word-stress-tokenizer / tokenizer.json
abletobetable's picture
Upload tokenizer
cdce4ea
raw
history blame
17.1 kB
{
"version": "1.0",
"truncation": null,
"padding": null,
"added_tokens": [
{
"id": 0,
"content": "<cls>",
"single_word": false,
"lstrip": false,
"rstrip": false,
"normalized": false,
"special": true
},
{
"id": 1,
"content": "<unk>",
"single_word": false,
"lstrip": false,
"rstrip": false,
"normalized": false,
"special": true
},
{
"id": 2,
"content": "<pad>",
"single_word": false,
"lstrip": false,
"rstrip": false,
"normalized": false,
"special": true
}
],
"normalizer": null,
"pre_tokenizer": null,
"post_processor": {
"type": "TemplateProcessing",
"single": [
{
"SpecialToken": {
"id": "<cls>",
"type_id": 0
}
},
{
"Sequence": {
"id": "A",
"type_id": 0
}
}
],
"pair": [
{
"Sequence": {
"id": "A",
"type_id": 0
}
},
{
"Sequence": {
"id": "B",
"type_id": 1
}
}
],
"special_tokens": {
"<cls>": {
"id": "<cls>",
"ids": [
0
],
"tokens": [
"<cls>"
]
}
}
},
"decoder": {
"type": "Metaspace",
"replacement": "▁",
"add_prefix_space": true
},
"model": {
"type": "Unigram",
"unk_id": 1,
"vocab": [
[
"<cls>",
0.0
],
[
"<unk>",
0.0
],
[
"<pad>",
0.0
],
[
"е",
-3.1636796363706257
],
[
"с",
-3.359173210921515
],
[
"м",
-3.476731179390411
],
[
"в",
-3.605386252605541
],
[
"и",
-3.651066013791489
],
[
"о",
-3.715245522180407
],
[
"л",
-3.817479263617182
],
[
"а",
-3.903019620887452
],
[
"т",
-3.9199928379587927
],
[
"у",
-3.930405667964292
],
[
"й",
-4.0541895018570315
],
[
"к",
-4.102994397537689
],
[
"х",
-4.103691610663567
],
[
"р",
-4.16366378424714
],
[
"я",
-4.185170395178137
],
[
"д",
-4.293637726858794
],
[
"по",
-4.318874404725774
],
[
"н",
-4.319659895767661
],
[
"ра",
-4.326548476322561
],
[
"ли",
-4.463338075253289
],
[
"на",
-4.548396485176252
],
[
"ка",
-4.557609822166175
],
[
"г",
-4.578202766924266
],
[
"п",
-4.5948832974279
],
[
"ла",
-4.6224297246247685
],
[
"з",
-4.668711901804954
],
[
"за",
-4.679470001112351
],
[
"ж",
-4.692888448272972
],
[
"ны",
-4.717019931310669
],
[
"ни",
-4.717178138341594
],
[
"б",
-4.724329749814941
],
[
"ре",
-4.81794062496984
],
[
"ся",
-4.837228337071681
],
[
"не",
-4.8478842031513585
],
[
"ч",
-4.887512861290738
],
[
"ро",
-4.926997101838513
],
[
"ло",
-4.974133703658508
],
[
"та",
-5.03710813523057
],
[
"ми",
-5.050561901133183
],
[
"вы",
-5.066939597672462
],
[
"ю",
-5.073388771984658
],
[
"ва",
-5.113926274278537
],
[
"ш",
-5.117244871533595
],
[
"ки",
-5.119883285875997
],
[
"но",
-5.1228343148009206
],
[
"ль",
-5.140078050336989
],
[
"во",
-5.172027764098098
],
[
"ет",
-5.177442189854446
],
[
"про",
-5.180757178756538
],
[
"до",
-5.195432996961143
],
[
"го",
-5.202898545471494
],
[
"ы",
-5.220288580282355
],
[
"те",
-5.223318808225277
],
[
"да",
-5.258319969867447
],
[
"сь",
-5.296768127931532
],
[
"при",
-5.300719847408466
],
[
"ст",
-5.312071600207292
],
[
"об",
-5.321440340809565
],
[
"ри",
-5.351944024735443
],
[
"ру",
-5.382717583024739
],
[
"де",
-5.423357093765036
],
[
"ти",
-5.430152754589945
],
[
"от",
-5.452013842602504
],
[
"па",
-5.461349448256415
],
[
"ди",
-5.462112441879385
],
[
"со",
-5.470733756935259
],
[
"ку",
-5.474588929119926
],
[
"ко",
-5.504473412608787
],
[
"то",
-5.504860293985371
],
[
"га",
-5.682461175015034
],
[
"бо",
-5.702286932153874
],
[
"ля",
-5.707186880493858
],
[
"чи",
-5.72614960880744
],
[
"ры",
-5.758139810477672
],
[
"ь",
-5.769722615320581
],
[
"раз",
-5.777657793853892
],
[
"ин",
-5.783559540075617
],
[
"ной",
-5.792938114764205
],
[
"ен",
-5.810652857528926
],
[
"под",
-5.843061465852913
],
[
"ан",
-5.865100339367238
],
[
"пи",
-5.879046574282175
],
[
"ар",
-5.893862057404115
],
[
"ть",
-5.897306442912062
],
[
"ф",
-5.9022291272716805
],
[
"жи",
-5.925370546484605
],
[
"ци",
-5.9281007525137515
],
[
"че",
-5.945348977901575
],
[
"ты",
-5.97216676011774
],
[
"ба",
-5.977038105260875
],
[
"вер",
-5.9834285662208355
],
[
"ная",
-5.990849208890617
],
[
"ча",
-6.013202406107327
],
[
"пе",
-6.032341480893758
],
[
"са",
-6.034450309096375
],
[
"ем",
-6.037807542535029
],
[
"ще",
-6.03837094183477
],
[
"ше",
-6.062354644467768
],
[
"из",
-6.066288952500397
],
[
"ств",
-6.069384268201751
],
[
"ного",
-6.082294432517263
],
[
"ско",
-6.08231914970808
],
[
"э",
-6.100856930219237
],
[
"ком",
-6.103344547667655
],
[
"ски",
-6.106879466116212
],
[
"ё",
-6.119189253137697
],
[
"ов",
-6.13781755905177
],
[
"би",
-6.178944074692677
],
[
"ста",
-6.185855399748315
],
[
"пу",
-6.204423533569656
],
[
"ков",
-6.213633369881909
],
[
"рас",
-6.2157363291207215
],
[
"пере",
-6.217680605082459
],
[
"ши",
-6.237591353842032
],
[
"ить",
-6.252525660416676
],
[
"ну",
-6.263488690915467
],
[
"ду",
-6.272614811909405
],
[
"це",
-6.29024569911721
],
[
"сто",
-6.298801808923643
],
[
"енны",
-6.302090293473123
],
[
"ц",
-6.308440789473085
],
[
"сти",
-6.334320371974204
],
[
"же",
-6.335635473316212
],
[
"ал",
-6.358520860862261
],
[
"му",
-6.364234522588494
],
[
"ров",
-6.383232073647163
],
[
"кон",
-6.39040970858249
],
[
"нов",
-6.403321122273848
],
[
"ном",
-6.404496872419564
],
[
"мен",
-6.406441341803423
],
[
"су",
-6.409942213032795
],
[
"ги",
-6.410630152721249
],
[
"ное",
-6.420773351327054
],
[
"ск",
-6.445759163556909
],
[
"тор",
-6.446487040563172
],
[
"щи",
-6.448299737869269
],
[
"тер",
-6.4611993948600475
],
[
"ке",
-6.476807169306805
],
[
"пред",
-6.480919764600883
],
[
"ту",
-6.510780721409588
],
[
"ша",
-6.5168657798038625
],
[
"бу",
-6.5276731424438506
],
[
"ения",
-6.5389266142007925
],
[
"ную",
-6.552064033974945
],
[
"пре",
-6.553167434011976
],
[
"вал",
-6.55725656498766
],
[
"ом",
-6.568516369824358
],
[
"ать",
-6.576879449964455
],
[
"кой",
-6.586751821830784
],
[
"ным",
-6.594442363464129
],
[
"ей",
-6.599767946641586
],
[
"ил",
-6.621395002677157
],
[
"лю",
-6.623376951115417
],
[
"ня",
-6.630620647600898
],
[
"фи",
-6.631450294454989
],
[
"щ",
-6.652009873573955
],
[
"лен",
-6.671904516820074
],
[
"ой",
-6.698990150040521
],
[
"ение",
-6.701814987349789
],
[
"тра",
-6.7350151247990135
],
[
"зна",
-6.7628336982097625
],
[
"нны",
-6.7636883592262285
],
[
"став",
-6.77706314032819
],
[
"ают",
-6.781120088269674
],
[
"ща",
-6.801003632576516
],
[
"нул",
-6.833802047509301
],
[
"ят",
-6.83599763637482
],
[
"ца",
-6.841362171180942
],
[
"вать",
-6.8454722753654105
],
[
"гу",
-6.855305485842227
],
[
"тель",
-6.865088978982656
],
[
"дер",
-6.869778946510175
],
[
"ую",
-6.8832869290382055
],
[
"зы",
-6.887226954371978
],
[
"ого",
-6.9096408817435835
],
[
"ды",
-6.917300681629015
],
[
"вали",
-6.92359862414774
],
[
"воз",
-6.924974892692155
],
[
"сту",
-6.938330038589572
],
[
"вши",
-6.944084484532088
],
[
"мер",
-6.949151102970278
],
[
"стр",
-6.964299475320226
],
[
"вед",
-6.973302237822134
],
[
"вод",
-7.007964932291481
],
[
"ская",
-7.0120720777731504
],
[
"лов",
-7.0172910864921105
],
[
"зи",
-7.018282598482371
],
[
"мат",
-7.03229387976162
],
[
"сы",
-7.045030730480992
],
[
"вает",
-7.061487717557232
],
[
"фе",
-7.061960024820603
],
[
"ность",
-7.062069414670493
],
[
"альны",
-7.066007981330634
],
[
"тельны",
-7.070899992884045
],
[
"ически",
-7.074910446308691
],
[
"кор",
-7.077010664573003
],
[
"стро",
-7.082808344718579
],
[
"бы",
-7.089034041901165
],
[
"енно",
-7.092846340929073
],
[
"ально",
-7.106798401325548
],
[
"жен",
-7.1158283656810495
],
[
"пер",
-7.140652274256048
],
[
"нь",
-7.146275463065722
],
[
"ство",
-7.191344050847165
],
[
"лени",
-7.192798006712744
],
[
"ками",
-7.223665578070645
],
[
"сер",
-7.273427212354388
],
[
"ция",
-7.309302043202202
],
[
"пы",
-7.311239771943472
],
[
"ическ",
-7.362626913781076
],
[
"ман",
-7.372506880655842
],
[
"цы",
-7.380877909236976
],
[
"лось",
-7.388421354263927
],
[
"фа",
-7.396928557939997
],
[
"полн",
-7.404319309649212
],
[
"бес",
-7.411486081472564
],
[
"ания",
-7.424202557681504
],
[
"бер",
-7.440982773293648
],
[
"итель",
-7.461792961820359
],
[
"вес",
-7.467919044018315
],
[
"мол",
-7.4733081240141805
],
[
"ант",
-7.478987381298584
],
[
"фор",
-7.484167266702335
],
[
"ключ",
-7.498912525005471
],
[
"ности",
-7.505175854659576
],
[
"ешь",
-7.508768160872985
],
[
"авто",
-7.538910090995354
],
[
"теле",
-7.540809540515366
],
[
"тур",
-7.54197419421619
],
[
"без",
-7.575807709091022
],
[
"вид",
-7.5878148311125475
],
[
"пле",
-7.60349359078537
],
[
"клад",
-7.615845269944962
],
[
"нуть",
-7.641275309552204
],
[
"ость",
-7.646320761168187
],
[
"ание",
-7.662126846510478
],
[
"очны",
-7.664907785029915
],
[
"луч",
-7.718489513949257
],
[
"олог",
-7.732846447649656
],
[
"ительно",
-7.74534876568144
],
[
"вел",
-7.818080089048796
],
[
"твор",
-7.820857524203315
],
[
"ющих",
-7.874749841019598
],
[
"ичны",
-7.924801361025125
],
[
"плат",
-7.96084116450813
],
[
"ировать",
-7.963776371734232
],
[
"благо",
-7.988391122025848
],
[
"ющий",
-8.024543177208486
],
[
"сход",
-8.135419072857303
],
[
"смотр",
-8.146977660829759
],
[
"следова",
-8.638979587708413
],
[
"электро",
-8.735356690654438
],
[
"ъ",
-9.097443078286028
]
]
}
}