", "answers_column": "answers"}], "builder_name": "xquad", "config_name": "xquad.ar", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"validation": {"name": "validation", "num_bytes": 1722799, "num_examples": 1190, "dataset_name": "xquad"}}, "download_checksums": {"https://github.com/deepmind/xquad/raw/master/xquad.ar.json": {"num_bytes": 1582988, "checksum": "abdabd7afed5c635d99cca0f3f0d0c9d9ed0bc77451e963c2e4e0638c29e486d"}, "https://github.com/deepmind/xquad/raw/master/xquad.de.json": {"num_bytes": 669810, "checksum": "990b5d746746ed65ed4702ea5f35f99ffa4e2f1c390c07d003642acd937916f9"}, "https://github.com/deepmind/xquad/raw/master/xquad.zh.json": {"num_bytes": 808652, "checksum": "691d0b3359bc6b8faa8de931dfdfe21d50a65861ae348e32a0d1a0190b0c8835"}, "https://github.com/deepmind/xquad/raw/master/xquad.vi.json": {"num_bytes": 911401, "checksum": "f619a1eb11fb42d3ab0834259e488a65f585447ef6154437bfb7199d85161a04"}, "https://github.com/deepmind/xquad/raw/master/xquad.en.json": {"num_bytes": 609383, "checksum": "e4c57d1c9143aaa1c5d265ba5987a65f4e69528d2a98f29d6e75019b10344f29"}, "https://github.com/deepmind/xquad/raw/master/xquad.es.json": {"num_bytes": 684322, "checksum": "dcbae93ec3a9f4b9e78fd834a171d6f96c1a875e10e15b7530b7e4ef4971e37e"}, "https://github.com/deepmind/xquad/raw/master/xquad.hi.json": {"num_bytes": 1680538, "checksum": "df2cce3532b37e9beb8979704b5c9a4bf874358f105395a298b89427b43b9d24"}, "https://github.com/deepmind/xquad/raw/master/xquad.el.json": {"num_bytes": 1918889, "checksum": "821cf0f88e73fa258fd2f548b19b6ec39f7025059e16f6f9fc8cd797c9c3663e"}, "https://github.com/deepmind/xquad/raw/master/xquad.th.json": {"num_bytes": 1809143, "checksum": "5cdda11d0e1e075f7872abf4e6ae830388ce7f617964d542308e9ae4257e0f43"}, "https://github.com/deepmind/xquad/raw/master/xquad.tr.json": {"num_bytes": 729506, "checksum": "92179a564774b7696100d144c1e10870d0a966b6fccbdd254a65b9d2ab1971cc"}, "https://github.com/deepmind/xquad/raw/master/xquad.ru.json": {"num_bytes": 1896368, "checksum": "208d5b1aa154c52b1b5c5eda16281e455e8fd198cdb9af3f469f0d6037d973bf"}, "https://github.com/deepmind/xquad/raw/master/xquad.ro.json": {"num_bytes": 661158, "checksum": "94f8d54e93ad2b7dd2e76dd921a666b5b8eff2d4379df9f179c42b445f09643f"}}, "download_size": 13962158, "post_processing_size": null, "dataset_size": 1722799, "size_in_bytes": 15684957}, "xquad.de": {"description": "XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering\nperformance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set\nof SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten languages: Spanish, German,\nGreek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel\nacross 12 languages.\n", "citation": "@article{Artetxe:etal:2019,\n author = {Mikel Artetxe and Sebastian Ruder and Dani Yogatama},\n title = {On the cross-lingual transferability of monolingual representations},\n journal = {CoRR},\n volume = {abs/1910.11856},\n year = {2019},\n archivePrefix = {arXiv},\n eprint = {1910.11856}\n}\n", "homepage": "https://github.com/deepmind/xquad", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "context": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "answers": {"feature": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "answer_start": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "task_templates": [{"task": "question-answering-extractive", "question_column": "question", "context_column": "context", "answers_column": "answers"}], "builder_name": "xquad", "config_name": "xquad.de", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"validation": {"name": "validation", "num_bytes": 1283301, "num_examples": 1190, "dataset_name": "xquad"}}, "download_checksums": {"https://github.com/deepmind/xquad/raw/master/xquad.ar.json": {"num_bytes": 1582988, "checksum": "abdabd7afed5c635d99cca0f3f0d0c9d9ed0bc77451e963c2e4e0638c29e486d"}, "https://github.com/deepmind/xquad/raw/master/xquad.de.json": {"num_bytes": 669810, "checksum": "990b5d746746ed65ed4702ea5f35f99ffa4e2f1c390c07d003642acd937916f9"}, "https://github.com/deepmind/xquad/raw/master/xquad.zh.json": {"num_bytes": 808652, "checksum": "691d0b3359bc6b8faa8de931dfdfe21d50a65861ae348e32a0d1a0190b0c8835"}, "https://github.com/deepmind/xquad/raw/master/xquad.vi.json": {"num_bytes": 911401, "checksum": "f619a1eb11fb42d3ab0834259e488a65f585447ef6154437bfb7199d85161a04"}, "https://github.com/deepmind/xquad/raw/master/xquad.en.json": {"num_bytes": 609383, "checksum": "e4c57d1c9143aaa1c5d265ba5987a65f4e69528d2a98f29d6e75019b10344f29"}, "https://github.com/deepmind/xquad/raw/master/xquad.es.json": {"num_bytes": 684322, "checksum": "dcbae93ec3a9f4b9e78fd834a171d6f96c1a875e10e15b7530b7e4ef4971e37e"}, "https://github.com/deepmind/xquad/raw/master/xquad.hi.json": {"num_bytes": 1680538, "checksum": "df2cce3532b37e9beb8979704b5c9a4bf874358f105395a298b89427b43b9d24"}, "https://github.com/deepmind/xquad/raw/master/xquad.el.json": {"num_bytes": 1918889, "checksum": "821cf0f88e73fa258fd2f548b19b6ec39f7025059e16f6f9fc8cd797c9c3663e"}, "https://github.com/deepmind/xquad/raw/master/xquad.th.json": {"num_bytes": 1809143, "checksum": "5cdda11d0e1e075f7872abf4e6ae830388ce7f617964d542308e9ae4257e0f43"}, "https://github.com/deepmind/xquad/raw/master/xquad.tr.json": {"num_bytes": 729506, "checksum": "92179a564774b7696100d144c1e10870d0a966b6fccbdd254a65b9d2ab1971cc"}, "https://github.com/deepmind/xquad/raw/master/xquad.ru.json": {"num_bytes": 1896368, "checksum": "208d5b1aa154c52b1b5c5eda16281e455e8fd198cdb9af3f469f0d6037d973bf"}, "https://github.com/deepmind/xquad/raw/master/xquad.ro.json": {"num_bytes": 661158, "checksum": "94f8d54e93ad2b7dd2e76dd921a666b5b8eff2d4379df9f179c42b445f09643f"}}, "download_size": 13962158, "post_processing_size": null, "dataset_size": 1283301, "size_in_bytes": 15245459}, "xquad.zh": {"description": "XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering\nperformance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set\nof SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten languages: Spanish, German,\nGreek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel\nacross 12 languages.\n", "citation": "@article{Artetxe:etal:2019,\n author = {Mikel Artetxe and Sebastian Ruder and Dani Yogatama},\n title = {On the cross-lingual transferability of monolingual representations},\n journal = {CoRR},\n volume = {abs/1910.11856},\n year = {2019},\n archivePrefix = {arXiv},\n eprint = {1910.11856}\n}\n", "homepage": "https://github.com/deepmind/xquad", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "context": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "answers": {"feature": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "answer_start": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "task_templates": [{"task": "question-answering-extractive", "question_column": "question", "context_column": "context", "answers_column": "answers"}], "builder_name": "xquad", "config_name": "xquad.zh", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"validation": {"name": "validation", "num_bytes": 984241, "num_examples": 1190, "dataset_name": "xquad"}}, "download_checksums": {"https://github.com/deepmind/xquad/raw/master/xquad.ar.json": {"num_bytes": 1582988, "checksum": "abdabd7afed5c635d99cca0f3f0d0c9d9ed0bc77451e963c2e4e0638c29e486d"}, "https://github.com/deepmind/xquad/raw/master/xquad.de.json": {"num_bytes": 669810, "checksum": "990b5d746746ed65ed4702ea5f35f99ffa4e2f1c390c07d003642acd937916f9"}, "https://github.com/deepmind/xquad/raw/master/xquad.zh.json": {"num_bytes": 808652, "checksum": "691d0b3359bc6b8faa8de931dfdfe21d50a65861ae348e32a0d1a0190b0c8835"}, "https://github.com/deepmind/xquad/raw/master/xquad.vi.json": {"num_bytes": 911401, "checksum": "f619a1eb11fb42d3ab0834259e488a65f585447ef6154437bfb7199d85161a04"}, "https://github.com/deepmind/xquad/raw/master/xquad.en.json": {"num_bytes": 609383, "checksum": "e4c57d1c9143aaa1c5d265ba5987a65f4e69528d2a98f29d6e75019b10344f29"}, "https://github.com/deepmind/xquad/raw/master/xquad.es.json": {"num_bytes": 684322, "checksum": "dcbae93ec3a9f4b9e78fd834a171d6f96c1a875e10e15b7530b7e4ef4971e37e"}, "https://github.com/deepmind/xquad/raw/master/xquad.hi.json": {"num_bytes": 1680538, "checksum": "df2cce3532b37e9beb8979704b5c9a4bf874358f105395a298b89427b43b9d24"}, "https://github.com/deepmind/xquad/raw/master/xquad.el.json": {"num_bytes": 1918889, "checksum": "821cf0f88e73fa258fd2f548b19b6ec39f7025059e16f6f9fc8cd797c9c3663e"}, "https://github.com/deepmind/xquad/raw/master/xquad.th.json": {"num_bytes": 1809143, "checksum": "5cdda11d0e1e075f7872abf4e6ae830388ce7f617964d542308e9ae4257e0f43"}, "https://github.com/deepmind/xquad/raw/master/xquad.tr.json": {"num_bytes": 729506, "checksum": "92179a564774b7696100d144c1e10870d0a966b6fccbdd254a65b9d2ab1971cc"}, "https://github.com/deepmind/xquad/raw/master/xquad.ru.json": {"num_bytes": 1896368, "checksum": "208d5b1aa154c52b1b5c5eda16281e455e8fd198cdb9af3f469f0d6037d973bf"}, "https://github.com/deepmind/xquad/raw/master/xquad.ro.json": {"num_bytes": 661158, "checksum": "94f8d54e93ad2b7dd2e76dd921a666b5b8eff2d4379df9f179c42b445f09643f"}}, "download_size": 13962158, "post_processing_size": null, "dataset_size": 984241, "size_in_bytes": 14946399}, "xquad.vi": {"description": "XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering\nperformance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set\nof SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten languages: Spanish, German,\nGreek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel\nacross 12 languages.\n", "citation": "@article{Artetxe:etal:2019,\n author = {Mikel Artetxe and Sebastian Ruder and Dani Yogatama},\n title = {On the cross-lingual transferability of monolingual representations},\n journal = {CoRR},\n volume = {abs/1910.11856},\n year = {2019},\n archivePrefix = {arXiv},\n eprint = {1910.11856}\n}\n", "homepage": "https://github.com/deepmind/xquad", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "context": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "answers": {"feature": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "answer_start": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "task_templates": [{"task": "question-answering-extractive", "question_column": "question", "context_column": "context", "answers_column": "answers"}], "builder_name": "xquad", "config_name": "xquad.vi", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"validation": {"name": "validation", "num_bytes": 1477239, "num_examples": 1190, "dataset_name": "xquad"}}, "download_checksums": {"https://github.com/deepmind/xquad/raw/master/xquad.ar.json": {"num_bytes": 1582988, "checksum": "abdabd7afed5c635d99cca0f3f0d0c9d9ed0bc77451e963c2e4e0638c29e486d"}, "https://github.com/deepmind/xquad/raw/master/xquad.de.json": {"num_bytes": 669810, "checksum": "990b5d746746ed65ed4702ea5f35f99ffa4e2f1c390c07d003642acd937916f9"}, "https://github.com/deepmind/xquad/raw/master/xquad.zh.json": {"num_bytes": 808652, "checksum": "691d0b3359bc6b8faa8de931dfdfe21d50a65861ae348e32a0d1a0190b0c8835"}, "https://github.com/deepmind/xquad/raw/master/xquad.vi.json": {"num_bytes": 911401, "checksum": "f619a1eb11fb42d3ab0834259e488a65f585447ef6154437bfb7199d85161a04"}, "https://github.com/deepmind/xquad/raw/master/xquad.en.json": {"num_bytes": 609383, "checksum": "e4c57d1c9143aaa1c5d265ba5987a65f4e69528d2a98f29d6e75019b10344f29"}, "https://github.com/deepmind/xquad/raw/master/xquad.es.json": {"num_bytes": 684322, "checksum": "dcbae93ec3a9f4b9e78fd834a171d6f96c1a875e10e15b7530b7e4ef4971e37e"}, "https://github.com/deepmind/xquad/raw/master/xquad.hi.json": {"num_bytes": 1680538, "checksum": "df2cce3532b37e9beb8979704b5c9a4bf874358f105395a298b89427b43b9d24"}, "https://github.com/deepmind/xquad/raw/master/xquad.el.json": {"num_bytes": 1918889, "checksum": "821cf0f88e73fa258fd2f548b19b6ec39f7025059e16f6f9fc8cd797c9c3663e"}, "https://github.com/deepmind/xquad/raw/master/xquad.th.json": {"num_bytes": 1809143, "checksum": "5cdda11d0e1e075f7872abf4e6ae830388ce7f617964d542308e9ae4257e0f43"}, "https://github.com/deepmind/xquad/raw/master/xquad.tr.json": {"num_bytes": 729506, "checksum": "92179a564774b7696100d144c1e10870d0a966b6fccbdd254a65b9d2ab1971cc"}, "https://github.com/deepmind/xquad/raw/master/xquad.ru.json": {"num_bytes": 1896368, "checksum": "208d5b1aa154c52b1b5c5eda16281e455e8fd198cdb9af3f469f0d6037d973bf"}, "https://github.com/deepmind/xquad/raw/master/xquad.ro.json": {"num_bytes": 661158, "checksum": "94f8d54e93ad2b7dd2e76dd921a666b5b8eff2d4379df9f179c42b445f09643f"}}, "download_size": 13962158, "post_processing_size": null, "dataset_size": 1477239, "size_in_bytes": 15439397}, "xquad.en": {"description": "XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering\nperformance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set\nof SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten languages: Spanish, German,\nGreek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel\nacross 12 languages.\n", "citation": "@article{Artetxe:etal:2019,\n author = {Mikel Artetxe and Sebastian Ruder and Dani Yogatama},\n title = {On the cross-lingual transferability of monolingual representations},\n journal = {CoRR},\n volume = {abs/1910.11856},\n year = {2019},\n archivePrefix = {arXiv},\n eprint = {1910.11856}\n}\n", "homepage": "https://github.com/deepmind/xquad", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "context": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "answers": {"feature": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "answer_start": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "task_templates": [{"task": "question-answering-extractive", "question_column": "question", "context_column": "context", "answers_column": "answers"}], "builder_name": "xquad", "config_name": "xquad.en", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"validation": {"name": "validation", "num_bytes": 1116123, "num_examples": 1190, "dataset_name": "xquad"}}, "download_checksums": {"https://github.com/deepmind/xquad/raw/master/xquad.ar.json": {"num_bytes": 1582988, "checksum": "abdabd7afed5c635d99cca0f3f0d0c9d9ed0bc77451e963c2e4e0638c29e486d"}, "https://github.com/deepmind/xquad/raw/master/xquad.de.json": {"num_bytes": 669810, "checksum": "990b5d746746ed65ed4702ea5f35f99ffa4e2f1c390c07d003642acd937916f9"}, "https://github.com/deepmind/xquad/raw/master/xquad.zh.json": {"num_bytes": 808652, "checksum": "691d0b3359bc6b8faa8de931dfdfe21d50a65861ae348e32a0d1a0190b0c8835"}, "https://github.com/deepmind/xquad/raw/master/xquad.vi.json": {"num_bytes": 911401, "checksum": "f619a1eb11fb42d3ab0834259e488a65f585447ef6154437bfb7199d85161a04"}, "https://github.com/deepmind/xquad/raw/master/xquad.en.json": {"num_bytes": 609383, "checksum": "e4c57d1c9143aaa1c5d265ba5987a65f4e69528d2a98f29d6e75019b10344f29"}, "https://github.com/deepmind/xquad/raw/master/xquad.es.json": {"num_bytes": 684322, "checksum": "dcbae93ec3a9f4b9e78fd834a171d6f96c1a875e10e15b7530b7e4ef4971e37e"}, "https://github.com/deepmind/xquad/raw/master/xquad.hi.json": {"num_bytes": 1680538, "checksum": "df2cce3532b37e9beb8979704b5c9a4bf874358f105395a298b89427b43b9d24"}, "https://github.com/deepmind/xquad/raw/master/xquad.el.json": {"num_bytes": 1918889, "checksum": "821cf0f88e73fa258fd2f548b19b6ec39f7025059e16f6f9fc8cd797c9c3663e"}, "https://github.com/deepmind/xquad/raw/master/xquad.th.json": {"num_bytes": 1809143, "checksum": "5cdda11d0e1e075f7872abf4e6ae830388ce7f617964d542308e9ae4257e0f43"}, "https://github.com/deepmind/xquad/raw/master/xquad.tr.json": {"num_bytes": 729506, "checksum": "92179a564774b7696100d144c1e10870d0a966b6fccbdd254a65b9d2ab1971cc"}, "https://github.com/deepmind/xquad/raw/master/xquad.ru.json": {"num_bytes": 1896368, "checksum": "208d5b1aa154c52b1b5c5eda16281e455e8fd198cdb9af3f469f0d6037d973bf"}, "https://github.com/deepmind/xquad/raw/master/xquad.ro.json": {"num_bytes": 661158, "checksum": "94f8d54e93ad2b7dd2e76dd921a666b5b8eff2d4379df9f179c42b445f09643f"}}, "download_size": 13962158, "post_processing_size": null, "dataset_size": 1116123, "size_in_bytes": 15078281}, "xquad.es": {"description": "XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering\nperformance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set\nof SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten languages: Spanish, German,\nGreek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel\nacross 12 languages.\n", "citation": "@article{Artetxe:etal:2019,\n author = {Mikel Artetxe and Sebastian Ruder and Dani Yogatama},\n title = {On the cross-lingual transferability of monolingual representations},\n journal = {CoRR},\n volume = {abs/1910.11856},\n year = {2019},\n archivePrefix = {arXiv},\n eprint = {1910.11856}\n}\n", "homepage": "https://github.com/deepmind/xquad", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "context": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "answers": {"feature": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "answer_start": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "task_templates": [{"task": "question-answering-extractive", "question_column": "question", "context_column": "context", "answers_column": "answers"}], "builder_name": "xquad", "config_name": "xquad.es", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"validation": {"name": "validation", "num_bytes": 1273499, "num_examples": 1190, "dataset_name": "xquad"}}, "download_checksums": {"https://github.com/deepmind/xquad/raw/master/xquad.ar.json": {"num_bytes": 1582988, "checksum": "abdabd7afed5c635d99cca0f3f0d0c9d9ed0bc77451e963c2e4e0638c29e486d"}, "https://github.com/deepmind/xquad/raw/master/xquad.de.json": {"num_bytes": 669810, "checksum": "990b5d746746ed65ed4702ea5f35f99ffa4e2f1c390c07d003642acd937916f9"}, "https://github.com/deepmind/xquad/raw/master/xquad.zh.json": {"num_bytes": 808652, "checksum": "691d0b3359bc6b8faa8de931dfdfe21d50a65861ae348e32a0d1a0190b0c8835"}, "https://github.com/deepmind/xquad/raw/master/xquad.vi.json": {"num_bytes": 911401, "checksum": "f619a1eb11fb42d3ab0834259e488a65f585447ef6154437bfb7199d85161a04"}, "https://github.com/deepmind/xquad/raw/master/xquad.en.json": {"num_bytes": 609383, "checksum": "e4c57d1c9143aaa1c5d265ba5987a65f4e69528d2a98f29d6e75019b10344f29"}, "https://github.com/deepmind/xquad/raw/master/xquad.es.json": {"num_bytes": 684322, "checksum": "dcbae93ec3a9f4b9e78fd834a171d6f96c1a875e10e15b7530b7e4ef4971e37e"}, "https://github.com/deepmind/xquad/raw/master/xquad.hi.json": {"num_bytes": 1680538, "checksum": "df2cce3532b37e9beb8979704b5c9a4bf874358f105395a298b89427b43b9d24"}, "https://github.com/deepmind/xquad/raw/master/xquad.el.json": {"num_bytes": 1918889, "checksum": "821cf0f88e73fa258fd2f548b19b6ec39f7025059e16f6f9fc8cd797c9c3663e"}, "https://github.com/deepmind/xquad/raw/master/xquad.th.json": {"num_bytes": 1809143, "checksum": "5cdda11d0e1e075f7872abf4e6ae830388ce7f617964d542308e9ae4257e0f43"}, "https://github.com/deepmind/xquad/raw/master/xquad.tr.json": {"num_bytes": 729506, "checksum": "92179a564774b7696100d144c1e10870d0a966b6fccbdd254a65b9d2ab1971cc"}, "https://github.com/deepmind/xquad/raw/master/xquad.ru.json": {"num_bytes": 1896368, "checksum": "208d5b1aa154c52b1b5c5eda16281e455e8fd198cdb9af3f469f0d6037d973bf"}, "https://github.com/deepmind/xquad/raw/master/xquad.ro.json": {"num_bytes": 661158, "checksum": "94f8d54e93ad2b7dd2e76dd921a666b5b8eff2d4379df9f179c42b445f09643f"}}, "download_size": 13962158, "post_processing_size": null, "dataset_size": 1273499, "size_in_bytes": 15235657}, "xquad.hi": {"description": "XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering\nperformance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set\nof SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten languages: Spanish, German,\nGreek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel\nacross 12 languages.\n", "citation": "@article{Artetxe:etal:2019,\n author = {Mikel Artetxe and Sebastian Ruder and Dani Yogatama},\n title = {On the cross-lingual transferability of monolingual representations},\n journal = {CoRR},\n volume = {abs/1910.11856},\n year = {2019},\n archivePrefix = {arXiv},\n eprint = {1910.11856}\n}\n", "homepage": "https://github.com/deepmind/xquad", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "context": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "answers": {"feature": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "answer_start": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "task_templates": [{"task": "question-answering-extractive", "question_column": "question", "context_column": "context", "answers_column": "answers"}], "builder_name": "xquad", "config_name": "xquad.hi", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"validation": {"name": "validation", "num_bytes": 2682975, "num_examples": 1190, "dataset_name": "xquad"}}, "download_checksums": {"https://github.com/deepmind/xquad/raw/master/xquad.ar.json": {"num_bytes": 1582988, "checksum": "abdabd7afed5c635d99cca0f3f0d0c9d9ed0bc77451e963c2e4e0638c29e486d"}, "https://github.com/deepmind/xquad/raw/master/xquad.de.json": {"num_bytes": 669810, "checksum": "990b5d746746ed65ed4702ea5f35f99ffa4e2f1c390c07d003642acd937916f9"}, "https://github.com/deepmind/xquad/raw/master/xquad.zh.json": {"num_bytes": 808652, "checksum": "691d0b3359bc6b8faa8de931dfdfe21d50a65861ae348e32a0d1a0190b0c8835"}, "https://github.com/deepmind/xquad/raw/master/xquad.vi.json": {"num_bytes": 911401, "checksum": "f619a1eb11fb42d3ab0834259e488a65f585447ef6154437bfb7199d85161a04"}, "https://github.com/deepmind/xquad/raw/master/xquad.en.json": {"num_bytes": 609383, "checksum": "e4c57d1c9143aaa1c5d265ba5987a65f4e69528d2a98f29d6e75019b10344f29"}, "https://github.com/deepmind/xquad/raw/master/xquad.es.json": {"num_bytes": 684322, "checksum": "dcbae93ec3a9f4b9e78fd834a171d6f96c1a875e10e15b7530b7e4ef4971e37e"}, "https://github.com/deepmind/xquad/raw/master/xquad.hi.json": {"num_bytes": 1680538, "checksum": "df2cce3532b37e9beb8979704b5c9a4bf874358f105395a298b89427b43b9d24"}, "https://github.com/deepmind/xquad/raw/master/xquad.el.json": {"num_bytes": 1918889, "checksum": "821cf0f88e73fa258fd2f548b19b6ec39f7025059e16f6f9fc8cd797c9c3663e"}, "https://github.com/deepmind/xquad/raw/master/xquad.th.json": {"num_bytes": 1809143, "checksum": "5cdda11d0e1e075f7872abf4e6ae830388ce7f617964d542308e9ae4257e0f43"}, "https://github.com/deepmind/xquad/raw/master/xquad.tr.json": {"num_bytes": 729506, "checksum": "92179a564774b7696100d144c1e10870d0a966b6fccbdd254a65b9d2ab1971cc"}, "https://github.com/deepmind/xquad/raw/master/xquad.ru.json": {"num_bytes": 1896368, "checksum": "208d5b1aa154c52b1b5c5eda16281e455e8fd198cdb9af3f469f0d6037d973bf"}, "https://github.com/deepmind/xquad/raw/master/xquad.ro.json": {"num_bytes": 661158, "checksum": "94f8d54e93ad2b7dd2e76dd921a666b5b8eff2d4379df9f179c42b445f09643f"}}, "download_size": 13962158, "post_processing_size": null, "dataset_size": 2682975, "size_in_bytes": 16645133}, "xquad.el": {"description": "XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering\nperformance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set\nof SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten languages: Spanish, German,\nGreek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel\nacross 12 languages.\n", "citation": "@article{Artetxe:etal:2019,\n author = {Mikel Artetxe and Sebastian Ruder and Dani Yogatama},\n title = {On the cross-lingual transferability of monolingual representations},\n journal = {CoRR},\n volume = {abs/1910.11856},\n year = {2019},\n archivePrefix = {arXiv},\n eprint = {1910.11856}\n}\n", "homepage": "https://github.com/deepmind/xquad", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "context": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "answers": {"feature": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "answer_start": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "task_templates": [{"task": "question-answering-extractive", "question_column": "question", "context_column": "context", "answers_column": "answers"}], "builder_name": "xquad", "config_name": "xquad.el", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"validation": {"name": "validation", "num_bytes": 2206690, "num_examples": 1190, "dataset_name": "xquad"}}, "download_checksums": {"https://github.com/deepmind/xquad/raw/master/xquad.ar.json": {"num_bytes": 1582988, "checksum": "abdabd7afed5c635d99cca0f3f0d0c9d9ed0bc77451e963c2e4e0638c29e486d"}, "https://github.com/deepmind/xquad/raw/master/xquad.de.json": {"num_bytes": 669810, "checksum": "990b5d746746ed65ed4702ea5f35f99ffa4e2f1c390c07d003642acd937916f9"}, "https://github.com/deepmind/xquad/raw/master/xquad.zh.json": {"num_bytes": 808652, "checksum": "691d0b3359bc6b8faa8de931dfdfe21d50a65861ae348e32a0d1a0190b0c8835"}, "https://github.com/deepmind/xquad/raw/master/xquad.vi.json": {"num_bytes": 911401, "checksum": "f619a1eb11fb42d3ab0834259e488a65f585447ef6154437bfb7199d85161a04"}, "https://github.com/deepmind/xquad/raw/master/xquad.en.json": {"num_bytes": 609383, "checksum": "e4c57d1c9143aaa1c5d265ba5987a65f4e69528d2a98f29d6e75019b10344f29"}, "https://github.com/deepmind/xquad/raw/master/xquad.es.json": {"num_bytes": 684322, "checksum": "dcbae93ec3a9f4b9e78fd834a171d6f96c1a875e10e15b7530b7e4ef4971e37e"}, "https://github.com/deepmind/xquad/raw/master/xquad.hi.json": {"num_bytes": 1680538, "checksum": "df2cce3532b37e9beb8979704b5c9a4bf874358f105395a298b89427b43b9d24"}, "https://github.com/deepmind/xquad/raw/master/xquad.el.json": {"num_bytes": 1918889, "checksum": "821cf0f88e73fa258fd2f548b19b6ec39f7025059e16f6f9fc8cd797c9c3663e"}, "https://github.com/deepmind/xquad/raw/master/xquad.th.json": {"num_bytes": 1809143, "checksum": "5cdda11d0e1e075f7872abf4e6ae830388ce7f617964d542308e9ae4257e0f43"}, "https://github.com/deepmind/xquad/raw/master/xquad.tr.json": {"num_bytes": 729506, "checksum": "92179a564774b7696100d144c1e10870d0a966b6fccbdd254a65b9d2ab1971cc"}, "https://github.com/deepmind/xquad/raw/master/xquad.ru.json": {"num_bytes": 1896368, "checksum": "208d5b1aa154c52b1b5c5eda16281e455e8fd198cdb9af3f469f0d6037d973bf"}, "https://github.com/deepmind/xquad/raw/master/xquad.ro.json": {"num_bytes": 661158, "checksum": "94f8d54e93ad2b7dd2e76dd921a666b5b8eff2d4379df9f179c42b445f09643f"}}, "download_size": 13962158, "post_processing_size": null, "dataset_size": 2206690, "size_in_bytes": 16168848}, "xquad.th": {"description": "XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering\nperformance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set\nof SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten languages: Spanish, German,\nGreek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel\nacross 12 languages.\n", "citation": "@article{Artetxe:etal:2019,\n author = {Mikel Artetxe and Sebastian Ruder and Dani Yogatama},\n title = {On the cross-lingual transferability of monolingual representations},\n journal = {CoRR},\n volume = {abs/1910.11856},\n year = {2019},\n archivePrefix = {arXiv},\n eprint = {1910.11856}\n}\n", "homepage": "https://github.com/deepmind/xquad", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "context": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "answers": {"feature": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "answer_start": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "task_templates": [{"task": "question-answering-extractive", "question_column": "question", "context_column": "context", "answers_column": "answers"}], "builder_name": "xquad", "config_name": "xquad.th", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"validation": {"name": "validation", "num_bytes": 2854959, "num_examples": 1190, "dataset_name": "xquad"}}, "download_checksums": {"https://github.com/deepmind/xquad/raw/master/xquad.ar.json": {"num_bytes": 1582988, "checksum": "abdabd7afed5c635d99cca0f3f0d0c9d9ed0bc77451e963c2e4e0638c29e486d"}, "https://github.com/deepmind/xquad/raw/master/xquad.de.json": {"num_bytes": 669810, "checksum": "990b5d746746ed65ed4702ea5f35f99ffa4e2f1c390c07d003642acd937916f9"}, "https://github.com/deepmind/xquad/raw/master/xquad.zh.json": {"num_bytes": 808652, "checksum": "691d0b3359bc6b8faa8de931dfdfe21d50a65861ae348e32a0d1a0190b0c8835"}, "https://github.com/deepmind/xquad/raw/master/xquad.vi.json": {"num_bytes": 911401, "checksum": "f619a1eb11fb42d3ab0834259e488a65f585447ef6154437bfb7199d85161a04"}, "https://github.com/deepmind/xquad/raw/master/xquad.en.json": {"num_bytes": 609383, "checksum": "e4c57d1c9143aaa1c5d265ba5987a65f4e69528d2a98f29d6e75019b10344f29"}, "https://github.com/deepmind/xquad/raw/master/xquad.es.json": {"num_bytes": 684322, "checksum": "dcbae93ec3a9f4b9e78fd834a171d6f96c1a875e10e15b7530b7e4ef4971e37e"}, "https://github.com/deepmind/xquad/raw/master/xquad.hi.json": {"num_bytes": 1680538, "checksum": "df2cce3532b37e9beb8979704b5c9a4bf874358f105395a298b89427b43b9d24"}, "https://github.com/deepmind/xquad/raw/master/xquad.el.json": {"num_bytes": 1918889, "checksum": "821cf0f88e73fa258fd2f548b19b6ec39f7025059e16f6f9fc8cd797c9c3663e"}, "https://github.com/deepmind/xquad/raw/master/xquad.th.json": {"num_bytes": 1809143, "checksum": "5cdda11d0e1e075f7872abf4e6ae830388ce7f617964d542308e9ae4257e0f43"}, "https://github.com/deepmind/xquad/raw/master/xquad.tr.json": {"num_bytes": 729506, "checksum": "92179a564774b7696100d144c1e10870d0a966b6fccbdd254a65b9d2ab1971cc"}, "https://github.com/deepmind/xquad/raw/master/xquad.ru.json": {"num_bytes": 1896368, "checksum": "208d5b1aa154c52b1b5c5eda16281e455e8fd198cdb9af3f469f0d6037d973bf"}, "https://github.com/deepmind/xquad/raw/master/xquad.ro.json": {"num_bytes": 661158, "checksum": "94f8d54e93ad2b7dd2e76dd921a666b5b8eff2d4379df9f179c42b445f09643f"}}, "download_size": 13962158, "post_processing_size": null, "dataset_size": 2854959, "size_in_bytes": 16817117}, "xquad.tr": {"description": "XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering\nperformance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set\nof SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten languages: Spanish, German,\nGreek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel\nacross 12 languages.\n", "citation": "@article{Artetxe:etal:2019,\n author = {Mikel Artetxe and Sebastian Ruder and Dani Yogatama},\n title = {On the cross-lingual transferability of monolingual representations},\n journal = {CoRR},\n volume = {abs/1910.11856},\n year = {2019},\n archivePrefix = {arXiv},\n eprint = {1910.11856}\n}\n", "homepage": "https://github.com/deepmind/xquad", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "context": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "answers": {"feature": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "answer_start": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "task_templates": [{"task": "question-answering-extractive", "question_column": "question", "context_column": "context", "answers_column": "answers"}], "builder_name": "xquad", "config_name": "xquad.tr", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"validation": {"name": "validation", "num_bytes": 1210763, "num_examples": 1190, "dataset_name": "xquad"}}, "download_checksums": {"https://github.com/deepmind/xquad/raw/master/xquad.ar.json": {"num_bytes": 1582988, "checksum": "abdabd7afed5c635d99cca0f3f0d0c9d9ed0bc77451e963c2e4e0638c29e486d"}, "https://github.com/deepmind/xquad/raw/master/xquad.de.json": {"num_bytes": 669810, "checksum": "990b5d746746ed65ed4702ea5f35f99ffa4e2f1c390c07d003642acd937916f9"}, "https://github.com/deepmind/xquad/raw/master/xquad.zh.json": {"num_bytes": 808652, "checksum": "691d0b3359bc6b8faa8de931dfdfe21d50a65861ae348e32a0d1a0190b0c8835"}, "https://github.com/deepmind/xquad/raw/master/xquad.vi.json": {"num_bytes": 911401, "checksum": "f619a1eb11fb42d3ab0834259e488a65f585447ef6154437bfb7199d85161a04"}, "https://github.com/deepmind/xquad/raw/master/xquad.en.json": {"num_bytes": 609383, "checksum": "e4c57d1c9143aaa1c5d265ba5987a65f4e69528d2a98f29d6e75019b10344f29"}, "https://github.com/deepmind/xquad/raw/master/xquad.es.json": {"num_bytes": 684322, "checksum": "dcbae93ec3a9f4b9e78fd834a171d6f96c1a875e10e15b7530b7e4ef4971e37e"}, "https://github.com/deepmind/xquad/raw/master/xquad.hi.json": {"num_bytes": 1680538, "checksum": "df2cce3532b37e9beb8979704b5c9a4bf874358f105395a298b89427b43b9d24"}, "https://github.com/deepmind/xquad/raw/master/xquad.el.json": {"num_bytes": 1918889, "checksum": "821cf0f88e73fa258fd2f548b19b6ec39f7025059e16f6f9fc8cd797c9c3663e"}, "https://github.com/deepmind/xquad/raw/master/xquad.th.json": {"num_bytes": 1809143, "checksum": "5cdda11d0e1e075f7872abf4e6ae830388ce7f617964d542308e9ae4257e0f43"}, "https://github.com/deepmind/xquad/raw/master/xquad.tr.json": {"num_bytes": 729506, "checksum": "92179a564774b7696100d144c1e10870d0a966b6fccbdd254a65b9d2ab1971cc"}, "https://github.com/deepmind/xquad/raw/master/xquad.ru.json": {"num_bytes": 1896368, "checksum": "208d5b1aa154c52b1b5c5eda16281e455e8fd198cdb9af3f469f0d6037d973bf"}, "https://github.com/deepmind/xquad/raw/master/xquad.ro.json": {"num_bytes": 661158, "checksum": "94f8d54e93ad2b7dd2e76dd921a666b5b8eff2d4379df9f179c42b445f09643f"}}, "download_size": 13962158, "post_processing_size": null, "dataset_size": 1210763, "size_in_bytes": 15172921}, "xquad.ru": {"description": "XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering\nperformance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set\nof SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten languages: Spanish, German,\nGreek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel\nacross 12 languages.\n", "citation": "@article{Artetxe:etal:2019,\n author = {Mikel Artetxe and Sebastian Ruder and Dani Yogatama},\n title = {On the cross-lingual transferability of monolingual representations},\n journal = {CoRR},\n volume = {abs/1910.11856},\n year = {2019},\n archivePrefix = {arXiv},\n eprint = {1910.11856}\n}\n", "homepage": "https://github.com/deepmind/xquad", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "context": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "answers": {"feature": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "answer_start": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "task_templates": [{"task": "question-answering-extractive", "question_column": "question", "context_column": "context", "answers_column": "answers"}], "builder_name": "xquad", "config_name": "xquad.ru", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"validation": {"name": "validation", "num_bytes": 2136990, "num_examples": 1190, "dataset_name": "xquad"}}, "download_checksums": {"https://github.com/deepmind/xquad/raw/master/xquad.ar.json": {"num_bytes": 1582988, "checksum": "abdabd7afed5c635d99cca0f3f0d0c9d9ed0bc77451e963c2e4e0638c29e486d"}, "https://github.com/deepmind/xquad/raw/master/xquad.de.json": {"num_bytes": 669810, "checksum": "990b5d746746ed65ed4702ea5f35f99ffa4e2f1c390c07d003642acd937916f9"}, "https://github.com/deepmind/xquad/raw/master/xquad.zh.json": {"num_bytes": 808652, "checksum": "691d0b3359bc6b8faa8de931dfdfe21d50a65861ae348e32a0d1a0190b0c8835"}, "https://github.com/deepmind/xquad/raw/master/xquad.vi.json": {"num_bytes": 911401, "checksum": "f619a1eb11fb42d3ab0834259e488a65f585447ef6154437bfb7199d85161a04"}, "https://github.com/deepmind/xquad/raw/master/xquad.en.json": {"num_bytes": 609383, "checksum": "e4c57d1c9143aaa1c5d265ba5987a65f4e69528d2a98f29d6e75019b10344f29"}, "https://github.com/deepmind/xquad/raw/master/xquad.es.json": {"num_bytes": 684322, "checksum": "dcbae93ec3a9f4b9e78fd834a171d6f96c1a875e10e15b7530b7e4ef4971e37e"}, "https://github.com/deepmind/xquad/raw/master/xquad.hi.json": {"num_bytes": 1680538, "checksum": "df2cce3532b37e9beb8979704b5c9a4bf874358f105395a298b89427b43b9d24"}, "https://github.com/deepmind/xquad/raw/master/xquad.el.json": {"num_bytes": 1918889, "checksum": "821cf0f88e73fa258fd2f548b19b6ec39f7025059e16f6f9fc8cd797c9c3663e"}, "https://github.com/deepmind/xquad/raw/master/xquad.th.json": {"num_bytes": 1809143, "checksum": "5cdda11d0e1e075f7872abf4e6ae830388ce7f617964d542308e9ae4257e0f43"}, "https://github.com/deepmind/xquad/raw/master/xquad.tr.json": {"num_bytes": 729506, "checksum": "92179a564774b7696100d144c1e10870d0a966b6fccbdd254a65b9d2ab1971cc"}, "https://github.com/deepmind/xquad/raw/master/xquad.ru.json": {"num_bytes": 1896368, "checksum": "208d5b1aa154c52b1b5c5eda16281e455e8fd198cdb9af3f469f0d6037d973bf"}, "https://github.com/deepmind/xquad/raw/master/xquad.ro.json": {"num_bytes": 661158, "checksum": "94f8d54e93ad2b7dd2e76dd921a666b5b8eff2d4379df9f179c42b445f09643f"}}, "download_size": 13962158, "post_processing_size": null, "dataset_size": 2136990, "size_in_bytes": 16099148}, "xquad.ro": {"description": "XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering\nperformance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set\nof SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten languages: Spanish, German,\nGreek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel\nacross 12 languages.\n", "citation": "@article{Artetxe:etal:2019,\n author = {Mikel Artetxe and Sebastian Ruder and Dani Yogatama},\n title = {On the cross-lingual transferability of monolingual representations},\n journal = {CoRR},\n volume = {abs/1910.11856},\n year = {2019},\n archivePrefix = {arXiv},\n eprint = {1910.11856}\n}\n", "homepage": "https://github.com/deepmind/xquad", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "context": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "answers": {"feature": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "answer_start": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "task_templates": [{"task": "question-answering-extractive", "question_column": "question", "context_column": "context", "answers_column": "answers"}], "builder_name": "xquad", "config_name": "xquad.ro", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"validation": {"name": "validation", "num_bytes": 1299450, "num_examples": 1190, "dataset_name": "xquad"}}, "download_checksums": {"https://github.com/deepmind/xquad/raw/master/xquad.ar.json": {"num_bytes": 1582988, "checksum": "abdabd7afed5c635d99cca0f3f0d0c9d9ed0bc77451e963c2e4e0638c29e486d"}, "https://github.com/deepmind/xquad/raw/master/xquad.de.json": {"num_bytes": 669810, "checksum": "990b5d746746ed65ed4702ea5f35f99ffa4e2f1c390c07d003642acd937916f9"}, "https://github.com/deepmind/xquad/raw/master/xquad.zh.json": {"num_bytes": 808652, "checksum": "691d0b3359bc6b8faa8de931dfdfe21d50a65861ae348e32a0d1a0190b0c8835"}, "https://github.com/deepmind/xquad/raw/master/xquad.vi.json": {"num_bytes": 911401, "checksum": "f619a1eb11fb42d3ab0834259e488a65f585447ef6154437bfb7199d85161a04"}, "https://github.com/deepmind/xquad/raw/master/xquad.en.json": {"num_bytes": 609383, "checksum": "e4c57d1c9143aaa1c5d265ba5987a65f4e69528d2a98f29d6e75019b10344f29"}, "https://github.com/deepmind/xquad/raw/master/xquad.es.json": {"num_bytes": 684322, "checksum": "dcbae93ec3a9f4b9e78fd834a171d6f96c1a875e10e15b7530b7e4ef4971e37e"}, "https://github.com/deepmind/xquad/raw/master/xquad.hi.json": {"num_bytes": 1680538, "checksum": "df2cce3532b37e9beb8979704b5c9a4bf874358f105395a298b89427b43b9d24"}, "https://github.com/deepmind/xquad/raw/master/xquad.el.json": {"num_bytes": 1918889, "checksum": "821cf0f88e73fa258fd2f548b19b6ec39f7025059e16f6f9fc8cd797c9c3663e"}, "https://github.com/deepmind/xquad/raw/master/xquad.th.json": {"num_bytes": 1809143, "checksum": "5cdda11d0e1e075f7872abf4e6ae830388ce7f617964d542308e9ae4257e0f43"}, "https://github.com/deepmind/xquad/raw/master/xquad.tr.json": {"num_bytes": 729506, "checksum": "92179a564774b7696100d144c1e10870d0a966b6fccbdd254a65b9d2ab1971cc"}, "https://github.com/deepmind/xquad/raw/master/xquad.ru.json": {"num_bytes": 1896368, "checksum": "208d5b1aa154c52b1b5c5eda16281e455e8fd198cdb9af3f469f0d6037d973bf"}, "https://github.com/deepmind/xquad/raw/master/xquad.ro.json": {"num_bytes": 661158, "checksum": "94f8d54e93ad2b7dd2e76dd921a666b5b8eff2d4379df9f179c42b445f09643f"}}, "download_size": 13962158, "post_processing_size": null, "dataset_size": 1299450, "size_in_bytes": 15261608}}
 
xquad.ar/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5af9e1a441be52d406aa2cf095aadd319fb4173463f363ee79ff3049ce037d0f
3
+ size 263002
xquad.de/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b52b5c144d88c0dc44096c26ecbc5a9b799416456a86e5aaa272171fb529c145
3
+ size 241957
xquad.el/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:26722d97e8e1d5cfa7b7c96065f2d231d84fc2a695e3feb797447955bac87e14
3
+ size 324379
xquad.en/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e769714c0b750622a2a70e3b5250f6040ef8578e859dff0b9faf4612287e0083
3
+ size 212372
xquad.es/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:69b6ad587e2ebcb0b00105f6736661154f9b1ce95eaa581bcac240d6edfc11ce
3
+ size 236874
xquad.hi/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e6c7be8500401994d4812db12380239b4ca844c21b650ebdd393d4aba045177e
3
+ size 322083
xquad.py DELETED
@@ -1,131 +0,0 @@
1
- """TODO(xquad): Add a description here."""
2
-
3
-
4
- import json
5
-
6
- import datasets
7
- from datasets.tasks import QuestionAnsweringExtractive
8
-
9
-
10
- _CITATION = """\
11
- @article{Artetxe:etal:2019,
12
- author = {Mikel Artetxe and Sebastian Ruder and Dani Yogatama},
13
- title = {On the cross-lingual transferability of monolingual representations},
14
- journal = {CoRR},
15
- volume = {abs/1910.11856},
16
- year = {2019},
17
- archivePrefix = {arXiv},
18
- eprint = {1910.11856}
19
- }
20
- """
21
-
22
- _DESCRIPTION = """\
23
- XQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering
24
- performance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set
25
- of SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten languages: Spanish, German,
26
- Greek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel
27
- across 12 languages.
28
- """
29
-
30
- _URL = "https://github.com/deepmind/xquad/raw/master/"
31
- _LANG = ["ar", "de", "zh", "vi", "en", "es", "hi", "el", "th", "tr", "ru", "ro"]
32
-
33
-
34
- class XquadConfig(datasets.BuilderConfig):
35
-
36
- """BuilderConfig for Xquad"""
37
-
38
- def __init__(self, lang, **kwargs):
39
- """
40
-
41
- Args:
42
- lang: string, language for the input text
43
- **kwargs: keyword arguments forwarded to super.
44
- """
45
- super(XquadConfig, self).__init__(version=datasets.Version("1.0.0", ""), **kwargs)
46
- self.lang = lang
47
-
48
-
49
- class Xquad(datasets.GeneratorBasedBuilder):
50
- """TODO(xquad): Short description of my dataset."""
51
-
52
- # TODO(xquad): Set up version.
53
- VERSION = datasets.Version("1.0.0")
54
- BUILDER_CONFIGS = [XquadConfig(name=f"xquad.{lang}", description=_DESCRIPTION, lang=lang) for lang in _LANG]
55
-
56
- def _info(self):
57
- # TODO(xquad): Specifies the datasets.DatasetInfo object
58
- return datasets.DatasetInfo(
59
- # This is the description that will appear on the datasets page.
60
- description=_DESCRIPTION,
61
- # datasets.features.FeatureConnectors
62
- features=datasets.Features(
63
- {
64
- "id": datasets.Value("string"),
65
- "context": datasets.Value("string"),
66
- "question": datasets.Value("string"),
67
- "answers": datasets.features.Sequence(
68
- {
69
- "text": datasets.Value("string"),
70
- "answer_start": datasets.Value("int32"),
71
- }
72
- ),
73
- # These are the features of your dataset like images, labels ...
74
- }
75
- ),
76
- # If there's a common (input, target) tuple from the features,
77
- # specify them here. They'll be used if as_supervised=True in
78
- # builder.as_dataset.
79
- supervised_keys=None,
80
- # Homepage of the dataset for documentation
81
- homepage="https://github.com/deepmind/xquad",
82
- citation=_CITATION,
83
- task_templates=[
84
- QuestionAnsweringExtractive(
85
- question_column="question", context_column="context", answers_column="answers"
86
- )
87
- ],
88
- )
89
-
90
- def _split_generators(self, dl_manager):
91
- """Returns SplitGenerators."""
92
- # TODO(xquad): Downloads the data and defines the splits
93
- # dl_manager is a datasets.download.DownloadManager that can be used to
94
- # download and extract URLs
95
- urls_to_download = {lang: _URL + f"xquad.{lang}.json" for lang in _LANG}
96
- downloaded_files = dl_manager.download_and_extract(urls_to_download)
97
-
98
- return [
99
- datasets.SplitGenerator(
100
- name=datasets.Split.VALIDATION,
101
- # These kwargs will be passed to _generate_examples
102
- gen_kwargs={"filepath": downloaded_files[self.config.lang]},
103
- ),
104
- ]
105
-
106
- def _generate_examples(self, filepath):
107
- """Yields examples."""
108
- # TODO(xquad): Yields (key, example) tuples from the dataset
109
- with open(filepath, encoding="utf-8") as f:
110
- xquad = json.load(f)
111
- id_ = 0
112
- for article in xquad["data"]:
113
- for paragraph in article["paragraphs"]:
114
- context = paragraph["context"].strip()
115
- for qa in paragraph["qas"]:
116
- question = qa["question"].strip()
117
- answer_starts = [answer["answer_start"] for answer in qa["answers"]]
118
- answers = [answer["text"].strip() for answer in qa["answers"]]
119
-
120
- # Features currently used are "context", "question", and "answers".
121
- # Others are extracted here for the ease of future expansions.
122
- yield id_, {
123
- "context": context,
124
- "question": question,
125
- "id": qa["id"],
126
- "answers": {
127
- "answer_start": answer_starts,
128
- "text": answers,
129
- },
130
- }
131
- id_ += 1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
xquad.ro/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3ed040cad2980e5d107dfa392a7b20be0352c85af52e1b1919fa3045ee665b95
3
+ size 244467
xquad.ru/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:90034e23fcdf29dbfc589951a160240cbf8e82afcb10995cfaf7e4ea98d10232
3
+ size 321728
xquad.th/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8fe604323dc84933b42c64f932ab5b3639de375750a20cb7c0937aced0f4bfe0
3
+ size 337307
xquad.tr/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6f788cc02246404a76bb890c3fc7beede9b6d4933f0fd23e841de95f7db3f5e9
3
+ size 228364
xquad.vi/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:52442ff8696ea0efa6595c67f2cdbd1a366e12cdb0dcaabbe0855bc4b5724ced
3
+ size 237644
xquad.zh/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6d9033e82b2a1dbf67b66025e84fab9a44e13dc91401e4914a9529db784146f6
3
+ size 205768