phucdev commited on
Commit
f3c058a
1 Parent(s): 95125d5

Upload dataloading script and README.md

Browse files
Files changed (2) hide show
  1. README.md +855 -0
  2. cross_ner.py +223 -0
README.md ADDED
@@ -0,0 +1,855 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ annotations_creators:
3
+ - expert-generated
4
+ language:
5
+ - en
6
+ language_creators:
7
+ - found
8
+ license: []
9
+ multilinguality:
10
+ - monolingual
11
+ pretty_name: CrossNER is a cross-domain dataset for named entity recognition
12
+ size_categories:
13
+ - 10K<n<100K
14
+ source_datasets:
15
+ - extended|conll2003
16
+ tags:
17
+ - cross domain
18
+ - ai
19
+ - news
20
+ - music
21
+ - literature
22
+ - politics
23
+ - science
24
+ task_categories:
25
+ - token-classification
26
+ task_ids:
27
+ - named-entity-recognition
28
+ dataset_info:
29
+ - config_name: ai
30
+ features:
31
+ - name: id
32
+ dtype: string
33
+ - name: tokens
34
+ sequence: string
35
+ - name: ner_tags
36
+ sequence:
37
+ class_label:
38
+ names:
39
+ '0': O
40
+ '1': B-academicjournal
41
+ '2': I-academicjournal
42
+ '3': B-album
43
+ '4': I-album
44
+ '5': B-algorithm
45
+ '6': I-algorithm
46
+ '7': B-astronomicalobject
47
+ '8': I-astronomicalobject
48
+ '9': B-award
49
+ '10': I-award
50
+ '11': B-band
51
+ '12': I-band
52
+ '13': B-book
53
+ '14': I-book
54
+ '15': B-chemicalcompound
55
+ '16': I-chemicalcompound
56
+ '17': B-chemicalelement
57
+ '18': I-chemicalelement
58
+ '19': B-conference
59
+ '20': I-conference
60
+ '21': B-country
61
+ '22': I-country
62
+ '23': B-discipline
63
+ '24': I-discipline
64
+ '25': B-election
65
+ '26': I-election
66
+ '27': B-enzyme
67
+ '28': I-enzyme
68
+ '29': B-event
69
+ '30': I-event
70
+ '31': B-field
71
+ '32': I-field
72
+ '33': B-literarygenre
73
+ '34': I-literarygenre
74
+ '35': B-location
75
+ '36': I-location
76
+ '37': B-magazine
77
+ '38': I-magazine
78
+ '39': B-metrics
79
+ '40': I-metrics
80
+ '41': B-misc
81
+ '42': I-misc
82
+ '43': B-musicalartist
83
+ '44': I-musicalartist
84
+ '45': B-musicalinstrument
85
+ '46': I-musicalinstrument
86
+ '47': B-musicgenre
87
+ '48': I-musicgenre
88
+ '49': B-organisation
89
+ '50': I-organisation
90
+ '51': B-person
91
+ '52': I-person
92
+ '53': B-poem
93
+ '54': I-poem
94
+ '55': B-politicalparty
95
+ '56': I-politicalparty
96
+ '57': B-politician
97
+ '58': I-politician
98
+ '59': B-product
99
+ '60': I-product
100
+ '61': B-programlang
101
+ '62': I-programlang
102
+ '63': B-protein
103
+ '64': I-protein
104
+ '65': B-researcher
105
+ '66': I-researcher
106
+ '67': B-scientist
107
+ '68': I-scientist
108
+ '69': B-song
109
+ '70': I-song
110
+ '71': B-task
111
+ '72': I-task
112
+ '73': B-theory
113
+ '74': I-theory
114
+ '75': B-university
115
+ '76': I-university
116
+ '77': B-writer
117
+ '78': I-writer
118
+ splits:
119
+ - name: train
120
+ num_bytes: 65080
121
+ num_examples: 100
122
+ - name: validation
123
+ num_bytes: 189453
124
+ num_examples: 350
125
+ - name: test
126
+ num_bytes: 225691
127
+ num_examples: 431
128
+ download_size: 289173
129
+ dataset_size: 480224
130
+ - config_name: literature
131
+ features:
132
+ - name: id
133
+ dtype: string
134
+ - name: tokens
135
+ sequence: string
136
+ - name: ner_tags
137
+ sequence:
138
+ class_label:
139
+ names:
140
+ '0': O
141
+ '1': B-academicjournal
142
+ '2': I-academicjournal
143
+ '3': B-album
144
+ '4': I-album
145
+ '5': B-algorithm
146
+ '6': I-algorithm
147
+ '7': B-astronomicalobject
148
+ '8': I-astronomicalobject
149
+ '9': B-award
150
+ '10': I-award
151
+ '11': B-band
152
+ '12': I-band
153
+ '13': B-book
154
+ '14': I-book
155
+ '15': B-chemicalcompound
156
+ '16': I-chemicalcompound
157
+ '17': B-chemicalelement
158
+ '18': I-chemicalelement
159
+ '19': B-conference
160
+ '20': I-conference
161
+ '21': B-country
162
+ '22': I-country
163
+ '23': B-discipline
164
+ '24': I-discipline
165
+ '25': B-election
166
+ '26': I-election
167
+ '27': B-enzyme
168
+ '28': I-enzyme
169
+ '29': B-event
170
+ '30': I-event
171
+ '31': B-field
172
+ '32': I-field
173
+ '33': B-literarygenre
174
+ '34': I-literarygenre
175
+ '35': B-location
176
+ '36': I-location
177
+ '37': B-magazine
178
+ '38': I-magazine
179
+ '39': B-metrics
180
+ '40': I-metrics
181
+ '41': B-misc
182
+ '42': I-misc
183
+ '43': B-musicalartist
184
+ '44': I-musicalartist
185
+ '45': B-musicalinstrument
186
+ '46': I-musicalinstrument
187
+ '47': B-musicgenre
188
+ '48': I-musicgenre
189
+ '49': B-organisation
190
+ '50': I-organisation
191
+ '51': B-person
192
+ '52': I-person
193
+ '53': B-poem
194
+ '54': I-poem
195
+ '55': B-politicalparty
196
+ '56': I-politicalparty
197
+ '57': B-politician
198
+ '58': I-politician
199
+ '59': B-product
200
+ '60': I-product
201
+ '61': B-programlang
202
+ '62': I-programlang
203
+ '63': B-protein
204
+ '64': I-protein
205
+ '65': B-researcher
206
+ '66': I-researcher
207
+ '67': B-scientist
208
+ '68': I-scientist
209
+ '69': B-song
210
+ '70': I-song
211
+ '71': B-task
212
+ '72': I-task
213
+ '73': B-theory
214
+ '74': I-theory
215
+ '75': B-university
216
+ '76': I-university
217
+ '77': B-writer
218
+ '78': I-writer
219
+ splits:
220
+ - name: train
221
+ num_bytes: 63181
222
+ num_examples: 100
223
+ - name: validation
224
+ num_bytes: 244076
225
+ num_examples: 400
226
+ - name: test
227
+ num_bytes: 270092
228
+ num_examples: 416
229
+ download_size: 334380
230
+ dataset_size: 577349
231
+ - config_name: music
232
+ features:
233
+ - name: id
234
+ dtype: string
235
+ - name: tokens
236
+ sequence: string
237
+ - name: ner_tags
238
+ sequence:
239
+ class_label:
240
+ names:
241
+ '0': O
242
+ '1': B-academicjournal
243
+ '2': I-academicjournal
244
+ '3': B-album
245
+ '4': I-album
246
+ '5': B-algorithm
247
+ '6': I-algorithm
248
+ '7': B-astronomicalobject
249
+ '8': I-astronomicalobject
250
+ '9': B-award
251
+ '10': I-award
252
+ '11': B-band
253
+ '12': I-band
254
+ '13': B-book
255
+ '14': I-book
256
+ '15': B-chemicalcompound
257
+ '16': I-chemicalcompound
258
+ '17': B-chemicalelement
259
+ '18': I-chemicalelement
260
+ '19': B-conference
261
+ '20': I-conference
262
+ '21': B-country
263
+ '22': I-country
264
+ '23': B-discipline
265
+ '24': I-discipline
266
+ '25': B-election
267
+ '26': I-election
268
+ '27': B-enzyme
269
+ '28': I-enzyme
270
+ '29': B-event
271
+ '30': I-event
272
+ '31': B-field
273
+ '32': I-field
274
+ '33': B-literarygenre
275
+ '34': I-literarygenre
276
+ '35': B-location
277
+ '36': I-location
278
+ '37': B-magazine
279
+ '38': I-magazine
280
+ '39': B-metrics
281
+ '40': I-metrics
282
+ '41': B-misc
283
+ '42': I-misc
284
+ '43': B-musicalartist
285
+ '44': I-musicalartist
286
+ '45': B-musicalinstrument
287
+ '46': I-musicalinstrument
288
+ '47': B-musicgenre
289
+ '48': I-musicgenre
290
+ '49': B-organisation
291
+ '50': I-organisation
292
+ '51': B-person
293
+ '52': I-person
294
+ '53': B-poem
295
+ '54': I-poem
296
+ '55': B-politicalparty
297
+ '56': I-politicalparty
298
+ '57': B-politician
299
+ '58': I-politician
300
+ '59': B-product
301
+ '60': I-product
302
+ '61': B-programlang
303
+ '62': I-programlang
304
+ '63': B-protein
305
+ '64': I-protein
306
+ '65': B-researcher
307
+ '66': I-researcher
308
+ '67': B-scientist
309
+ '68': I-scientist
310
+ '69': B-song
311
+ '70': I-song
312
+ '71': B-task
313
+ '72': I-task
314
+ '73': B-theory
315
+ '74': I-theory
316
+ '75': B-university
317
+ '76': I-university
318
+ '77': B-writer
319
+ '78': I-writer
320
+ splits:
321
+ - name: train
322
+ num_bytes: 65077
323
+ num_examples: 100
324
+ - name: validation
325
+ num_bytes: 259702
326
+ num_examples: 380
327
+ - name: test
328
+ num_bytes: 327195
329
+ num_examples: 465
330
+ download_size: 414065
331
+ dataset_size: 651974
332
+ - config_name: conll2003
333
+ features:
334
+ - name: id
335
+ dtype: string
336
+ - name: tokens
337
+ sequence: string
338
+ - name: ner_tags
339
+ sequence:
340
+ class_label:
341
+ names:
342
+ '0': O
343
+ '1': B-academicjournal
344
+ '2': I-academicjournal
345
+ '3': B-album
346
+ '4': I-album
347
+ '5': B-algorithm
348
+ '6': I-algorithm
349
+ '7': B-astronomicalobject
350
+ '8': I-astronomicalobject
351
+ '9': B-award
352
+ '10': I-award
353
+ '11': B-band
354
+ '12': I-band
355
+ '13': B-book
356
+ '14': I-book
357
+ '15': B-chemicalcompound
358
+ '16': I-chemicalcompound
359
+ '17': B-chemicalelement
360
+ '18': I-chemicalelement
361
+ '19': B-conference
362
+ '20': I-conference
363
+ '21': B-country
364
+ '22': I-country
365
+ '23': B-discipline
366
+ '24': I-discipline
367
+ '25': B-election
368
+ '26': I-election
369
+ '27': B-enzyme
370
+ '28': I-enzyme
371
+ '29': B-event
372
+ '30': I-event
373
+ '31': B-field
374
+ '32': I-field
375
+ '33': B-literarygenre
376
+ '34': I-literarygenre
377
+ '35': B-location
378
+ '36': I-location
379
+ '37': B-magazine
380
+ '38': I-magazine
381
+ '39': B-metrics
382
+ '40': I-metrics
383
+ '41': B-misc
384
+ '42': I-misc
385
+ '43': B-musicalartist
386
+ '44': I-musicalartist
387
+ '45': B-musicalinstrument
388
+ '46': I-musicalinstrument
389
+ '47': B-musicgenre
390
+ '48': I-musicgenre
391
+ '49': B-organisation
392
+ '50': I-organisation
393
+ '51': B-person
394
+ '52': I-person
395
+ '53': B-poem
396
+ '54': I-poem
397
+ '55': B-politicalparty
398
+ '56': I-politicalparty
399
+ '57': B-politician
400
+ '58': I-politician
401
+ '59': B-product
402
+ '60': I-product
403
+ '61': B-programlang
404
+ '62': I-programlang
405
+ '63': B-protein
406
+ '64': I-protein
407
+ '65': B-researcher
408
+ '66': I-researcher
409
+ '67': B-scientist
410
+ '68': I-scientist
411
+ '69': B-song
412
+ '70': I-song
413
+ '71': B-task
414
+ '72': I-task
415
+ '73': B-theory
416
+ '74': I-theory
417
+ '75': B-university
418
+ '76': I-university
419
+ '77': B-writer
420
+ '78': I-writer
421
+ splits:
422
+ - name: train
423
+ num_bytes: 3561081
424
+ num_examples: 14041
425
+ - name: validation
426
+ num_bytes: 891431
427
+ num_examples: 3250
428
+ - name: test
429
+ num_bytes: 811470
430
+ num_examples: 3453
431
+ download_size: 2694794
432
+ dataset_size: 5263982
433
+ - config_name: politics
434
+ features:
435
+ - name: id
436
+ dtype: string
437
+ - name: tokens
438
+ sequence: string
439
+ - name: ner_tags
440
+ sequence:
441
+ class_label:
442
+ names:
443
+ '0': O
444
+ '1': B-academicjournal
445
+ '2': I-academicjournal
446
+ '3': B-album
447
+ '4': I-album
448
+ '5': B-algorithm
449
+ '6': I-algorithm
450
+ '7': B-astronomicalobject
451
+ '8': I-astronomicalobject
452
+ '9': B-award
453
+ '10': I-award
454
+ '11': B-band
455
+ '12': I-band
456
+ '13': B-book
457
+ '14': I-book
458
+ '15': B-chemicalcompound
459
+ '16': I-chemicalcompound
460
+ '17': B-chemicalelement
461
+ '18': I-chemicalelement
462
+ '19': B-conference
463
+ '20': I-conference
464
+ '21': B-country
465
+ '22': I-country
466
+ '23': B-discipline
467
+ '24': I-discipline
468
+ '25': B-election
469
+ '26': I-election
470
+ '27': B-enzyme
471
+ '28': I-enzyme
472
+ '29': B-event
473
+ '30': I-event
474
+ '31': B-field
475
+ '32': I-field
476
+ '33': B-literarygenre
477
+ '34': I-literarygenre
478
+ '35': B-location
479
+ '36': I-location
480
+ '37': B-magazine
481
+ '38': I-magazine
482
+ '39': B-metrics
483
+ '40': I-metrics
484
+ '41': B-misc
485
+ '42': I-misc
486
+ '43': B-musicalartist
487
+ '44': I-musicalartist
488
+ '45': B-musicalinstrument
489
+ '46': I-musicalinstrument
490
+ '47': B-musicgenre
491
+ '48': I-musicgenre
492
+ '49': B-organisation
493
+ '50': I-organisation
494
+ '51': B-person
495
+ '52': I-person
496
+ '53': B-poem
497
+ '54': I-poem
498
+ '55': B-politicalparty
499
+ '56': I-politicalparty
500
+ '57': B-politician
501
+ '58': I-politician
502
+ '59': B-product
503
+ '60': I-product
504
+ '61': B-programlang
505
+ '62': I-programlang
506
+ '63': B-protein
507
+ '64': I-protein
508
+ '65': B-researcher
509
+ '66': I-researcher
510
+ '67': B-scientist
511
+ '68': I-scientist
512
+ '69': B-song
513
+ '70': I-song
514
+ '71': B-task
515
+ '72': I-task
516
+ '73': B-theory
517
+ '74': I-theory
518
+ '75': B-university
519
+ '76': I-university
520
+ '77': B-writer
521
+ '78': I-writer
522
+ splits:
523
+ - name: train
524
+ num_bytes: 143507
525
+ num_examples: 200
526
+ - name: validation
527
+ num_bytes: 422760
528
+ num_examples: 541
529
+ - name: test
530
+ num_bytes: 472690
531
+ num_examples: 651
532
+ download_size: 724168
533
+ dataset_size: 1038957
534
+ - config_name: science
535
+ features:
536
+ - name: id
537
+ dtype: string
538
+ - name: tokens
539
+ sequence: string
540
+ - name: ner_tags
541
+ sequence:
542
+ class_label:
543
+ names:
544
+ '0': O
545
+ '1': B-academicjournal
546
+ '2': I-academicjournal
547
+ '3': B-album
548
+ '4': I-album
549
+ '5': B-algorithm
550
+ '6': I-algorithm
551
+ '7': B-astronomicalobject
552
+ '8': I-astronomicalobject
553
+ '9': B-award
554
+ '10': I-award
555
+ '11': B-band
556
+ '12': I-band
557
+ '13': B-book
558
+ '14': I-book
559
+ '15': B-chemicalcompound
560
+ '16': I-chemicalcompound
561
+ '17': B-chemicalelement
562
+ '18': I-chemicalelement
563
+ '19': B-conference
564
+ '20': I-conference
565
+ '21': B-country
566
+ '22': I-country
567
+ '23': B-discipline
568
+ '24': I-discipline
569
+ '25': B-election
570
+ '26': I-election
571
+ '27': B-enzyme
572
+ '28': I-enzyme
573
+ '29': B-event
574
+ '30': I-event
575
+ '31': B-field
576
+ '32': I-field
577
+ '33': B-literarygenre
578
+ '34': I-literarygenre
579
+ '35': B-location
580
+ '36': I-location
581
+ '37': B-magazine
582
+ '38': I-magazine
583
+ '39': B-metrics
584
+ '40': I-metrics
585
+ '41': B-misc
586
+ '42': I-misc
587
+ '43': B-musicalartist
588
+ '44': I-musicalartist
589
+ '45': B-musicalinstrument
590
+ '46': I-musicalinstrument
591
+ '47': B-musicgenre
592
+ '48': I-musicgenre
593
+ '49': B-organisation
594
+ '50': I-organisation
595
+ '51': B-person
596
+ '52': I-person
597
+ '53': B-poem
598
+ '54': I-poem
599
+ '55': B-politicalparty
600
+ '56': I-politicalparty
601
+ '57': B-politician
602
+ '58': I-politician
603
+ '59': B-product
604
+ '60': I-product
605
+ '61': B-programlang
606
+ '62': I-programlang
607
+ '63': B-protein
608
+ '64': I-protein
609
+ '65': B-researcher
610
+ '66': I-researcher
611
+ '67': B-scientist
612
+ '68': I-scientist
613
+ '69': B-song
614
+ '70': I-song
615
+ '71': B-task
616
+ '72': I-task
617
+ '73': B-theory
618
+ '74': I-theory
619
+ '75': B-university
620
+ '76': I-university
621
+ '77': B-writer
622
+ '78': I-writer
623
+ splits:
624
+ - name: train
625
+ num_bytes: 121928
626
+ num_examples: 200
627
+ - name: validation
628
+ num_bytes: 276118
629
+ num_examples: 450
630
+ - name: test
631
+ num_bytes: 334181
632
+ num_examples: 543
633
+ download_size: 485191
634
+ dataset_size: 732227
635
+ ---
636
+ # Dataset Card for CrossRE
637
+ ## Table of Contents
638
+ - [Table of Contents](#table-of-contents)
639
+ - [Dataset Description](#dataset-description)
640
+ - [Dataset Summary](#dataset-summary)
641
+ - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
642
+ - [Languages](#languages)
643
+ - [Dataset Structure](#dataset-structure)
644
+ - [Data Instances](#data-instances)
645
+ - [Data Fields](#data-fields)
646
+ - [Data Splits](#data-splits)
647
+ - [Dataset Creation](#dataset-creation)
648
+ - [Curation Rationale](#curation-rationale)
649
+ - [Source Data](#source-data)
650
+ - [Annotations](#annotations)
651
+ - [Personal and Sensitive Information](#personal-and-sensitive-information)
652
+ - [Considerations for Using the Data](#considerations-for-using-the-data)
653
+ - [Social Impact of Dataset](#social-impact-of-dataset)
654
+ - [Discussion of Biases](#discussion-of-biases)
655
+ - [Other Known Limitations](#other-known-limitations)
656
+ - [Additional Information](#additional-information)
657
+ - [Dataset Curators](#dataset-curators)
658
+ - [Licensing Information](#licensing-information)
659
+ - [Citation Information](#citation-information)
660
+ - [Contributions](#contributions)
661
+
662
+ ## Dataset Description
663
+ - **Repository:** [CrossNER](https://github.com/zliucr/CrossNER)
664
+ - **Paper:** [CrossNER: Evaluating Cross-Domain Named Entity Recognition](https://arxiv.org/abs/2012.04373)
665
+
666
+ ### Dataset Summary
667
+ CrossNER is a fully-labeled collected of named entity recognition (NER) data spanning over five diverse domains
668
+ (Politics, Natural Science, Music, Literature, and Artificial Intelligence) with specialized entity categories for
669
+ different domains. Additionally, CrossNER also includes unlabeled domain-related corpora for the corresponding five
670
+ domains.
671
+
672
+ For details, see the paper:
673
+ [CrossNER: Evaluating Cross-Domain Named Entity Recognition](https://arxiv.org/abs/2012.04373)
674
+
675
+ ### Supported Tasks and Leaderboards
676
+
677
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
678
+
679
+ ### Languages
680
+
681
+ The language data in CrossNER is in English (BCP-47 en)
682
+
683
+ ## Dataset Structure
684
+
685
+ ### Data Instances
686
+
687
+ #### conll2003
688
+ - **Size of downloaded dataset files:** 2.69 MB
689
+ - **Size of the generated dataset:** 5.26 MB
690
+
691
+ An example of 'train' looks as follows:
692
+ ```json
693
+ {
694
+ "id": "0",
695
+ "tokens": ["EU", "rejects", "German", "call", "to", "boycott", "British", "lamb", "."],
696
+ "ner_tags": [49, 0, 41, 0, 0, 0, 41, 0, 0]
697
+ }
698
+ ```
699
+
700
+ #### politics
701
+ - **Size of downloaded dataset files:** 0.72 MB
702
+ - **Size of the generated dataset:** 1.04 MB
703
+
704
+ An example of 'train' looks as follows:
705
+ ```json
706
+ {
707
+ "id": "0",
708
+ "tokens": ["Parties", "with", "mainly", "Eurosceptic", "views", "are", "the", "ruling", "United", "Russia", ",", "and", "opposition", "parties", "the", "Communist", "Party", "of", "the", "Russian", "Federation", "and", "Liberal", "Democratic", "Party", "of", "Russia", "."],
709
+ "ner_tags": [0, 0, 0, 0, 0, 0, 0, 0, 55, 56, 0, 0, 0, 0, 0, 55, 56, 56, 56, 56, 56, 0, 55, 56, 56, 56, 56, 0]
710
+ }
711
+ ```
712
+
713
+ #### science
714
+ - **Size of downloaded dataset files:** 0.49 MB
715
+ - **Size of the generated dataset:** 0.73 MB
716
+
717
+ An example of 'train' looks as follows:
718
+ ```json
719
+ {
720
+ "id": "0",
721
+ "tokens": ["They", "may", "also", "use", "Adenosine", "triphosphate", ",", "Nitric", "oxide", ",", "and", "ROS", "for", "signaling", "in", "the", "same", "ways", "that", "animals", "do", "."],
722
+ "ner_tags": [0, 0, 0, 0, 15, 16, 0, 15, 16, 0, 0, 15, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
723
+ }
724
+ ```
725
+
726
+ #### music
727
+ - **Size of downloaded dataset files:** 0.41 MB
728
+ - **Size of the generated dataset:** 0.65 MB
729
+
730
+ An example of 'train' looks as follows:
731
+ ```json
732
+ {
733
+ "id": "0",
734
+ "tokens": ["In", "2003", ",", "the", "Stade", "de", "France", "was", "the", "primary", "site", "of", "the", "2003", "World", "Championships", "in", "Athletics", "."],
735
+ "ner_tags": [0, 0, 0, 0, 35, 36, 36, 0, 0, 0, 0, 0, 0, 29, 30, 30, 30, 30, 0]
736
+ }
737
+ ```
738
+
739
+ #### literature
740
+ - **Size of downloaded dataset files:** 0.33 MB
741
+ - **Size of the generated dataset:** 0.58 MB
742
+
743
+ An example of 'train' looks as follows:
744
+ ```json
745
+ {
746
+ "id": "0",
747
+ "tokens": ["In", "1351", ",", "during", "the", "reign", "of", "Emperor", "Toghon", "Temür", "of", "the", "Yuan", "dynasty", ",", "93rd-generation", "descendant", "Kong", "Huan", "(", "孔浣", ")", "'", "s", "2nd", "son", "Kong", "Shao", "(", "孔昭", ")", "moved", "from", "China", "to", "Korea", "during", "the", "Goryeo", ",", "and", "was", "received", "courteously", "by", "Princess", "Noguk", "(", "the", "Mongolian-born", "wife", "of", "the", "future", "king", "Gongmin", ")", "."],
748
+ "ner_tags": [0, 0, 0, 0, 0, 0, 0, 51, 52, 52, 0, 0, 21, 22, 0, 0, 0, 77, 78, 0, 77, 0, 0, 0, 0, 0, 77, 78, 0, 77, 0, 0, 0, 21, 0, 21, 0, 0, 41, 0, 0, 0, 0, 0, 0, 51, 52, 0, 0, 41, 0, 0, 0, 0, 0, 51, 0, 0]
749
+ }
750
+ ```
751
+
752
+ #### ai
753
+ - **Size of downloaded dataset files:** 0.29 MB
754
+ - **Size of the generated dataset:** 0.48 MB
755
+
756
+ An example of 'train' looks as follows:
757
+ ```json
758
+ {
759
+ "id": "0",
760
+ "tokens": ["Popular", "approaches", "of", "opinion-based", "recommender", "system", "utilize", "various", "techniques", "including", "text", "mining", ",", "information", "retrieval", ",", "sentiment", "analysis", "(", "see", "also", "Multimodal", "sentiment", "analysis", ")", "and", "deep", "learning", "X.Y.", "Feng", ",", "H.", "Zhang", ",", "Y.J.", "Ren", ",", "P.H.", "Shang", ",", "Y.", "Zhu", ",", "Y.C.", "Liang", ",", "R.C.", "Guan", ",", "D.", "Xu", ",", "(", "2019", ")", ",", ",", "21", "(", "5", ")", ":", "e12957", "."],
761
+ "ner_tags": [0, 0, 0, 59, 60, 60, 0, 0, 0, 0, 31, 32, 0, 71, 72, 0, 71, 72, 0, 0, 0, 71, 72, 72, 0, 0, 31, 32, 65, 66, 0, 65, 66, 0, 65, 66, 0, 65, 66, 0, 65, 66, 0, 65, 66, 0, 65, 66, 0, 65, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
762
+ }
763
+ ```
764
+
765
+ ### Data Fields
766
+
767
+ The data fields are the same among all splits.
768
+ - `id`: the instance id of this sentence, a `string` feature.
769
+ - `tokens`: the list of tokens of this sentence, a `list` of `string` features.
770
+ - `ner_tags`: the list of entity tags, a `list` of classification labels.
771
+
772
+ ```json
773
+ {"O": 0, "B-academicjournal": 1, "I-academicjournal": 2, "B-album": 3, "I-album": 4, "B-algorithm": 5, "I-algorithm": 6, "B-astronomicalobject": 7, "I-astronomicalobject": 8, "B-award": 9, "I-award": 10, "B-band": 11, "I-band": 12, "B-book": 13, "I-book": 14, "B-chemicalcompound": 15, "I-chemicalcompound": 16, "B-chemicalelement": 17, "I-chemicalelement": 18, "B-conference": 19, "I-conference": 20, "B-country": 21, "I-country": 22, "B-discipline": 23, "I-discipline": 24, "B-election": 25, "I-election": 26, "B-enzyme": 27, "I-enzyme": 28, "B-event": 29, "I-event": 30, "B-field": 31, "I-field": 32, "B-literarygenre": 33, "I-literarygenre": 34, "B-location": 35, "I-location": 36, "B-magazine": 37, "I-magazine": 38, "B-metrics": 39, "I-metrics": 40, "B-misc": 41, "I-misc": 42, "B-musicalartist": 43, "I-musicalartist": 44, "B-musicalinstrument": 45, "I-musicalinstrument": 46, "B-musicgenre": 47, "I-musicgenre": 48, "B-organisation": 49, "I-organisation": 50, "B-person": 51, "I-person": 52, "B-poem": 53, "I-poem": 54, "B-politicalparty": 55, "I-politicalparty": 56, "B-politician": 57, "I-politician": 58, "B-product": 59, "I-product": 60, "B-programlang": 61, "I-programlang": 62, "B-protein": 63, "I-protein": 64, "B-researcher": 65, "I-researcher": 66, "B-scientist": 67, "I-scientist": 68, "B-song": 69, "I-song": 70, "B-task": 71, "I-task": 72, "B-theory": 73, "I-theory": 74, "B-university": 75, "I-university": 76, "B-writer": 77, "I-writer": 78}
774
+ ```
775
+
776
+ ### Data Splits
777
+ | | Train | Dev | Test |
778
+ |--------------|--------|-------|-------|
779
+ | conll2003 | 14,987 | 3,466 | 3,684 |
780
+ | politics | 200 | 541 | 651 |
781
+ | science | 200 | 450 | 543 |
782
+ | music | 100 | 380 | 456 |
783
+ | literature | 100 | 400 | 416 |
784
+ | ai | 100 | 350 | 431 |
785
+
786
+ ## Dataset Creation
787
+
788
+ ### Curation Rationale
789
+
790
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
791
+
792
+ ### Source Data
793
+
794
+ #### Initial Data Collection and Normalization
795
+
796
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
797
+
798
+ #### Who are the source language producers?
799
+
800
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
801
+
802
+ ### Annotations
803
+
804
+ #### Annotation process
805
+
806
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
807
+
808
+ #### Who are the annotators?
809
+
810
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
811
+
812
+ ### Personal and Sensitive Information
813
+
814
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
815
+
816
+ ## Considerations for Using the Data
817
+
818
+ ### Social Impact of Dataset
819
+
820
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
821
+
822
+ ### Discussion of Biases
823
+
824
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
825
+
826
+ ### Other Known Limitations
827
+
828
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
829
+
830
+ ## Additional Information
831
+
832
+ ### Dataset Curators
833
+
834
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
835
+
836
+ ### Licensing Information
837
+
838
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
839
+
840
+ ### Citation Information
841
+
842
+ ```
843
+ @article{liu2020crossner,
844
+ title={CrossNER: Evaluating Cross-Domain Named Entity Recognition},
845
+ author={Zihan Liu and Yan Xu and Tiezheng Yu and Wenliang Dai and Ziwei Ji and Samuel Cahyawijaya and Andrea Madotto and Pascale Fung},
846
+ year={2020},
847
+ eprint={2012.04373},
848
+ archivePrefix={arXiv},
849
+ primaryClass={cs.CL}
850
+ }
851
+ ```
852
+
853
+ ### Contributions
854
+
855
+ Thanks to [@phucdev](https://github.com/phucdev) for adding this dataset.
cross_ner.py ADDED
@@ -0,0 +1,223 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
2
+ #
3
+ # Licensed under the Apache License, Version 2.0 (the "License");
4
+ # you may not use this file except in compliance with the License.
5
+ # You may obtain a copy of the License at
6
+ #
7
+ # http://www.apache.org/licenses/LICENSE-2.0
8
+ #
9
+ # Unless required by applicable law or agreed to in writing, software
10
+ # distributed under the License is distributed on an "AS IS" BASIS,
11
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12
+ # See the License for the specific language governing permissions and
13
+ # limitations under the License.
14
+ """CrossNER is a cross-domain dataset for named entity recognition"""
15
+
16
+
17
+ import json
18
+ import os
19
+
20
+ import datasets
21
+
22
+
23
+ _CITATION = """\
24
+ @article{liu2020crossner,
25
+ title={CrossNER: Evaluating Cross-Domain Named Entity Recognition},
26
+ author={Zihan Liu and Yan Xu and Tiezheng Yu and Wenliang Dai and Ziwei Ji and Samuel Cahyawijaya and Andrea Madotto and Pascale Fung},
27
+ year={2020},
28
+ eprint={2012.04373},
29
+ archivePrefix={arXiv},
30
+ primaryClass={cs.CL}
31
+ }
32
+ """
33
+
34
+ _DESCRIPTION = """\
35
+ CrossNER is a fully-labeled collected of named entity recognition (NER) data spanning over five diverse domains
36
+ (Politics, Natural Science, Music, Literature, and Artificial Intelligence) with specialized entity categories for
37
+ different domains. Additionally, CrossNER also includes unlabeled domain-related corpora for the corresponding five
38
+ domains.
39
+
40
+ For details, see the paper:
41
+ [CrossNER: Evaluating Cross-Domain Named Entity Recognition](https://arxiv.org/abs/2012.04373)
42
+ """
43
+
44
+ _HOMEPAGE = "https://github.com/zliucr/CrossNER"
45
+
46
+ # TODO: Add the licence for the dataset here if you can find it
47
+ _LICENSE = ""
48
+
49
+ # The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
50
+ # This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
51
+ _URLS = {
52
+ "conll2003": {
53
+ "train": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/conll2003/train.txt",
54
+ "validation": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/conll2003/dev.txt",
55
+ "test": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/conll2003/test.txt",
56
+ },
57
+ "politics": {
58
+ "train": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/politics/train.txt",
59
+ "validation": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/politics/dev.txt",
60
+ "test": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/politics/test.txt",
61
+ },
62
+ "science": {
63
+ "train": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/science/train.txt",
64
+ "validation": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/science/dev.txt",
65
+ "test": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/science/test.txt",
66
+ },
67
+ "music": {
68
+ "train": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/music/train.txt",
69
+ "validation": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/music/dev.txt",
70
+ "test": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/music/test.txt",
71
+ },
72
+ "literature": {
73
+ "train": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/literature/train.txt",
74
+ "validation": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/literature/dev.txt",
75
+ "test": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/literature/test.txt",
76
+ },
77
+ "ai": {
78
+ "train": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/ai/train.txt",
79
+ "validation": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/ai/dev.txt",
80
+ "test": "https://raw.githubusercontent.com/zliucr/CrossNER/main/ner_data/ai/test.txt",
81
+ },
82
+ }
83
+
84
+
85
+ _CLASS_LABELS = [
86
+ "O",
87
+ "B-academicjournal", "I-academicjournal",
88
+ "B-album", "I-album",
89
+ "B-algorithm", "I-algorithm",
90
+ "B-astronomicalobject", "I-astronomicalobject",
91
+ "B-award", "I-award",
92
+ "B-band", "I-band",
93
+ "B-book", "I-book",
94
+ "B-chemicalcompound", "I-chemicalcompound",
95
+ "B-chemicalelement", "I-chemicalelement",
96
+ "B-conference", "I-conference",
97
+ "B-country", "I-country",
98
+ "B-discipline", "I-discipline",
99
+ "B-election", "I-election",
100
+ "B-enzyme", "I-enzyme",
101
+ "B-event", "I-event",
102
+ "B-field", "I-field",
103
+ "B-literarygenre", "I-literarygenre",
104
+ "B-location", "I-location",
105
+ "B-magazine", "I-magazine",
106
+ "B-metrics", "I-metrics",
107
+ "B-misc", "I-misc",
108
+ "B-musicalartist", "I-musicalartist",
109
+ "B-musicalinstrument", "I-musicalinstrument",
110
+ "B-musicgenre", "I-musicgenre",
111
+ "B-organisation", "I-organisation",
112
+ "B-person", "I-person",
113
+ "B-poem", "I-poem",
114
+ "B-politicalparty", "I-politicalparty",
115
+ "B-politician", "I-politician",
116
+ "B-product", "I-product",
117
+ "B-programlang", "I-programlang",
118
+ "B-protein", "I-protein",
119
+ "B-researcher", "I-researcher",
120
+ "B-scientist", "I-scientist",
121
+ "B-song", "I-song",
122
+ "B-task", "I-task",
123
+ "B-theory", "I-theory",
124
+ "B-university", "I-university",
125
+ "B-writer", "I-writer",
126
+ ]
127
+
128
+
129
+ class CrossNER(datasets.GeneratorBasedBuilder):
130
+ """CrossNER is a cross-domain dataset for named entity recognition"""
131
+
132
+ VERSION = datasets.Version("1.1.0")
133
+
134
+ # This is an example of a dataset with multiple configurations.
135
+ # If you don't want/need to define several sub-sets in your dataset,
136
+ # just remove the BUILDER_CONFIG_CLASS and the BUILDER_CONFIGS attributes.
137
+
138
+ # If you need to make complex sub-parts in the datasets with configurable options
139
+ # You can create your own builder configuration class to store attribute, inheriting from datasets.BuilderConfig
140
+ # BUILDER_CONFIG_CLASS = MyBuilderConfig
141
+
142
+ # You will be able to load one or the other configurations in the following list with
143
+ # data = datasets.load_dataset('my_dataset', 'first_domain')
144
+ # data = datasets.load_dataset('my_dataset', 'second_domain')
145
+ BUILDER_CONFIGS = [
146
+ datasets.BuilderConfig(name="conll2003", version=VERSION,
147
+ description="This part of CrossNER covers data from the news domain"),
148
+ datasets.BuilderConfig(name="politics", version=VERSION,
149
+ description="This part of CrossNER covers data from the politics domain"),
150
+ datasets.BuilderConfig(name="science", version=VERSION,
151
+ description="This part of CrossNER covers data from the science domain"),
152
+ datasets.BuilderConfig(name="music", version=VERSION,
153
+ description="This part of CrossNER covers data from the music domain"),
154
+ datasets.BuilderConfig(name="literature", version=VERSION,
155
+ description="This part of CrossNER covers data from the literature domain"),
156
+ datasets.BuilderConfig(name="ai", version=VERSION,
157
+ description="This part of CrossNER covers data from the AI domain"),
158
+ ]
159
+
160
+ def _info(self):
161
+ features = datasets.Features(
162
+ {
163
+ "id": datasets.Value("string"),
164
+ "tokens": datasets.Sequence(datasets.Value("string")),
165
+ "ner_tags": datasets.Sequence(datasets.features.ClassLabel(names=_CLASS_LABELS)),
166
+ }
167
+ )
168
+ return datasets.DatasetInfo(
169
+ # This is the description that will appear on the datasets page.
170
+ description=_DESCRIPTION,
171
+ # This defines the different columns of the dataset and their types
172
+ features=features, # Here we define them above because they are different between the two configurations
173
+ # If there's a common (input, target) tuple from the features, uncomment supervised_keys line below and
174
+ # specify them. They'll be used if as_supervised=True in builder.as_dataset.
175
+ # supervised_keys=("sentence", "label"),
176
+ # Homepage of the dataset for documentation
177
+ homepage=_HOMEPAGE,
178
+ # License for the dataset if available
179
+ license=_LICENSE,
180
+ # Citation for the dataset
181
+ citation=_CITATION,
182
+ )
183
+
184
+ def _split_generators(self, dl_manager):
185
+ # If several configurations are possible (listed in BUILDER_CONFIGS), the configuration selected by the user is in self.config.name
186
+
187
+ # dl_manager is a datasets.download.DownloadManager that can be used to download and extract URLS
188
+ # It can accept any type or nested list/dict and will give back the same structure with the url replaced with path to local files.
189
+ # By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
190
+ urls = _URLS[self.config.name]
191
+ downloaded_files = dl_manager.download_and_extract(urls)
192
+ return [datasets.SplitGenerator(name=i, gen_kwargs={"filepath": downloaded_files[str(i)]})
193
+ for i in [datasets.Split.TRAIN, datasets.Split.VALIDATION, datasets.Split.TEST]]
194
+
195
+ # method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
196
+ def _generate_examples(self, filepath):
197
+ # The `key` is for legacy reasons (tfds) and is not important in itself, but must be unique for each example.
198
+ with open(filepath, encoding="utf-8") as f:
199
+ guid = 0
200
+ tokens = []
201
+ ner_tags = []
202
+ for line in f:
203
+ if line == "" or line == "\n":
204
+ if tokens:
205
+ yield guid, {
206
+ "id": str(guid),
207
+ "tokens": tokens,
208
+ "ner_tags": ner_tags,
209
+ }
210
+ guid += 1
211
+ tokens = []
212
+ ner_tags = []
213
+ else:
214
+ splits = line.split("\t")
215
+ tokens.append(splits[0])
216
+ ner_tags.append(splits[1].rstrip())
217
+ # last example
218
+ if tokens:
219
+ yield guid, {
220
+ "id": str(guid),
221
+ "tokens": tokens,
222
+ "ner_tags": ner_tags,
223
+ }