Hugging Face's logo

AI4Protein
/
deep_bpe_400

Feature Extraction
Transformers
PyTorch
roformer
fill-mask
Model card Files
xet
Community
3
GinnM commited on
Commit
016c3b0
·
1 Parent(s): b04b5df

Upload tokenizer

Browse files
Files changed (3) hide show
  1. special_tokens_map.json +9 -0
  2. tokenizer.json +916 -0
  3. tokenizer_config.json +5 -0
special_tokens_map.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<cls>",
3
+ "cls_token": "<cls>",
4
+ "eos_token": "<sep>",
5
+ "mask_token": "<mask>",
6
+ "pad_token": "<pad>",
7
+ "sep_token": "<sep>",
8
+ "unk_token": "<unk>"
9
+ }
tokenizer.json ADDED
@@ -0,0 +1,916 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "1.0",
3
+ "truncation": null,
4
+ "padding": null,
5
+ "added_tokens": [
6
+ {
7
+ "id": 0,
8
+ "content": "<pad>",
9
+ "single_word": false,
10
+ "lstrip": false,
11
+ "rstrip": false,
12
+ "normalized": false,
13
+ "special": true
14
+ },
15
+ {
16
+ "id": 1,
17
+ "content": "<cls>",
18
+ "single_word": false,
19
+ "lstrip": false,
20
+ "rstrip": false,
21
+ "normalized": false,
22
+ "special": true
23
+ },
24
+ {
25
+ "id": 2,
26
+ "content": "<sep>",
27
+ "single_word": false,
28
+ "lstrip": false,
29
+ "rstrip": false,
30
+ "normalized": false,
31
+ "special": true
32
+ },
33
+ {
34
+ "id": 3,
35
+ "content": "<unk>",
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": null,
53
+ "pre_tokenizer": null,
54
+ "post_processor": {
55
+ "type": "TemplateProcessing",
56
+ "single": [
57
+ {
58
+ "SpecialToken": {
59
+ "id": "<cls>",
60
+ "type_id": 0
61
+ }
62
+ },
63
+ {
64
+ "Sequence": {
65
+ "id": "A",
66
+ "type_id": 0
67
+ }
68
+ },
69
+ {
70
+ "SpecialToken": {
71
+ "id": "<sep>",
72
+ "type_id": 0
73
+ }
74
+ }
75
+ ],
76
+ "pair": [
77
+ {
78
+ "SpecialToken": {
79
+ "id": "<cls>",
80
+ "type_id": 0
81
+ }
82
+ },
83
+ {
84
+ "Sequence": {
85
+ "id": "A",
86
+ "type_id": 0
87
+ }
88
+ },
89
+ {
90
+ "SpecialToken": {
91
+ "id": "<sep>",
92
+ "type_id": 0
93
+ }
94
+ },
95
+ {
96
+ "Sequence": {
97
+ "id": "B",
98
+ "type_id": 1
99
+ }
100
+ },
101
+ {
102
+ "SpecialToken": {
103
+ "id": "<sep>",
104
+ "type_id": 1
105
+ }
106
+ }
107
+ ],
108
+ "special_tokens": {
109
+ "<cls>": {
110
+ "id": "<cls>",
111
+ "ids": [
112
+ 1
113
+ ],
114
+ "tokens": [
115
+ "<cls>"
116
+ ]
117
+ },
118
+ "<sep>": {
119
+ "id": "<sep>",
120
+ "ids": [
121
+ 2
122
+ ],
123
+ "tokens": [
124
+ "<sep>"
125
+ ]
126
+ }
127
+ }
128
+ },
129
+ "decoder": {
130
+ "type": "BPEDecoder",
131
+ "suffix": "</w>"
132
+ },
133
+ "model": {
134
+ "type": "BPE",
135
+ "dropout": null,
136
+ "unk_token": null,
137
+ "continuing_subword_prefix": null,
138
+ "end_of_word_suffix": null,
139
+ "fuse_unk": false,
140
+ "byte_fallback": false,
141
+ "vocab": {
142
+ "<pad>": 0,
143
+ "<cls>": 1,
144
+ "<sep>": 2,
145
+ "<unk>": 3,
146
+ "<mask>": 4,
147
+ "A": 5,
148
+ "B": 6,
149
+ "C": 7,
150
+ "D": 8,
151
+ "E": 9,
152
+ "F": 10,
153
+ "G": 11,
154
+ "H": 12,
155
+ "I": 13,
156
+ "K": 14,
157
+ "L": 15,
158
+ "M": 16,
159
+ "N": 17,
160
+ "O": 18,
161
+ "P": 19,
162
+ "Q": 20,
163
+ "R": 21,
164
+ "S": 22,
165
+ "T": 23,
166
+ "U": 24,
167
+ "V": 25,
168
+ "W": 26,
169
+ "X": 27,
170
+ "Y": 28,
171
+ "Z": 29,
172
+ "AA": 30,
173
+ "LL": 31,
174
+ "SS": 32,
175
+ "LA": 33,
176
+ "LS": 34,
177
+ "GG": 35,
178
+ "LV": 36,
179
+ "LR": 37,
180
+ "LE": 38,
181
+ "LG": 39,
182
+ "LT": 40,
183
+ "AV": 41,
184
+ "AG": 42,
185
+ "LP": 43,
186
+ "LD": 44,
187
+ "LI": 45,
188
+ "AS": 46,
189
+ "EE": 47,
190
+ "RR": 48,
191
+ "LK": 49,
192
+ "VV": 50,
193
+ "AT": 51,
194
+ "AE": 52,
195
+ "SG": 53,
196
+ "DG": 54,
197
+ "AR": 55,
198
+ "KK": 56,
199
+ "AI": 57,
200
+ "AP": 58,
201
+ "TT": 59,
202
+ "LN": 60,
203
+ "LF": 61,
204
+ "LQ": 62,
205
+ "AD": 63,
206
+ "SV": 64,
207
+ "ST": 65,
208
+ "SI": 66,
209
+ "SP": 67,
210
+ "DE": 68,
211
+ "TG": 69,
212
+ "TV": 70,
213
+ "SR": 71,
214
+ "SE": 72,
215
+ "DV": 73,
216
+ "PP": 74,
217
+ "KE": 75,
218
+ "KI": 76,
219
+ "DI": 77,
220
+ "GV": 78,
221
+ "AQ": 79,
222
+ "AF": 80,
223
+ "RV": 81,
224
+ "TI": 82,
225
+ "NI": 83,
226
+ "PV": 84,
227
+ "RE": 85,
228
+ "SF": 86,
229
+ "KV": 87,
230
+ "NG": 88,
231
+ "LY": 89,
232
+ "RI": 90,
233
+ "PG": 91,
234
+ "RG": 92,
235
+ "DD": 93,
236
+ "PE": 94,
237
+ "AK": 95,
238
+ "NV": 96,
239
+ "SD": 97,
240
+ "TE": 98,
241
+ "QQ": 99,
242
+ "NE": 100,
243
+ "SK": 101,
244
+ "LH": 102,
245
+ "SQ": 103,
246
+ "GI": 104,
247
+ "FV": 105,
248
+ "TP": 106,
249
+ "GE": 107,
250
+ "SN": 108,
251
+ "FI": 109,
252
+ "AY": 110,
253
+ "AN": 111,
254
+ "QV": 112,
255
+ "RK": 113,
256
+ "SY": 114,
257
+ "IV": 115,
258
+ "FE": 116,
259
+ "DP": 117,
260
+ "RP": 118,
261
+ "IE": 119,
262
+ "QE": 120,
263
+ "FG": 121,
264
+ "LM": 122,
265
+ "NN": 123,
266
+ "KG": 124,
267
+ "RT": 125,
268
+ "RD": 126,
269
+ "RQ": 127,
270
+ "II": 128,
271
+ "TD": 129,
272
+ "AH": 130,
273
+ "AM": 131,
274
+ "YV": 132,
275
+ "RF": 133,
276
+ "EV": 134,
277
+ "KN": 135,
278
+ "KD": 136,
279
+ "IG": 137,
280
+ "LC": 138,
281
+ "KT": 139,
282
+ "KP": 140,
283
+ "RY": 141,
284
+ "SH": 142,
285
+ "QG": 143,
286
+ "NP": 144,
287
+ "QI": 145,
288
+ "FD": 146,
289
+ "YG": 147,
290
+ "RN": 148,
291
+ "YE": 149,
292
+ "TF": 150,
293
+ "MV": 151,
294
+ "RH": 152,
295
+ "LW": 153,
296
+ "QP": 154,
297
+ "TN": 155,
298
+ "DF": 156,
299
+ "KQ": 157,
300
+ "YI": 158,
301
+ "ME": 159,
302
+ "VE": 160,
303
+ "LAA": 161,
304
+ "PD": 162,
305
+ "FF": 163,
306
+ "PI": 164,
307
+ "AC": 165,
308
+ "ND": 166,
309
+ "VG": 167,
310
+ "TY": 168,
311
+ "TQ": 169,
312
+ "SM": 170,
313
+ "AW": 171,
314
+ "KY": 172,
315
+ "HG": 173,
316
+ "SC": 174,
317
+ "HV": 175,
318
+ "ID": 176,
319
+ "KF": 177,
320
+ "EG": 178,
321
+ "KR": 179,
322
+ "EI": 180,
323
+ "MG": 181,
324
+ "TR": 182,
325
+ "NF": 183,
326
+ "SW": 184,
327
+ "TK": 185,
328
+ "YD": 186,
329
+ "NQ": 187,
330
+ "ALL": 188,
331
+ "MI": 189,
332
+ "PT": 190,
333
+ "NY": 191,
334
+ "QD": 192,
335
+ "GD": 193,
336
+ "NT": 194,
337
+ "HE": 195,
338
+ "VI": 196,
339
+ "PF": 197,
340
+ "ALA": 198,
341
+ "RS": 199,
342
+ "HI": 200,
343
+ "CG": 201,
344
+ "TS": 202,
345
+ "HP": 203,
346
+ "YF": 204,
347
+ "NK": 205,
348
+ "VD": 206,
349
+ "RW": 207,
350
+ "PQ": 208,
351
+ "RM": 209,
352
+ "LSS": 210,
353
+ "ED": 211,
354
+ "MD": 212,
355
+ "TH": 213,
356
+ "YY": 214,
357
+ "RC": 215,
358
+ "QK": 216,
359
+ "QF": 217,
360
+ "QT": 218,
361
+ "GF": 219,
362
+ "IP": 220,
363
+ "MP": 221,
364
+ "VP": 222,
365
+ "SA": 223,
366
+ "IF": 224,
367
+ "EK": 225,
368
+ "HD": 226,
369
+ "QR": 227,
370
+ "SLL": 228,
371
+ "QN": 229,
372
+ "HF": 230,
373
+ "VF": 231,
374
+ "YP": 232,
375
+ "GT": 233,
376
+ "DT": 234,
377
+ "GP": 235,
378
+ "MK": 236,
379
+ "CV": 237,
380
+ "MT": 238,
381
+ "AGG": 239,
382
+ "VT": 240,
383
+ "DY": 241,
384
+ "ALR": 242,
385
+ "ALV": 243,
386
+ "NR": 244,
387
+ "DK": 245,
388
+ "IT": 246,
389
+ "GK": 247,
390
+ "FP": 248,
391
+ "QY": 249,
392
+ "GR": 250,
393
+ "DR": 251,
394
+ "SAA": 252,
395
+ "ER": 253,
396
+ "LGG": 254,
397
+ "ALS": 255,
398
+ "EP": 256,
399
+ "ALG": 257,
400
+ "FT": 258,
401
+ "ASS": 259,
402
+ "HH": 260,
403
+ "ET": 261,
404
+ "EQ": 262,
405
+ "DN": 263,
406
+ "IK": 264,
407
+ "WV": 265,
408
+ "YT": 266,
409
+ "IN": 267,
410
+ "SLS": 268,
411
+ "MN": 269,
412
+ "QH": 270,
413
+ "ALE": 271,
414
+ "GN": 272,
415
+ "MF": 273,
416
+ "GY": 274,
417
+ "IY": 275,
418
+ "KH": 276,
419
+ "EN": 277,
420
+ "EF": 278,
421
+ "DQ": 279,
422
+ "PR": 280,
423
+ "KS": 281,
424
+ "GQ": 282,
425
+ "AAV": 283,
426
+ "YN": 284,
427
+ "ALT": 285,
428
+ "ARR": 286,
429
+ "AAG": 287,
430
+ "PN": 288,
431
+ "SLA": 289,
432
+ "RLL": 290,
433
+ "FY": 291,
434
+ "FN": 292,
435
+ "PK": 293,
436
+ "MQ": 294,
437
+ "ALP": 295,
438
+ "MR": 296,
439
+ "ALD": 297,
440
+ "CI": 298,
441
+ "SGG": 299,
442
+ "WI": 300,
443
+ "HT": 301,
444
+ "CE": 302,
445
+ "WG": 303,
446
+ "YK": 304,
447
+ "YQ": 305,
448
+ "CP": 306,
449
+ "HR": 307,
450
+ "VR": 308,
451
+ "SAS": 309,
452
+ "WE": 310,
453
+ "HQ": 311,
454
+ "ALI": 312,
455
+ "YR": 313,
456
+ "IQ": 314,
457
+ "FK": 315,
458
+ "VK": 316,
459
+ "ALK": 317,
460
+ "ADG": 318,
461
+ "SLV": 319,
462
+ "VN": 320,
463
+ "CD": 321,
464
+ "XX": 322,
465
+ "AEE": 323,
466
+ "SLP": 324,
467
+ "LAG": 325,
468
+ "IR": 326,
469
+ "VQ": 327,
470
+ "FQ": 328,
471
+ "TW": 329,
472
+ "SLT": 330,
473
+ "SLG": 331,
474
+ "TC": 332,
475
+ "PY": 333,
476
+ "SLR": 334,
477
+ "LAE": 335,
478
+ "LLV": 336,
479
+ "HY": 337,
480
+ "LLG": 338,
481
+ "LAV": 339,
482
+ "FR": 340,
483
+ "SLE": 341,
484
+ "KM": 342,
485
+ "HN": 343,
486
+ "SAV": 344,
487
+ "SAG": 345,
488
+ "DW": 346,
489
+ "LLE": 347,
490
+ "ALQ": 348,
491
+ "TM": 349,
492
+ "SLI": 350,
493
+ "AAE": 351,
494
+ "SAT": 352,
495
+ "CF": 353,
496
+ "DH": 354,
497
+ "SLD": 355,
498
+ "HK": 356,
499
+ "STT": 357,
500
+ "SLK": 358,
501
+ "EY": 359,
502
+ "GGG": 360,
503
+ "VY": 361,
504
+ "KW": 362,
505
+ "LSG": 363,
506
+ "AAAA": 364,
507
+ "PH": 365,
508
+ "ALF": 366,
509
+ "MM": 367,
510
+ "SAP": 368,
511
+ "SVV": 369,
512
+ "LEE": 370,
513
+ "RAA": 371,
514
+ "WD": 372,
515
+ "NH": 373,
516
+ "CT": 374,
517
+ "MY": 375,
518
+ "ALN": 376,
519
+ "AVV": 377,
520
+ "RLR": 378,
521
+ "ASG": 379,
522
+ "WT": 380,
523
+ "SEE": 381,
524
+ "SRR": 382,
525
+ "DM": 383,
526
+ "LSV": 384,
527
+ "KC": 385,
528
+ "TAA": 386,
529
+ "WQ": 387,
530
+ "IH": 388,
531
+ "AKK": 389,
532
+ "LSE": 390,
533
+ "SSG": 391,
534
+ "LVV": 392,
535
+ "AGV": 393,
536
+ "WP": 394,
537
+ "LLD": 395,
538
+ "SLF": 396,
539
+ "WF": 397,
540
+ "WR": 398,
541
+ "GH": 399
542
+ },
543
+ "merges": [
544
+ "A A",
545
+ "L L",
546
+ "S S",
547
+ "L A",
548
+ "L S",
549
+ "G G",
550
+ "L V",
551
+ "L R",
552
+ "L E",
553
+ "L G",
554
+ "L T",
555
+ "A V",
556
+ "A G",
557
+ "L P",
558
+ "L D",
559
+ "L I",
560
+ "A S",
561
+ "E E",
562
+ "R R",
563
+ "L K",
564
+ "V V",
565
+ "A T",
566
+ "A E",
567
+ "S G",
568
+ "D G",
569
+ "A R",
570
+ "K K",
571
+ "A I",
572
+ "A P",
573
+ "T T",
574
+ "L N",
575
+ "L F",
576
+ "L Q",
577
+ "A D",
578
+ "S V",
579
+ "S T",
580
+ "S I",
581
+ "S P",
582
+ "D E",
583
+ "T G",
584
+ "T V",
585
+ "S R",
586
+ "S E",
587
+ "D V",
588
+ "P P",
589
+ "K E",
590
+ "K I",
591
+ "D I",
592
+ "G V",
593
+ "A Q",
594
+ "A F",
595
+ "R V",
596
+ "T I",
597
+ "N I",
598
+ "P V",
599
+ "R E",
600
+ "S F",
601
+ "K V",
602
+ "N G",
603
+ "L Y",
604
+ "R I",
605
+ "P G",
606
+ "R G",
607
+ "D D",
608
+ "P E",
609
+ "A K",
610
+ "N V",
611
+ "S D",
612
+ "T E",
613
+ "Q Q",
614
+ "N E",
615
+ "S K",
616
+ "L H",
617
+ "S Q",
618
+ "G I",
619
+ "F V",
620
+ "T P",
621
+ "G E",
622
+ "S N",
623
+ "F I",
624
+ "A Y",
625
+ "A N",
626
+ "Q V",
627
+ "R K",
628
+ "S Y",
629
+ "I V",
630
+ "F E",
631
+ "D P",
632
+ "R P",
633
+ "I E",
634
+ "Q E",
635
+ "F G",
636
+ "L M",
637
+ "N N",
638
+ "K G",
639
+ "R T",
640
+ "R D",
641
+ "R Q",
642
+ "I I",
643
+ "T D",
644
+ "A H",
645
+ "A M",
646
+ "Y V",
647
+ "R F",
648
+ "E V",
649
+ "K N",
650
+ "K D",
651
+ "I G",
652
+ "L C",
653
+ "K T",
654
+ "K P",
655
+ "R Y",
656
+ "S H",
657
+ "Q G",
658
+ "N P",
659
+ "Q I",
660
+ "F D",
661
+ "Y G",
662
+ "R N",
663
+ "Y E",
664
+ "T F",
665
+ "M V",
666
+ "R H",
667
+ "L W",
668
+ "Q P",
669
+ "T N",
670
+ "D F",
671
+ "K Q",
672
+ "Y I",
673
+ "M E",
674
+ "V E",
675
+ "L AA",
676
+ "P D",
677
+ "F F",
678
+ "P I",
679
+ "A C",
680
+ "N D",
681
+ "V G",
682
+ "T Y",
683
+ "T Q",
684
+ "S M",
685
+ "A W",
686
+ "K Y",
687
+ "H G",
688
+ "S C",
689
+ "H V",
690
+ "I D",
691
+ "K F",
692
+ "E G",
693
+ "K R",
694
+ "E I",
695
+ "M G",
696
+ "T R",
697
+ "N F",
698
+ "S W",
699
+ "T K",
700
+ "Y D",
701
+ "N Q",
702
+ "A LL",
703
+ "M I",
704
+ "P T",
705
+ "N Y",
706
+ "Q D",
707
+ "G D",
708
+ "N T",
709
+ "H E",
710
+ "V I",
711
+ "P F",
712
+ "A LA",
713
+ "R S",
714
+ "H I",
715
+ "C G",
716
+ "T S",
717
+ "H P",
718
+ "Y F",
719
+ "N K",
720
+ "V D",
721
+ "R W",
722
+ "P Q",
723
+ "R M",
724
+ "L SS",
725
+ "E D",
726
+ "M D",
727
+ "T H",
728
+ "Y Y",
729
+ "R C",
730
+ "Q K",
731
+ "Q F",
732
+ "Q T",
733
+ "G F",
734
+ "I P",
735
+ "M P",
736
+ "V P",
737
+ "S A",
738
+ "I F",
739
+ "E K",
740
+ "H D",
741
+ "Q R",
742
+ "S LL",
743
+ "Q N",
744
+ "H F",
745
+ "V F",
746
+ "Y P",
747
+ "G T",
748
+ "D T",
749
+ "G P",
750
+ "M K",
751
+ "C V",
752
+ "M T",
753
+ "A GG",
754
+ "V T",
755
+ "D Y",
756
+ "A LR",
757
+ "A LV",
758
+ "N R",
759
+ "D K",
760
+ "I T",
761
+ "G K",
762
+ "F P",
763
+ "Q Y",
764
+ "G R",
765
+ "D R",
766
+ "S AA",
767
+ "E R",
768
+ "L GG",
769
+ "A LS",
770
+ "E P",
771
+ "A LG",
772
+ "F T",
773
+ "A SS",
774
+ "H H",
775
+ "E T",
776
+ "E Q",
777
+ "D N",
778
+ "I K",
779
+ "W V",
780
+ "Y T",
781
+ "I N",
782
+ "S LS",
783
+ "M N",
784
+ "Q H",
785
+ "A LE",
786
+ "G N",
787
+ "M F",
788
+ "G Y",
789
+ "I Y",
790
+ "K H",
791
+ "E N",
792
+ "E F",
793
+ "D Q",
794
+ "P R",
795
+ "K S",
796
+ "G Q",
797
+ "AA V",
798
+ "Y N",
799
+ "A LT",
800
+ "A RR",
801
+ "AA G",
802
+ "P N",
803
+ "S LA",
804
+ "R LL",
805
+ "F Y",
806
+ "F N",
807
+ "P K",
808
+ "M Q",
809
+ "A LP",
810
+ "M R",
811
+ "A LD",
812
+ "C I",
813
+ "S GG",
814
+ "W I",
815
+ "H T",
816
+ "C E",
817
+ "W G",
818
+ "Y K",
819
+ "Y Q",
820
+ "C P",
821
+ "H R",
822
+ "V R",
823
+ "S AS",
824
+ "W E",
825
+ "H Q",
826
+ "A LI",
827
+ "Y R",
828
+ "I Q",
829
+ "F K",
830
+ "V K",
831
+ "A LK",
832
+ "A DG",
833
+ "S LV",
834
+ "V N",
835
+ "C D",
836
+ "X X",
837
+ "A EE",
838
+ "S LP",
839
+ "LA G",
840
+ "I R",
841
+ "V Q",
842
+ "F Q",
843
+ "T W",
844
+ "S LT",
845
+ "S LG",
846
+ "T C",
847
+ "P Y",
848
+ "S LR",
849
+ "LA E",
850
+ "LL V",
851
+ "H Y",
852
+ "LL G",
853
+ "LA V",
854
+ "F R",
855
+ "S LE",
856
+ "K M",
857
+ "H N",
858
+ "S AV",
859
+ "S AG",
860
+ "D W",
861
+ "LL E",
862
+ "A LQ",
863
+ "T M",
864
+ "S LI",
865
+ "AA E",
866
+ "S AT",
867
+ "C F",
868
+ "D H",
869
+ "S LD",
870
+ "H K",
871
+ "S TT",
872
+ "S LK",
873
+ "E Y",
874
+ "GG G",
875
+ "V Y",
876
+ "K W",
877
+ "LS G",
878
+ "AA AA",
879
+ "P H",
880
+ "A LF",
881
+ "M M",
882
+ "S AP",
883
+ "S VV",
884
+ "LE E",
885
+ "R AA",
886
+ "W D",
887
+ "N H",
888
+ "C T",
889
+ "M Y",
890
+ "A LN",
891
+ "AV V",
892
+ "R LR",
893
+ "AS G",
894
+ "W T",
895
+ "S EE",
896
+ "S RR",
897
+ "D M",
898
+ "LS V",
899
+ "K C",
900
+ "T AA",
901
+ "W Q",
902
+ "I H",
903
+ "A KK",
904
+ "LS E",
905
+ "SS G",
906
+ "LV V",
907
+ "AG V",
908
+ "W P",
909
+ "LL D",
910
+ "S LF",
911
+ "W F",
912
+ "W R",
913
+ "G H"
914
+ ]
915
+ }
916
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "clean_up_tokenization_spaces": true,
3
+ "model_max_length": 1000000000000000019884624838656,
4
+ "tokenizer_class": "PreTrainedTokenizerFast"
5
+ }