lhoestq HF staff commited on
Commit
a4ac6a2
1 Parent(s): 9e12063

add dataset_info in dataset metadata

Browse files
Files changed (1) hide show
  1. README.md +321 -1
README.md CHANGED
@@ -28,6 +28,326 @@ task_ids:
28
  - word-sense-disambiguation
29
  - coreference-resolution
30
  - extractive-qa
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
31
  ---
32
 
33
  # Dataset Card for "super_glue"
@@ -290,4 +610,4 @@ get the correct citation for each contained dataset.
290
 
291
  ### Contributions
292
 
293
- Thanks to [@thomwolf](https://github.com/thomwolf), [@lewtun](https://github.com/lewtun), [@patrickvonplaten](https://github.com/patrickvonplaten) for adding this dataset.
 
28
  - word-sense-disambiguation
29
  - coreference-resolution
30
  - extractive-qa
31
+ dataset_info:
32
+ - config_name: boolq
33
+ features:
34
+ - name: question
35
+ dtype: string
36
+ - name: passage
37
+ dtype: string
38
+ - name: idx
39
+ dtype: int32
40
+ - name: label
41
+ dtype:
42
+ class_label:
43
+ names:
44
+ 0: 'False'
45
+ 1: 'True'
46
+ splits:
47
+ - name: test
48
+ num_bytes: 2107997
49
+ num_examples: 3245
50
+ - name: train
51
+ num_bytes: 6179206
52
+ num_examples: 9427
53
+ - name: validation
54
+ num_bytes: 2118505
55
+ num_examples: 3270
56
+ download_size: 4118001
57
+ dataset_size: 10405708
58
+ - config_name: cb
59
+ features:
60
+ - name: premise
61
+ dtype: string
62
+ - name: hypothesis
63
+ dtype: string
64
+ - name: idx
65
+ dtype: int32
66
+ - name: label
67
+ dtype:
68
+ class_label:
69
+ names:
70
+ 0: entailment
71
+ 1: contradiction
72
+ 2: neutral
73
+ splits:
74
+ - name: test
75
+ num_bytes: 93660
76
+ num_examples: 250
77
+ - name: train
78
+ num_bytes: 87218
79
+ num_examples: 250
80
+ - name: validation
81
+ num_bytes: 21894
82
+ num_examples: 56
83
+ download_size: 75482
84
+ dataset_size: 202772
85
+ - config_name: copa
86
+ features:
87
+ - name: premise
88
+ dtype: string
89
+ - name: choice1
90
+ dtype: string
91
+ - name: choice2
92
+ dtype: string
93
+ - name: question
94
+ dtype: string
95
+ - name: idx
96
+ dtype: int32
97
+ - name: label
98
+ dtype:
99
+ class_label:
100
+ names:
101
+ 0: choice1
102
+ 1: choice2
103
+ splits:
104
+ - name: test
105
+ num_bytes: 60303
106
+ num_examples: 500
107
+ - name: train
108
+ num_bytes: 49599
109
+ num_examples: 400
110
+ - name: validation
111
+ num_bytes: 12586
112
+ num_examples: 100
113
+ download_size: 43986
114
+ dataset_size: 122488
115
+ - config_name: multirc
116
+ features:
117
+ - name: paragraph
118
+ dtype: string
119
+ - name: question
120
+ dtype: string
121
+ - name: answer
122
+ dtype: string
123
+ - name: idx
124
+ struct:
125
+ - name: paragraph
126
+ dtype: int32
127
+ - name: question
128
+ dtype: int32
129
+ - name: answer
130
+ dtype: int32
131
+ - name: label
132
+ dtype:
133
+ class_label:
134
+ names:
135
+ 0: 'False'
136
+ 1: 'True'
137
+ splits:
138
+ - name: test
139
+ num_bytes: 14996451
140
+ num_examples: 9693
141
+ - name: train
142
+ num_bytes: 46213579
143
+ num_examples: 27243
144
+ - name: validation
145
+ num_bytes: 7758918
146
+ num_examples: 4848
147
+ download_size: 1116225
148
+ dataset_size: 68968948
149
+ - config_name: record
150
+ features:
151
+ - name: passage
152
+ dtype: string
153
+ - name: query
154
+ dtype: string
155
+ - name: entities
156
+ sequence: string
157
+ - name: entity_spans
158
+ sequence:
159
+ - name: text
160
+ dtype: string
161
+ - name: start
162
+ dtype: int32
163
+ - name: end
164
+ dtype: int32
165
+ - name: answers
166
+ sequence: string
167
+ - name: idx
168
+ struct:
169
+ - name: passage
170
+ dtype: int32
171
+ - name: query
172
+ dtype: int32
173
+ splits:
174
+ - name: test
175
+ num_bytes: 17200575
176
+ num_examples: 10000
177
+ - name: train
178
+ num_bytes: 179232052
179
+ num_examples: 100730
180
+ - name: validation
181
+ num_bytes: 17479084
182
+ num_examples: 10000
183
+ download_size: 51757880
184
+ dataset_size: 213911711
185
+ - config_name: rte
186
+ features:
187
+ - name: premise
188
+ dtype: string
189
+ - name: hypothesis
190
+ dtype: string
191
+ - name: idx
192
+ dtype: int32
193
+ - name: label
194
+ dtype:
195
+ class_label:
196
+ names:
197
+ 0: entailment
198
+ 1: not_entailment
199
+ splits:
200
+ - name: test
201
+ num_bytes: 975799
202
+ num_examples: 3000
203
+ - name: train
204
+ num_bytes: 848745
205
+ num_examples: 2490
206
+ - name: validation
207
+ num_bytes: 90899
208
+ num_examples: 277
209
+ download_size: 750920
210
+ dataset_size: 1915443
211
+ - config_name: wic
212
+ features:
213
+ - name: word
214
+ dtype: string
215
+ - name: sentence1
216
+ dtype: string
217
+ - name: sentence2
218
+ dtype: string
219
+ - name: start1
220
+ dtype: int32
221
+ - name: start2
222
+ dtype: int32
223
+ - name: end1
224
+ dtype: int32
225
+ - name: end2
226
+ dtype: int32
227
+ - name: idx
228
+ dtype: int32
229
+ - name: label
230
+ dtype:
231
+ class_label:
232
+ names:
233
+ 0: 'False'
234
+ 1: 'True'
235
+ splits:
236
+ - name: test
237
+ num_bytes: 180593
238
+ num_examples: 1400
239
+ - name: train
240
+ num_bytes: 665183
241
+ num_examples: 5428
242
+ - name: validation
243
+ num_bytes: 82623
244
+ num_examples: 638
245
+ download_size: 396213
246
+ dataset_size: 928399
247
+ - config_name: wsc
248
+ features:
249
+ - name: text
250
+ dtype: string
251
+ - name: span1_index
252
+ dtype: int32
253
+ - name: span2_index
254
+ dtype: int32
255
+ - name: span1_text
256
+ dtype: string
257
+ - name: span2_text
258
+ dtype: string
259
+ - name: idx
260
+ dtype: int32
261
+ - name: label
262
+ dtype:
263
+ class_label:
264
+ names:
265
+ 0: 'False'
266
+ 1: 'True'
267
+ splits:
268
+ - name: test
269
+ num_bytes: 31572
270
+ num_examples: 146
271
+ - name: train
272
+ num_bytes: 89883
273
+ num_examples: 554
274
+ - name: validation
275
+ num_bytes: 21637
276
+ num_examples: 104
277
+ download_size: 32751
278
+ dataset_size: 143092
279
+ - config_name: wsc.fixed
280
+ features:
281
+ - name: text
282
+ dtype: string
283
+ - name: span1_index
284
+ dtype: int32
285
+ - name: span2_index
286
+ dtype: int32
287
+ - name: span1_text
288
+ dtype: string
289
+ - name: span2_text
290
+ dtype: string
291
+ - name: idx
292
+ dtype: int32
293
+ - name: label
294
+ dtype:
295
+ class_label:
296
+ names:
297
+ 0: 'False'
298
+ 1: 'True'
299
+ splits:
300
+ - name: test
301
+ num_bytes: 31568
302
+ num_examples: 146
303
+ - name: train
304
+ num_bytes: 89883
305
+ num_examples: 554
306
+ - name: validation
307
+ num_bytes: 21637
308
+ num_examples: 104
309
+ download_size: 32751
310
+ dataset_size: 143088
311
+ - config_name: axb
312
+ features:
313
+ - name: sentence1
314
+ dtype: string
315
+ - name: sentence2
316
+ dtype: string
317
+ - name: idx
318
+ dtype: int32
319
+ - name: label
320
+ dtype:
321
+ class_label:
322
+ names:
323
+ 0: entailment
324
+ 1: not_entailment
325
+ splits:
326
+ - name: test
327
+ num_bytes: 238392
328
+ num_examples: 1104
329
+ download_size: 33950
330
+ dataset_size: 238392
331
+ - config_name: axg
332
+ features:
333
+ - name: premise
334
+ dtype: string
335
+ - name: hypothesis
336
+ dtype: string
337
+ - name: idx
338
+ dtype: int32
339
+ - name: label
340
+ dtype:
341
+ class_label:
342
+ names:
343
+ 0: entailment
344
+ 1: not_entailment
345
+ splits:
346
+ - name: test
347
+ num_bytes: 53581
348
+ num_examples: 356
349
+ download_size: 10413
350
+ dataset_size: 53581
351
  ---
352
 
353
  # Dataset Card for "super_glue"
 
610
 
611
  ### Contributions
612
 
613
+ Thanks to [@thomwolf](https://github.com/thomwolf), [@lewtun](https://github.com/lewtun), [@patrickvonplaten](https://github.com/patrickvonplaten) for adding this dataset.