Hugging Face's logo

ICEF-NLP
/
bcms-bertic-comtext-sr-legal-msd-ekavica

Token Classification
Transformers
Safetensors
Serbian
electra
legal
Inference Endpoints
Model card Files Community
vukbatanovic commited on
Commit
11d2de0
•
1 Parent(s): f58d674

Initial commit

Browse files
Files changed (5) hide show
  1. config.json +910 -0
  2. model.safetensors +3 -0
  3. special_tokens_map.json +7 -0
  4. tokenizer_config.json +58 -0
  5. vocab.txt +0 -0
config.json ADDED
@@ -0,0 +1,910 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "classla/bcms-bertic",
3
+ "architectures": [
4
+ "ElectraForTokenClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "classifier_dropout": null,
8
+ "embedding_size": 768,
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 768,
12
+ "id2label": {
13
+ "0": "Xf",
14
+ "1": "Y",
15
+ "2": "Ncmsg",
16
+ "3": "Npmsg",
17
+ "4": "Mdm",
18
+ "5": "Mdc",
19
+ "6": "Agpmsny",
20
+ "7": "Npmsn",
21
+ "8": "Sl",
22
+ "9": "Ncmsl",
23
+ "10": "Cc",
24
+ "11": "Z",
25
+ "12": "Ncnsn",
26
+ "13": "Sg",
27
+ "14": "Agpmsgy",
28
+ "15": "Ncmpl",
29
+ "16": "Ncmpg",
30
+ "17": "Vmr3p",
31
+ "18": "Ncmsan",
32
+ "19": "Ncmsn",
33
+ "20": "Mdo",
34
+ "21": "Ncfsg",
35
+ "22": "Vmr3s",
36
+ "23": "Px--sa",
37
+ "24": "Rgp",
38
+ "25": "Agpfsny",
39
+ "26": "Ncfsn",
40
+ "27": "Agpmsdy",
41
+ "28": "Agpmsly",
42
+ "29": "Sa",
43
+ "30": "Ncmpa",
44
+ "31": "Pi-mpa",
45
+ "32": "Var3s",
46
+ "33": "Appmsnn",
47
+ "34": "Pd-msn",
48
+ "35": "Ncnsa",
49
+ "36": "Agpfsgy",
50
+ "37": "Ncfsl",
51
+ "38": "Cs",
52
+ "39": "Agpmsnn",
53
+ "40": "Si",
54
+ "41": "Ncfpi",
55
+ "42": "Pd-msg",
56
+ "43": "Ncmsi",
57
+ "44": "Ncnsg",
58
+ "45": "Agpfpgy",
59
+ "46": "Ncfpg",
60
+ "47": "Npmsl",
61
+ "48": "Ncnsi",
62
+ "49": "Aspfsny",
63
+ "50": "Agpmpay",
64
+ "51": "Npfsg",
65
+ "52": "Pi-msay",
66
+ "53": "Npfsn",
67
+ "54": "Agpfpny",
68
+ "55": "Ncfpn",
69
+ "56": "Ncmsd",
70
+ "57": "Agpfpay",
71
+ "58": "Ncfpa",
72
+ "59": "Pi-fpn",
73
+ "60": "Agpnsiy",
74
+ "61": "Pi-fpa",
75
+ "62": "Var3p",
76
+ "63": "Appfpny",
77
+ "64": "Agpfpiy",
78
+ "65": "Qq",
79
+ "66": "Qo",
80
+ "67": "Pi-msl",
81
+ "68": "Pi3n-n",
82
+ "69": "Agpfsay",
83
+ "70": "Ncfsa",
84
+ "71": "Agpmply",
85
+ "72": "Agpfply",
86
+ "73": "Ncfpl",
87
+ "74": "Ncnpl",
88
+ "75": "Pd-nsi",
89
+ "76": "Vmp-sn",
90
+ "77": "Ncnsd",
91
+ "78": "X",
92
+ "79": "Vmn",
93
+ "80": "Pi-msan",
94
+ "81": "Vmp-sm",
95
+ "82": "Agpmpgy",
96
+ "83": "Agpnsgy",
97
+ "84": "Agpfsiy",
98
+ "85": "Ncfsi",
99
+ "86": "Qz",
100
+ "87": "Agcmsiy",
101
+ "88": "Pi-fsa",
102
+ "89": "Sd",
103
+ "90": "Px-fpd",
104
+ "91": "Ncfpd",
105
+ "92": "Agpnsny",
106
+ "93": "Pi-mpd",
107
+ "94": "Px-mpd",
108
+ "95": "Ncmpd",
109
+ "96": "Agpmpdy",
110
+ "97": "Agpnpiy",
111
+ "98": "Mlc",
112
+ "99": "Pi-nsd",
113
+ "100": "Appfsny",
114
+ "101": "Pd-msl",
115
+ "102": "Pd-fsn",
116
+ "103": "Mrc",
117
+ "104": "Pd-nsd",
118
+ "105": "Ncnpd",
119
+ "106": "Appmpny",
120
+ "107": "Van",
121
+ "108": "Agpmpiy",
122
+ "109": "Px-mpi",
123
+ "110": "Ncmpi",
124
+ "111": "Pi-nsa",
125
+ "112": "Pd-npg",
126
+ "113": "Ncnpg",
127
+ "114": "Agsfsly",
128
+ "115": "Agpfsly",
129
+ "116": "Pi-fsn",
130
+ "117": "Pi-msg",
131
+ "118": "Ncmpn",
132
+ "119": "Pi-fsg",
133
+ "120": "Pp3msi",
134
+ "121": "Agpmsiy",
135
+ "122": "Pd-nsn",
136
+ "123": "Px-fsa",
137
+ "124": "Vaa3s",
138
+ "125": "Vmp-sf",
139
+ "126": "Ncnpn",
140
+ "127": "Agpnsay",
141
+ "128": "Pd-fpn",
142
+ "129": "Vmp-pf",
143
+ "130": "Pd-fpg",
144
+ "131": "Pd-msd",
145
+ "132": "Vap-pf",
146
+ "133": "Agpmsann",
147
+ "134": "Pp3-pa",
148
+ "135": "Mlofsg",
149
+ "136": "Mlofsn",
150
+ "137": "Pi-fsi",
151
+ "138": "Pd-fsi",
152
+ "139": "Pd-nsa",
153
+ "140": "Rgc",
154
+ "141": "Pd-fpa",
155
+ "142": "Px-msg",
156
+ "143": "Pi-fsl",
157
+ "144": "Mlofsa",
158
+ "145": "Px-fsl",
159
+ "146": "Pd-msan",
160
+ "147": "Agpnpay",
161
+ "148": "Ncnpa",
162
+ "149": "Rr",
163
+ "150": "Pd-fsg",
164
+ "151": "Pi-nsg",
165
+ "152": "Pd-msi",
166
+ "153": "Mlcf-g",
167
+ "154": "Ncmsay",
168
+ "155": "Agpnply",
169
+ "156": "Appnsny",
170
+ "157": "Pi-nsn",
171
+ "158": "Agcmsly",
172
+ "159": "Ncnsl",
173
+ "160": "Npfsl",
174
+ "161": "Pi-mpi",
175
+ "162": "Pi-msi",
176
+ "163": "Xw",
177
+ "164": "Npfsa",
178
+ "165": "Pd-fsa",
179
+ "166": "Mlofpd",
180
+ "167": "Npmsay",
181
+ "168": "Npmsan",
182
+ "169": "Pi-msn",
183
+ "170": "Mlomsg",
184
+ "171": "Px-mpa",
185
+ "172": "Agpmsayn",
186
+ "173": "Pi-msd",
187
+ "174": "Ps3msd",
188
+ "175": "Pi-mpn",
189
+ "176": "Agpmpny",
190
+ "177": "Px-msl",
191
+ "178": "Rgs",
192
+ "179": "Pd-nsl",
193
+ "180": "Agcmpny",
194
+ "181": "Pi-mpg",
195
+ "182": "Pi-npg",
196
+ "183": "Agpnpgy",
197
+ "184": "Px-nsa",
198
+ "185": "Vmf3s",
199
+ "186": "Vmf3p",
200
+ "187": "Ps3nsn",
201
+ "188": "Pi-npn",
202
+ "189": "Mlonpg",
203
+ "190": "Agpnsly",
204
+ "191": "Pi-fpl",
205
+ "192": "Ps3fpg",
206
+ "193": "Pp3msa",
207
+ "194": "Px-msan",
208
+ "195": "Px-fsg",
209
+ "196": "Pi-npa",
210
+ "197": "Px-fpg",
211
+ "198": "Pp3msg",
212
+ "199": "Px-mpg",
213
+ "200": "Pi-fpg",
214
+ "201": "Ps3fsg",
215
+ "202": "Ps3fsd",
216
+ "203": "Agpfsdy",
217
+ "204": "Ncfsd",
218
+ "205": "Agsmsny",
219
+ "206": "Npfsd",
220
+ "207": "Npmsd",
221
+ "208": "Px-msi",
222
+ "209": "Agcnsgy",
223
+ "210": "Agcmsny",
224
+ "211": "Mlomsi",
225
+ "212": "Var1s",
226
+ "213": "Agpmsayy",
227
+ "214": "Vmp-pm",
228
+ "215": "Npnsl",
229
+ "216": "Npnsg",
230
+ "217": "Mlomsn",
231
+ "218": "Mro",
232
+ "219": "Agsmsly",
233
+ "220": "Mds",
234
+ "221": "Npnsn",
235
+ "222": "Pi3n-g",
236
+ "223": "Pi3n-i",
237
+ "224": "Mlcf-n",
238
+ "225": "Pd-nsg",
239
+ "226": "Pp3msn",
240
+ "227": "Pd-fsl",
241
+ "228": "Agcmsayn",
242
+ "229": "Vap-sm",
243
+ "230": "Pi3n-a",
244
+ "231": "Pp3msd",
245
+ "232": "Ps3mpa",
246
+ "233": "Agpmsany",
247
+ "234": "Pi3m-a",
248
+ "235": "Vap-sn",
249
+ "236": "Mlomsl",
250
+ "237": "Vaa3p",
251
+ "238": "Ps3fsa",
252
+ "239": "Pi3m-g",
253
+ "240": "Mlonsg",
254
+ "241": "Ps3fsn",
255
+ "242": "Pi-nsl",
256
+ "243": "Mlcmsan",
257
+ "244": "Mlonsd",
258
+ "245": "Agcfsay",
259
+ "246": "Px-nsi",
260
+ "247": "Ps3nsa",
261
+ "248": "Vap-pn",
262
+ "249": "Mlcfsa",
263
+ "250": "Appnpny",
264
+ "251": "Ncnpi",
265
+ "252": "Pi-fpd",
266
+ "253": "Agpmsgn",
267
+ "254": "Px-fpa",
268
+ "255": "Mlcfsg",
269
+ "256": "Pp3fsa",
270
+ "257": "Mlonpd",
271
+ "258": "Pi-npl",
272
+ "259": "Vaf3s",
273
+ "260": "Mlcfsn",
274
+ "261": "Mlofsd",
275
+ "262": "Agcfsny",
276
+ "263": "Ps3fsi",
277
+ "264": "Agpnsdy",
278
+ "265": "Pp3fsd",
279
+ "266": "Mlonpn",
280
+ "267": "Agpnpny",
281
+ "268": "Mlonpa",
282
+ "269": "Mlonsa",
283
+ "270": "Ps3nsg",
284
+ "271": "Mlonpi",
285
+ "272": "Px-fsi",
286
+ "273": "Mlcmsay",
287
+ "274": "Vap-pm",
288
+ "275": "Ps3msan",
289
+ "276": "Mlcmsn",
290
+ "277": "Pi3n-l",
291
+ "278": "Mlomsan",
292
+ "279": "Mlcmsl",
293
+ "280": "Mlofsl",
294
+ "281": "Pp3fsl",
295
+ "282": "Ps3nsl",
296
+ "283": "Agpfpdy",
297
+ "284": "Pi-mpl",
298
+ "285": "Pd-mpg",
299
+ "286": "Ps3npg",
300
+ "287": "Ps3npa",
301
+ "288": "Agcmsgy",
302
+ "289": "Pi-npi",
303
+ "290": "Mcd",
304
+ "291": "Agcfsgy",
305
+ "292": "Px-npa",
306
+ "293": "Pp3-pd",
307
+ "294": "Mlcmsg",
308
+ "295": "Agsfsay",
309
+ "296": "Mlomsd",
310
+ "297": "Vap-sf",
311
+ "298": "Ps3msg",
312
+ "299": "Ps3mpi",
313
+ "300": "Pd-mpn",
314
+ "301": "Vaf3p",
315
+ "302": "Pp3-pl",
316
+ "303": "Mlompd",
317
+ "304": "Pd-mpl",
318
+ "305": "Ps3mpd",
319
+ "306": "Agpnpdy",
320
+ "307": "Pi-npd",
321
+ "308": "Agsfsgy",
322
+ "309": "Mlsf-a",
323
+ "310": "Px-npl",
324
+ "311": "Ps3mpg",
325
+ "312": "Pd-mpa",
326
+ "313": "Ps3npl",
327
+ "314": "Pi-fpi",
328
+ "315": "Agcnsny",
329
+ "316": "Agcnsay",
330
+ "317": "Agsmsayn",
331
+ "318": "Pp3fsn",
332
+ "319": "Agsfsny",
333
+ "320": "Npmpn",
334
+ "321": "Pd-mpi",
335
+ "322": "Agcfpny",
336
+ "323": "Agcfsiy",
337
+ "324": "Agcfpay",
338
+ "325": "Agcfpiy",
339
+ "326": "Px-msay",
340
+ "327": "Pp3-pg",
341
+ "328": "Pp3nsl",
342
+ "329": "Pp3fsg",
343
+ "330": "Mlcf-a",
344
+ "331": "Pi3m-n",
345
+ "332": "Agcmply",
346
+ "333": "Ps3mpn",
347
+ "334": "Mlonsn",
348
+ "335": "Agcfsly",
349
+ "336": "Ps3fpl",
350
+ "337": "Vmp-pn",
351
+ "338": "Pi-nsi",
352
+ "339": "Pp3msl",
353
+ "340": "Mlomsay",
354
+ "341": "Mlsm-g",
355
+ "342": "Mlsm-n",
356
+ "343": "Mlonsl",
357
+ "344": "Pp3mpn",
358
+ "345": "Px-nsg",
359
+ "346": "Aspmsgy",
360
+ "347": "Npfsi",
361
+ "348": "Pd-fpl",
362
+ "349": "Aspmsly",
363
+ "350": "Ps3npn",
364
+ "351": "Mlsf-n",
365
+ "352": "Pp3-pi",
366
+ "353": "Px-npg",
367
+ "354": "Npmsi",
368
+ "355": "Vmr1p",
369
+ "356": "Aspfsly",
370
+ "357": "Ps3msn",
371
+ "358": "Pd-npn",
372
+ "359": "Appfsay",
373
+ "360": "Appfpay",
374
+ "361": "Ps3npi",
375
+ "362": "Pd-fpi",
376
+ "363": "Pp3fpn",
377
+ "364": "Mlcf-d",
378
+ "365": "Pi-fsd",
379
+ "366": "Mlcfpg",
380
+ "367": "Pq3n-n",
381
+ "368": "Aspfsay",
382
+ "369": "Mlcnsa",
383
+ "370": "Mlompn",
384
+ "371": "Vmr1s",
385
+ "372": "Pp1-sn",
386
+ "373": "Pp3nsn",
387
+ "374": "Aspmsiy",
388
+ "375": "Aspfsgy",
389
+ "376": "Agsmsiy",
390
+ "377": "Ps1msn",
391
+ "378": "Mlcfsi",
392
+ "379": "Ps3msl",
393
+ "380": "Pd-npa",
394
+ "381": "Px--sd",
395
+ "382": "Aspmpiy",
396
+ "383": "Aspnpgy",
397
+ "384": "Mlsf-g",
398
+ "385": "Mlompg",
399
+ "386": "Agpnsgn",
400
+ "387": "Px--sg",
401
+ "388": "Mlofsi",
402
+ "389": "Pp1-sd",
403
+ "390": "Agsmsgy",
404
+ "391": "Ps3fpn",
405
+ "392": "Aspnsny",
406
+ "393": "Ps3msi",
407
+ "394": "Ps3fpa",
408
+ "395": "Pd-fsd",
409
+ "396": "Appfsly",
410
+ "397": "Px-npi",
411
+ "398": "Appmpn",
412
+ "399": "Px-nsl",
413
+ "400": "Mls",
414
+ "401": "Mlsn-n",
415
+ "402": "Ps3nsd",
416
+ "403": "Px-fsd",
417
+ "404": "Mlsn-g",
418
+ "405": "Pp3fsi",
419
+ "406": "Aspmsgn",
420
+ "407": "Mlcfsl",
421
+ "408": "Mlcfsd",
422
+ "409": "Px-fpl",
423
+ "410": "Pp1-pi",
424
+ "411": "Pp2-sd",
425
+ "412": "Var2s",
426
+ "413": "Vmr2s",
427
+ "414": "Mlcnsn",
428
+ "415": "Agcfpgy",
429
+ "416": "Pp3nsi",
430
+ "417": "Ps3nsi",
431
+ "418": "Mlcnsg",
432
+ "419": "Vma3s",
433
+ "420": "Pi3m-d",
434
+ "421": "Pi3n-d",
435
+ "422": "Pp3nsa",
436
+ "423": "Px-msd",
437
+ "424": "Ps3fsl",
438
+ "425": "Ps3fpd",
439
+ "426": "Aspfsiy",
440
+ "427": "Aspnpny",
441
+ "428": "Aspfpny",
442
+ "429": "Pq-npn",
443
+ "430": "Pd-msay",
444
+ "431": "Aspnsay",
445
+ "432": "Aspnsgy",
446
+ "433": "Aspmsdy",
447
+ "434": "Mlcmsd",
448
+ "435": "Mlonsi",
449
+ "436": "Mlompl",
450
+ "437": "Agcmpgy"
451
+ },
452
+ "initializer_range": 0.02,
453
+ "intermediate_size": 3072,
454
+ "label2id": {
455
+ "Agcfpay": 324,
456
+ "Agcfpgy": 415,
457
+ "Agcfpiy": 325,
458
+ "Agcfpny": 322,
459
+ "Agcfsay": 245,
460
+ "Agcfsgy": 291,
461
+ "Agcfsiy": 323,
462
+ "Agcfsly": 335,
463
+ "Agcfsny": 262,
464
+ "Agcmpgy": 437,
465
+ "Agcmply": 332,
466
+ "Agcmpny": 180,
467
+ "Agcmsayn": 228,
468
+ "Agcmsgy": 288,
469
+ "Agcmsiy": 87,
470
+ "Agcmsly": 158,
471
+ "Agcmsny": 210,
472
+ "Agcnsay": 316,
473
+ "Agcnsgy": 209,
474
+ "Agcnsny": 315,
475
+ "Agpfpay": 57,
476
+ "Agpfpdy": 283,
477
+ "Agpfpgy": 45,
478
+ "Agpfpiy": 64,
479
+ "Agpfply": 72,
480
+ "Agpfpny": 54,
481
+ "Agpfsay": 69,
482
+ "Agpfsdy": 203,
483
+ "Agpfsgy": 36,
484
+ "Agpfsiy": 84,
485
+ "Agpfsly": 115,
486
+ "Agpfsny": 25,
487
+ "Agpmpay": 50,
488
+ "Agpmpdy": 96,
489
+ "Agpmpgy": 82,
490
+ "Agpmpiy": 108,
491
+ "Agpmply": 71,
492
+ "Agpmpny": 176,
493
+ "Agpmsann": 133,
494
+ "Agpmsany": 233,
495
+ "Agpmsayn": 172,
496
+ "Agpmsayy": 213,
497
+ "Agpmsdy": 27,
498
+ "Agpmsgn": 253,
499
+ "Agpmsgy": 14,
500
+ "Agpmsiy": 121,
501
+ "Agpmsly": 28,
502
+ "Agpmsnn": 39,
503
+ "Agpmsny": 6,
504
+ "Agpnpay": 147,
505
+ "Agpnpdy": 306,
506
+ "Agpnpgy": 183,
507
+ "Agpnpiy": 97,
508
+ "Agpnply": 155,
509
+ "Agpnpny": 267,
510
+ "Agpnsay": 127,
511
+ "Agpnsdy": 264,
512
+ "Agpnsgn": 386,
513
+ "Agpnsgy": 83,
514
+ "Agpnsiy": 60,
515
+ "Agpnsly": 190,
516
+ "Agpnsny": 92,
517
+ "Agsfsay": 295,
518
+ "Agsfsgy": 308,
519
+ "Agsfsly": 114,
520
+ "Agsfsny": 319,
521
+ "Agsmsayn": 317,
522
+ "Agsmsgy": 390,
523
+ "Agsmsiy": 376,
524
+ "Agsmsly": 219,
525
+ "Agsmsny": 205,
526
+ "Appfpay": 360,
527
+ "Appfpny": 63,
528
+ "Appfsay": 359,
529
+ "Appfsly": 396,
530
+ "Appfsny": 100,
531
+ "Appmpn": 398,
532
+ "Appmpny": 106,
533
+ "Appmsnn": 33,
534
+ "Appnpny": 250,
535
+ "Appnsny": 156,
536
+ "Aspfpny": 428,
537
+ "Aspfsay": 368,
538
+ "Aspfsgy": 375,
539
+ "Aspfsiy": 426,
540
+ "Aspfsly": 356,
541
+ "Aspfsny": 49,
542
+ "Aspmpiy": 382,
543
+ "Aspmsdy": 433,
544
+ "Aspmsgn": 406,
545
+ "Aspmsgy": 346,
546
+ "Aspmsiy": 374,
547
+ "Aspmsly": 349,
548
+ "Aspnpgy": 383,
549
+ "Aspnpny": 427,
550
+ "Aspnsay": 431,
551
+ "Aspnsgy": 432,
552
+ "Aspnsny": 392,
553
+ "Cc": 10,
554
+ "Cs": 38,
555
+ "Mcd": 290,
556
+ "Mdc": 5,
557
+ "Mdm": 4,
558
+ "Mdo": 20,
559
+ "Mds": 220,
560
+ "Mlc": 98,
561
+ "Mlcf-a": 330,
562
+ "Mlcf-d": 364,
563
+ "Mlcf-g": 153,
564
+ "Mlcf-n": 224,
565
+ "Mlcfpg": 366,
566
+ "Mlcfsa": 249,
567
+ "Mlcfsd": 408,
568
+ "Mlcfsg": 255,
569
+ "Mlcfsi": 378,
570
+ "Mlcfsl": 407,
571
+ "Mlcfsn": 260,
572
+ "Mlcmsan": 243,
573
+ "Mlcmsay": 273,
574
+ "Mlcmsd": 434,
575
+ "Mlcmsg": 294,
576
+ "Mlcmsl": 279,
577
+ "Mlcmsn": 276,
578
+ "Mlcnsa": 369,
579
+ "Mlcnsg": 418,
580
+ "Mlcnsn": 414,
581
+ "Mlofpd": 166,
582
+ "Mlofsa": 144,
583
+ "Mlofsd": 261,
584
+ "Mlofsg": 135,
585
+ "Mlofsi": 388,
586
+ "Mlofsl": 280,
587
+ "Mlofsn": 136,
588
+ "Mlompd": 303,
589
+ "Mlompg": 385,
590
+ "Mlompl": 436,
591
+ "Mlompn": 370,
592
+ "Mlomsan": 278,
593
+ "Mlomsay": 340,
594
+ "Mlomsd": 296,
595
+ "Mlomsg": 170,
596
+ "Mlomsi": 211,
597
+ "Mlomsl": 236,
598
+ "Mlomsn": 217,
599
+ "Mlonpa": 268,
600
+ "Mlonpd": 257,
601
+ "Mlonpg": 189,
602
+ "Mlonpi": 271,
603
+ "Mlonpn": 266,
604
+ "Mlonsa": 269,
605
+ "Mlonsd": 244,
606
+ "Mlonsg": 240,
607
+ "Mlonsi": 435,
608
+ "Mlonsl": 343,
609
+ "Mlonsn": 334,
610
+ "Mls": 400,
611
+ "Mlsf-a": 309,
612
+ "Mlsf-g": 384,
613
+ "Mlsf-n": 351,
614
+ "Mlsm-g": 341,
615
+ "Mlsm-n": 342,
616
+ "Mlsn-g": 404,
617
+ "Mlsn-n": 401,
618
+ "Mrc": 103,
619
+ "Mro": 218,
620
+ "Ncfpa": 58,
621
+ "Ncfpd": 91,
622
+ "Ncfpg": 46,
623
+ "Ncfpi": 41,
624
+ "Ncfpl": 73,
625
+ "Ncfpn": 55,
626
+ "Ncfsa": 70,
627
+ "Ncfsd": 204,
628
+ "Ncfsg": 21,
629
+ "Ncfsi": 85,
630
+ "Ncfsl": 37,
631
+ "Ncfsn": 26,
632
+ "Ncmpa": 30,
633
+ "Ncmpd": 95,
634
+ "Ncmpg": 16,
635
+ "Ncmpi": 110,
636
+ "Ncmpl": 15,
637
+ "Ncmpn": 118,
638
+ "Ncmsan": 18,
639
+ "Ncmsay": 154,
640
+ "Ncmsd": 56,
641
+ "Ncmsg": 2,
642
+ "Ncmsi": 43,
643
+ "Ncmsl": 9,
644
+ "Ncmsn": 19,
645
+ "Ncnpa": 148,
646
+ "Ncnpd": 105,
647
+ "Ncnpg": 113,
648
+ "Ncnpi": 251,
649
+ "Ncnpl": 74,
650
+ "Ncnpn": 126,
651
+ "Ncnsa": 35,
652
+ "Ncnsd": 77,
653
+ "Ncnsg": 44,
654
+ "Ncnsi": 48,
655
+ "Ncnsl": 159,
656
+ "Ncnsn": 12,
657
+ "Npfsa": 164,
658
+ "Npfsd": 206,
659
+ "Npfsg": 51,
660
+ "Npfsi": 347,
661
+ "Npfsl": 160,
662
+ "Npfsn": 53,
663
+ "Npmpn": 320,
664
+ "Npmsan": 168,
665
+ "Npmsay": 167,
666
+ "Npmsd": 207,
667
+ "Npmsg": 3,
668
+ "Npmsi": 354,
669
+ "Npmsl": 47,
670
+ "Npmsn": 7,
671
+ "Npnsg": 216,
672
+ "Npnsl": 215,
673
+ "Npnsn": 221,
674
+ "Pd-fpa": 141,
675
+ "Pd-fpg": 130,
676
+ "Pd-fpi": 362,
677
+ "Pd-fpl": 348,
678
+ "Pd-fpn": 128,
679
+ "Pd-fsa": 165,
680
+ "Pd-fsd": 395,
681
+ "Pd-fsg": 150,
682
+ "Pd-fsi": 138,
683
+ "Pd-fsl": 227,
684
+ "Pd-fsn": 102,
685
+ "Pd-mpa": 312,
686
+ "Pd-mpg": 285,
687
+ "Pd-mpi": 321,
688
+ "Pd-mpl": 304,
689
+ "Pd-mpn": 300,
690
+ "Pd-msan": 146,
691
+ "Pd-msay": 430,
692
+ "Pd-msd": 131,
693
+ "Pd-msg": 42,
694
+ "Pd-msi": 152,
695
+ "Pd-msl": 101,
696
+ "Pd-msn": 34,
697
+ "Pd-npa": 380,
698
+ "Pd-npg": 112,
699
+ "Pd-npn": 358,
700
+ "Pd-nsa": 139,
701
+ "Pd-nsd": 104,
702
+ "Pd-nsg": 225,
703
+ "Pd-nsi": 75,
704
+ "Pd-nsl": 179,
705
+ "Pd-nsn": 122,
706
+ "Pi-fpa": 61,
707
+ "Pi-fpd": 252,
708
+ "Pi-fpg": 200,
709
+ "Pi-fpi": 314,
710
+ "Pi-fpl": 191,
711
+ "Pi-fpn": 59,
712
+ "Pi-fsa": 88,
713
+ "Pi-fsd": 365,
714
+ "Pi-fsg": 119,
715
+ "Pi-fsi": 137,
716
+ "Pi-fsl": 143,
717
+ "Pi-fsn": 116,
718
+ "Pi-mpa": 31,
719
+ "Pi-mpd": 93,
720
+ "Pi-mpg": 181,
721
+ "Pi-mpi": 161,
722
+ "Pi-mpl": 284,
723
+ "Pi-mpn": 175,
724
+ "Pi-msan": 80,
725
+ "Pi-msay": 52,
726
+ "Pi-msd": 173,
727
+ "Pi-msg": 117,
728
+ "Pi-msi": 162,
729
+ "Pi-msl": 67,
730
+ "Pi-msn": 169,
731
+ "Pi-npa": 196,
732
+ "Pi-npd": 307,
733
+ "Pi-npg": 182,
734
+ "Pi-npi": 289,
735
+ "Pi-npl": 258,
736
+ "Pi-npn": 188,
737
+ "Pi-nsa": 111,
738
+ "Pi-nsd": 99,
739
+ "Pi-nsg": 151,
740
+ "Pi-nsi": 338,
741
+ "Pi-nsl": 242,
742
+ "Pi-nsn": 157,
743
+ "Pi3m-a": 234,
744
+ "Pi3m-d": 420,
745
+ "Pi3m-g": 239,
746
+ "Pi3m-n": 331,
747
+ "Pi3n-a": 230,
748
+ "Pi3n-d": 421,
749
+ "Pi3n-g": 222,
750
+ "Pi3n-i": 223,
751
+ "Pi3n-l": 277,
752
+ "Pi3n-n": 68,
753
+ "Pp1-pi": 410,
754
+ "Pp1-sd": 389,
755
+ "Pp1-sn": 372,
756
+ "Pp2-sd": 411,
757
+ "Pp3-pa": 134,
758
+ "Pp3-pd": 293,
759
+ "Pp3-pg": 327,
760
+ "Pp3-pi": 352,
761
+ "Pp3-pl": 302,
762
+ "Pp3fpn": 363,
763
+ "Pp3fsa": 256,
764
+ "Pp3fsd": 265,
765
+ "Pp3fsg": 329,
766
+ "Pp3fsi": 405,
767
+ "Pp3fsl": 281,
768
+ "Pp3fsn": 318,
769
+ "Pp3mpn": 344,
770
+ "Pp3msa": 193,
771
+ "Pp3msd": 231,
772
+ "Pp3msg": 198,
773
+ "Pp3msi": 120,
774
+ "Pp3msl": 339,
775
+ "Pp3msn": 226,
776
+ "Pp3nsa": 422,
777
+ "Pp3nsi": 416,
778
+ "Pp3nsl": 328,
779
+ "Pp3nsn": 373,
780
+ "Pq-npn": 429,
781
+ "Pq3n-n": 367,
782
+ "Ps1msn": 377,
783
+ "Ps3fpa": 394,
784
+ "Ps3fpd": 425,
785
+ "Ps3fpg": 192,
786
+ "Ps3fpl": 336,
787
+ "Ps3fpn": 391,
788
+ "Ps3fsa": 238,
789
+ "Ps3fsd": 202,
790
+ "Ps3fsg": 201,
791
+ "Ps3fsi": 263,
792
+ "Ps3fsl": 424,
793
+ "Ps3fsn": 241,
794
+ "Ps3mpa": 232,
795
+ "Ps3mpd": 305,
796
+ "Ps3mpg": 311,
797
+ "Ps3mpi": 299,
798
+ "Ps3mpn": 333,
799
+ "Ps3msan": 275,
800
+ "Ps3msd": 174,
801
+ "Ps3msg": 298,
802
+ "Ps3msi": 393,
803
+ "Ps3msl": 379,
804
+ "Ps3msn": 357,
805
+ "Ps3npa": 287,
806
+ "Ps3npg": 286,
807
+ "Ps3npi": 361,
808
+ "Ps3npl": 313,
809
+ "Ps3npn": 350,
810
+ "Ps3nsa": 247,
811
+ "Ps3nsd": 402,
812
+ "Ps3nsg": 270,
813
+ "Ps3nsi": 417,
814
+ "Ps3nsl": 282,
815
+ "Ps3nsn": 187,
816
+ "Px--sa": 23,
817
+ "Px--sd": 381,
818
+ "Px--sg": 387,
819
+ "Px-fpa": 254,
820
+ "Px-fpd": 90,
821
+ "Px-fpg": 197,
822
+ "Px-fpl": 409,
823
+ "Px-fsa": 123,
824
+ "Px-fsd": 403,
825
+ "Px-fsg": 195,
826
+ "Px-fsi": 272,
827
+ "Px-fsl": 145,
828
+ "Px-mpa": 171,
829
+ "Px-mpd": 94,
830
+ "Px-mpg": 199,
831
+ "Px-mpi": 109,
832
+ "Px-msan": 194,
833
+ "Px-msay": 326,
834
+ "Px-msd": 423,
835
+ "Px-msg": 142,
836
+ "Px-msi": 208,
837
+ "Px-msl": 177,
838
+ "Px-npa": 292,
839
+ "Px-npg": 353,
840
+ "Px-npi": 397,
841
+ "Px-npl": 310,
842
+ "Px-nsa": 184,
843
+ "Px-nsg": 345,
844
+ "Px-nsi": 246,
845
+ "Px-nsl": 399,
846
+ "Qo": 66,
847
+ "Qq": 65,
848
+ "Qz": 86,
849
+ "Rgc": 140,
850
+ "Rgp": 24,
851
+ "Rgs": 178,
852
+ "Rr": 149,
853
+ "Sa": 29,
854
+ "Sd": 89,
855
+ "Sg": 13,
856
+ "Si": 40,
857
+ "Sl": 8,
858
+ "Vaa3p": 237,
859
+ "Vaa3s": 124,
860
+ "Vaf3p": 301,
861
+ "Vaf3s": 259,
862
+ "Van": 107,
863
+ "Vap-pf": 132,
864
+ "Vap-pm": 274,
865
+ "Vap-pn": 248,
866
+ "Vap-sf": 297,
867
+ "Vap-sm": 229,
868
+ "Vap-sn": 235,
869
+ "Var1s": 212,
870
+ "Var2s": 412,
871
+ "Var3p": 62,
872
+ "Var3s": 32,
873
+ "Vma3s": 419,
874
+ "Vmf3p": 186,
875
+ "Vmf3s": 185,
876
+ "Vmn": 79,
877
+ "Vmp-pf": 129,
878
+ "Vmp-pm": 214,
879
+ "Vmp-pn": 337,
880
+ "Vmp-sf": 125,
881
+ "Vmp-sm": 81,
882
+ "Vmp-sn": 76,
883
+ "Vmr1p": 355,
884
+ "Vmr1s": 371,
885
+ "Vmr2s": 413,
886
+ "Vmr3p": 17,
887
+ "Vmr3s": 22,
888
+ "X": 78,
889
+ "Xf": 0,
890
+ "Xw": 163,
891
+ "Y": 1,
892
+ "Z": 11
893
+ },
894
+ "layer_norm_eps": 1e-12,
895
+ "max_position_embeddings": 512,
896
+ "model_type": "electra",
897
+ "num_attention_heads": 12,
898
+ "num_hidden_layers": 12,
899
+ "pad_token_id": 0,
900
+ "position_embedding_type": "absolute",
901
+ "summary_activation": "gelu",
902
+ "summary_last_dropout": 0.1,
903
+ "summary_type": "first",
904
+ "summary_use_proj": true,
905
+ "torch_dtype": "float32",
906
+ "transformers_version": "4.36.2",
907
+ "type_vocab_size": 2,
908
+ "use_cache": true,
909
+ "vocab_size": 32000
910
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a1a54980e8ba060acc170be7c6cb08140de2613bdf8eec5a0ecb3eaa6f020cdb
3
+ size 441478248
special_tokens_map.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": "[CLS]",
3
+ "mask_token": "[MASK]",
4
+ "pad_token": "[PAD]",
5
+ "sep_token": "[SEP]",
6
+ "unk_token": "[UNK]"
7
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "[PAD]",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "[UNK]",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "[CLS]",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "[SEP]",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "4": {
36
+ "content": "[MASK]",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "clean_up_tokenization_spaces": true,
45
+ "cls_token": "[CLS]",
46
+ "do_basic_tokenize": true,
47
+ "do_lower_case": false,
48
+ "mask_token": "[MASK]",
49
+ "max_len": 512,
50
+ "model_max_length": 512,
51
+ "never_split": null,
52
+ "pad_token": "[PAD]",
53
+ "sep_token": "[SEP]",
54
+ "strip_accents": false,
55
+ "tokenize_chinese_chars": true,
56
+ "tokenizer_class": "ElectraTokenizer",
57
+ "unk_token": "[UNK]"
58
+ }
vocab.txt ADDED
The diff for this file is too large to render. See raw diff