hsohn3 commited on
Commit
ebd1fd1
1 Parent(s): 20e246a

Upload tokenizer.json

Browse files
Files changed (1) hide show
  1. tokenizer.json +511 -0
tokenizer.json ADDED
@@ -0,0 +1,511 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "1.0",
3
+ "truncation": null,
4
+ "padding": null,
5
+ "added_tokens": [
6
+ {
7
+ "id": 0,
8
+ "content": "[UNK]",
9
+ "single_word": false,
10
+ "lstrip": false,
11
+ "rstrip": false,
12
+ "normalized": false,
13
+ "special": true
14
+ },
15
+ {
16
+ "id": 1,
17
+ "content": "[PAD]",
18
+ "single_word": false,
19
+ "lstrip": false,
20
+ "rstrip": false,
21
+ "normalized": false,
22
+ "special": true
23
+ },
24
+ {
25
+ "id": 2,
26
+ "content": "[CLS]",
27
+ "single_word": false,
28
+ "lstrip": false,
29
+ "rstrip": false,
30
+ "normalized": false,
31
+ "special": true
32
+ },
33
+ {
34
+ "id": 3,
35
+ "content": "[SEP]",
36
+ "single_word": false,
37
+ "lstrip": false,
38
+ "rstrip": false,
39
+ "normalized": false,
40
+ "special": true
41
+ },
42
+ {
43
+ "id": 4,
44
+ "content": "[MASK]",
45
+ "single_word": false,
46
+ "lstrip": false,
47
+ "rstrip": false,
48
+ "normalized": false,
49
+ "special": true
50
+ }
51
+ ],
52
+ "normalizer": {
53
+ "type": "BertNormalizer",
54
+ "clean_text": true,
55
+ "handle_chinese_chars": true,
56
+ "strip_accents": null,
57
+ "lowercase": true
58
+ },
59
+ "pre_tokenizer": {
60
+ "type": "WhitespaceSplit"
61
+ },
62
+ "post_processor": {
63
+ "type": "TemplateProcessing",
64
+ "single": [
65
+ {
66
+ "SpecialToken": {
67
+ "id": "[CLS]",
68
+ "type_id": 0
69
+ }
70
+ },
71
+ {
72
+ "Sequence": {
73
+ "id": "A",
74
+ "type_id": 0
75
+ }
76
+ },
77
+ {
78
+ "SpecialToken": {
79
+ "id": "[SEP]",
80
+ "type_id": 0
81
+ }
82
+ }
83
+ ],
84
+ "pair": [
85
+ {
86
+ "SpecialToken": {
87
+ "id": "[CLS]",
88
+ "type_id": 0
89
+ }
90
+ },
91
+ {
92
+ "Sequence": {
93
+ "id": "A",
94
+ "type_id": 0
95
+ }
96
+ },
97
+ {
98
+ "SpecialToken": {
99
+ "id": "[SEP]",
100
+ "type_id": 0
101
+ }
102
+ },
103
+ {
104
+ "Sequence": {
105
+ "id": "B",
106
+ "type_id": 1
107
+ }
108
+ },
109
+ {
110
+ "SpecialToken": {
111
+ "id": "[SEP]",
112
+ "type_id": 1
113
+ }
114
+ }
115
+ ],
116
+ "special_tokens": {
117
+ "[CLS]": {
118
+ "id": "[CLS]",
119
+ "ids": [
120
+ 2
121
+ ],
122
+ "tokens": [
123
+ "[CLS]"
124
+ ]
125
+ },
126
+ "[SEP]": {
127
+ "id": "[SEP]",
128
+ "ids": [
129
+ 3
130
+ ],
131
+ "tokens": [
132
+ "[SEP]"
133
+ ]
134
+ }
135
+ }
136
+ },
137
+ "decoder": {
138
+ "type": "WordPiece",
139
+ "prefix": "##",
140
+ "cleanup": true
141
+ },
142
+ "model": {
143
+ "type": "WordLevel",
144
+ "vocab": {
145
+ "[UNK]": 0,
146
+ "[PAD]": 1,
147
+ "[CLS]": 2,
148
+ "[SEP]": 3,
149
+ "[MASK]": 4,
150
+ "[1hr]": 5,
151
+ "hr=n": 6,
152
+ "rr=n": 7,
153
+ "sbp=n": 8,
154
+ "spo2=n": 9,
155
+ "temp=n": 10,
156
+ "fio2=n": 11,
157
+ "map=n": 12,
158
+ "dbp=n": 13,
159
+ "hr=h": 14,
160
+ "hr=vh": 15,
161
+ "map=h": 16,
162
+ "dbp=h": 17,
163
+ "sbp=h": 18,
164
+ "hr=l": 19,
165
+ "map=vh": 20,
166
+ "dbp=vh": 21,
167
+ "temp=h": 22,
168
+ "sbp=vh": 23,
169
+ "spo2=vh": 24,
170
+ "spo2=n>n": 25,
171
+ "spo2=l": 26,
172
+ "rr=vh": 27,
173
+ "of=n": 28,
174
+ "hr=n>n": 29,
175
+ "temp=vh": 30,
176
+ "rr=n>n": 31,
177
+ "dbp=vl": 32,
178
+ "fio2=h": 33,
179
+ "[2hr]": 34,
180
+ "fio2=vh": 35,
181
+ "rr=h": 36,
182
+ "map=vl": 37,
183
+ "dbp=l": 38,
184
+ "sbp=vl": 39,
185
+ "spo2=n>l": 40,
186
+ "map=l": 41,
187
+ "spo2=l>n": 42,
188
+ "hr=vl": 43,
189
+ "map=n>n": 44,
190
+ "sbp=l": 45,
191
+ "sbp=n>n": 46,
192
+ "dbp=n>n": 47,
193
+ "plt=n": 48,
194
+ "[3hr]": 49,
195
+ "hr=n>l": 50,
196
+ "[4hr]": 51,
197
+ "hr=l>n": 52,
198
+ "cr=n": 53,
199
+ "wbc=n": 54,
200
+ "rr=n>h": 55,
201
+ "bun=n": 56,
202
+ "rr=h>n": 57,
203
+ "hr=h>n": 58,
204
+ "hr=n>h": 59,
205
+ "spo2=vl": 60,
206
+ "of=h": 61,
207
+ "spo2=h>n": 62,
208
+ "spo2=n>h": 63,
209
+ "hr=l>l": 64,
210
+ "spo2=h": 65,
211
+ "rr=h>h": 66,
212
+ "map=l>n": 67,
213
+ "map=n>l": 68,
214
+ "spo2=l>l": 69,
215
+ "sbp=h>n": 70,
216
+ "map=h>n": 71,
217
+ "dbp=n>l": 72,
218
+ "dbp=l>n": 73,
219
+ "spo2=vh>vh": 74,
220
+ "sbp=l>n": 75,
221
+ "hr=h>h": 76,
222
+ "dbp=h>n": 77,
223
+ "sbp=n>l": 78,
224
+ "of=vh": 79,
225
+ "map=n>h": 80,
226
+ "sbp=n>h": 81,
227
+ "dbp=n>h": 82,
228
+ "spo2=n>vh": 83,
229
+ "spo2=vh>n": 84,
230
+ "spo2=h>vh": 85,
231
+ "spo2=vh>h": 86,
232
+ "[5hr]": 87,
233
+ "spo2=vl>l": 88,
234
+ "rr=l>n": 89,
235
+ "spo2=l>vl": 90,
236
+ "spo2=h>h": 91,
237
+ "hr=vl>l": 92,
238
+ "hr=l>vl": 93,
239
+ "rr=n>l": 94,
240
+ "of=l": 95,
241
+ "spo2=vl>n": 96,
242
+ "temp=l": 97,
243
+ "spo2=n>vl": 98,
244
+ "rr=h>vh": 99,
245
+ "rr=l": 100,
246
+ "rr=vh>h": 101,
247
+ "hr=vl>vl": 102,
248
+ "dbp=l>l": 103,
249
+ "hr=vh>h": 104,
250
+ "map=l>l": 105,
251
+ "sbp=vh>h": 106,
252
+ "plt=l": 107,
253
+ "map=vh>h": 108,
254
+ "hr=h>vh": 109,
255
+ "bun=h": 110,
256
+ "rr=vl>n": 111,
257
+ "wbc=h": 112,
258
+ "dbp=vh>h": 113,
259
+ "sbp=h>h": 114,
260
+ "sbp=h>vh": 115,
261
+ "map=h>h": 116,
262
+ "sbp=l>l": 117,
263
+ "map=h>vh": 118,
264
+ "dbp=vl>l": 119,
265
+ "plt=h": 120,
266
+ "dbp=vh>n": 121,
267
+ "bun=l": 122,
268
+ "[6hr]": 123,
269
+ "dbp=l>vl": 124,
270
+ "dbp=h>h": 125,
271
+ "dbp=h>vh": 126,
272
+ "spo2=vl>vl": 127,
273
+ "wbc=l": 128,
274
+ "map=vl>l": 129,
275
+ "rr=vl": 130,
276
+ "map=vh>n": 131,
277
+ "cr=h": 132,
278
+ "sbp=vh>n": 133,
279
+ "rr=n>vl": 134,
280
+ "rr=vh>vh": 135,
281
+ "map=l>vl": 136,
282
+ "sbp=vl>l": 137,
283
+ "dbp=n>vh": 138,
284
+ "bun=vh": 139,
285
+ "rr=l>l": 140,
286
+ "wbc=vh": 141,
287
+ "hr=vl>n": 142,
288
+ "sbp=l>vl": 143,
289
+ "rr=vl>l": 144,
290
+ "sbp=n>vh": 145,
291
+ "lt=n": 146,
292
+ "map=n>vh": 147,
293
+ "map=vl>n": 148,
294
+ "[7hr]": 149,
295
+ "hr=vh>vh": 150,
296
+ "map=n>vl": 151,
297
+ "hr=n>vl": 152,
298
+ "plt=vl": 153,
299
+ "rr=vl>vl": 154,
300
+ "temp=vl": 155,
301
+ "br=n": 156,
302
+ "dbp=vl>n": 157,
303
+ "dbp=n>vl": 158,
304
+ "dbp=vl>vl": 159,
305
+ "cr=vh": 160,
306
+ "sbp=vl>n": 161,
307
+ "rr=l>vl": 162,
308
+ "[8hr]": 163,
309
+ "sbp=vh>vh": 164,
310
+ "sbp=n>vl": 165,
311
+ "plt=vh": 166,
312
+ "bun=vl": 167,
313
+ "wbc=vl": 168,
314
+ "dbp=vh>vh": 169,
315
+ "map=vh>vh": 170,
316
+ "rr=vh>n": 171,
317
+ "rr=n>vh": 172,
318
+ "map=vl>vl": 173,
319
+ "cr=l": 174,
320
+ "sbp=vl>vl": 175,
321
+ "fio2=n>h": 176,
322
+ "cr=vl": 177,
323
+ "temp=n>h": 178,
324
+ "[9hr]": 179,
325
+ "temp=l>n": 180,
326
+ "hr=n>vh": 181,
327
+ "hr=vh>n": 182,
328
+ "temp=h>n": 183,
329
+ "spo2=vh>l": 184,
330
+ "spo2=h>l": 185,
331
+ "spo2=l>vh": 186,
332
+ "spo2=l>h": 187,
333
+ "spo2=vl>vh": 188,
334
+ "temp=h>vh": 189,
335
+ "[10hr]": 190,
336
+ "spo2=vh>vl": 191,
337
+ "lt=h": 192,
338
+ "map=h>l": 193,
339
+ "temp=n>l": 194,
340
+ "spo2=vl>h": 195,
341
+ "dbp=h>l": 196,
342
+ "sbp=h>l": 197,
343
+ "rr=h>l": 198,
344
+ "rr=l>h": 199,
345
+ "temp=vh>h": 200,
346
+ "temp=vl>l": 201,
347
+ "map=l>h": 202,
348
+ "br=h": 203,
349
+ "bd=n": 204,
350
+ "[11hr]": 205,
351
+ "spo2=h>vl": 206,
352
+ "lt=l": 207,
353
+ "hr=l>h": 208,
354
+ "dbp=l>h": 209,
355
+ "rr=vl>h": 210,
356
+ "sbp=l>h": 211,
357
+ "temp=n>n": 212,
358
+ "lt=vh": 213,
359
+ "dbp=vh>l": 214,
360
+ "[12hr]": 215,
361
+ "rr=h>vl": 216,
362
+ "map=vh>l": 217,
363
+ "temp=l>vl": 218,
364
+ "temp=h>h": 219,
365
+ "hr=h>l": 220,
366
+ "br=vh": 221,
367
+ "rr=vh>vl": 222,
368
+ "hr=vl>h": 223,
369
+ "rr=vl>vh": 224,
370
+ "crp=n": 225,
371
+ "lt=vl": 226,
372
+ "sbp=vh>l": 227,
373
+ "br=l": 228,
374
+ "fio2=h>n": 229,
375
+ "br=vl": 230,
376
+ "hr=h>vl": 231,
377
+ "temp=vl>n": 232,
378
+ "[13hr]": 233,
379
+ "map=h>vl": 234,
380
+ "dbp=l>vh": 235,
381
+ "dbp=h>vl": 236,
382
+ "map=l>vh": 237,
383
+ "temp=l>l": 238,
384
+ "of=vl": 239,
385
+ "map=vh>vl": 240,
386
+ "hr=vl>vh": 241,
387
+ "dbp=vh>vl": 242,
388
+ "dbp=vl>h": 243,
389
+ "map=vl>h": 244,
390
+ "[999hr]": 245,
391
+ "hr=l>vh": 246,
392
+ "sbp=l>vh": 247,
393
+ "sbp=h>vl": 248,
394
+ "crp=l": 249,
395
+ "hr=vh>l": 250,
396
+ "sr=n": 251,
397
+ "sbp=vl>h": 252,
398
+ "hr=vh>vl": 253,
399
+ "[24hr]": 254,
400
+ "sbp=vh>vl": 255,
401
+ "temp=n>vh": 256,
402
+ "bd=l": 257,
403
+ "dbp=vl>vh": 258,
404
+ "[23hr]": 259,
405
+ "[14hr]": 260,
406
+ "rr=vh>l": 261,
407
+ "[21hr]": 262,
408
+ "temp=vh>n": 263,
409
+ "[22hr]": 264,
410
+ "[20hr]": 265,
411
+ "bd=h": 266,
412
+ "temp=vh>vh": 267,
413
+ "map=vl>vh": 268,
414
+ "[15hr]": 269,
415
+ "[18hr]": 270,
416
+ "fio2=vh>h": 271,
417
+ "bd=vh": 272,
418
+ "temp=n>vl": 273,
419
+ "fio2=n>n": 274,
420
+ "[19hr]": 275,
421
+ "crp=h": 276,
422
+ "rr=l>vh": 277,
423
+ "[16hr]": 278,
424
+ "of=n>l": 279,
425
+ "of=vh>h": 280,
426
+ "fio2=h>vh": 281,
427
+ "temp=l>h": 282,
428
+ "sbp=vl>vh": 283,
429
+ "[17hr]": 284,
430
+ "fio2=n>vh": 285,
431
+ "temp=vl>vl": 286,
432
+ "of=h>n": 287,
433
+ "of=h>vh": 288,
434
+ "bd=vl": 289,
435
+ "fio2=h>h": 290,
436
+ "crp=vh": 291,
437
+ "of=n>h": 292,
438
+ "fio2=vh>n": 293,
439
+ "sr=l": 294,
440
+ "of=l>n": 295,
441
+ "temp=vl>h": 296,
442
+ "temp=h>l": 297,
443
+ "sr=h": 298,
444
+ "sr=vl": 299,
445
+ "sr=vh": 300,
446
+ "of=vh>n": 301,
447
+ "pct=n": 302,
448
+ "of=n>vh": 303,
449
+ "temp=h>vl": 304,
450
+ "temp=vh>vl": 305,
451
+ "temp=vh>l": 306,
452
+ "temp=l>vh": 307,
453
+ "temp=vl>vh": 308,
454
+ "fio2=vh>vh": 309,
455
+ "of=n>n": 310,
456
+ "fio2=vl": 311,
457
+ "plt=l>vl": 312,
458
+ "of=h>h": 313,
459
+ "of=h>l": 314,
460
+ "pct=h": 315,
461
+ "plt=vl>l": 316,
462
+ "of=l>l": 317,
463
+ "of=vh>vh": 318,
464
+ "of=l>vl": 319,
465
+ "of=vl>n": 320,
466
+ "pct=vh": 321,
467
+ "plt=l>n": 322,
468
+ "plt=n>l": 323,
469
+ "plt=vl>n": 324,
470
+ "of=vl>l": 325,
471
+ "of=l>h": 326,
472
+ "of=l>vh": 327,
473
+ "of=n>vl": 328,
474
+ "bd=l>n": 329,
475
+ "bun=l>n": 330,
476
+ "bun=n>h": 331,
477
+ "cr=n>h": 332,
478
+ "fio2=n>vl": 333,
479
+ "fio2=vh>vl": 334,
480
+ "lt=h>n": 335,
481
+ "lt=n>h": 336,
482
+ "bd=vl>l": 337,
483
+ "br=n>n": 338,
484
+ "bun=h>vh": 339,
485
+ "bun=n>l": 340,
486
+ "cr=h>vh": 341,
487
+ "cr=n>l": 342,
488
+ "cr=n>vl": 343,
489
+ "lt=l>n": 344,
490
+ "lt=n>l": 345,
491
+ "lt=n>vl": 346,
492
+ "lt=vh>h": 347,
493
+ "lt=vl>l": 348,
494
+ "of=h>vl": 349,
495
+ "of=vh>l": 350,
496
+ "of=vh>vl": 351,
497
+ "of=vl>h": 352,
498
+ "plt=h>n": 353,
499
+ "plt=l>l": 354,
500
+ "wbc=h>n": 355,
501
+ "wbc=h>vh": 356,
502
+ "wbc=l>vl": 357,
503
+ "wbc=n>h": 358,
504
+ "wbc=n>l": 359,
505
+ "wbc=vh>h": 360,
506
+ "wbc=vl>l": 361,
507
+ "wbc=vl>vh": 362
508
+ },
509
+ "unk_token": "[UNK]"
510
+ }
511
+ }