panyutsriwirote commited on
Commit
c1e2cba
1 Parent(s): 0f69c49
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
added_tokens.json ADDED
@@ -0,0 +1,837 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "‼": 248733,
3
+ "⁉": 248850,
4
+ "⌛": 248707,
5
+ "⌨": 248450,
6
+ "⏏": 248572,
7
+ "⏫": 248890,
8
+ "⏭": 249212,
9
+ "⏮": 248920,
10
+ "⏯": 248576,
11
+ "⏲": 248922,
12
+ "⏸": 249193,
13
+ "⏹": 249213,
14
+ "Ⓜ": 248957,
15
+ "♉": 248638,
16
+ "♊": 248741,
17
+ "♌": 249062,
18
+ "♍": 248609,
19
+ "♎": 248611,
20
+ "♏": 249129,
21
+ "♑": 248994,
22
+ "♓": 248790,
23
+ "♾": 248725,
24
+ "♿": 248454,
25
+ "⚒": 248845,
26
+ "⚖": 248534,
27
+ "⚗": 248943,
28
+ "⚙": 248676,
29
+ "⚛": 248761,
30
+ "⚧": 248971,
31
+ "⚰": 248930,
32
+ "⚱": 249010,
33
+ "⛈": 249232,
34
+ "⛎": 248516,
35
+ "⛏": 248490,
36
+ "⛑": 248575,
37
+ "⛩": 248953,
38
+ "⛪": 249120,
39
+ "⛲": 249141,
40
+ "⛴": 248779,
41
+ "⛷": 248649,
42
+ "⛸": 248693,
43
+ "⛹": 249260,
44
+ "⛽": 249248,
45
+ "✝": 248664,
46
+ "➗": 248803,
47
+ "➰": 248665,
48
+ "➿": 248458,
49
+ "⬛": 249000,
50
+ "⬜": 248821,
51
+ "〽": 248791,
52
+ "㊗": 248521,
53
+ "㊙": 249060,
54
+ "🀄": 248929,
55
+ "🃏": 248468,
56
+ "🅿": 248755,
57
+ "🆎": 248619,
58
+ "🆒": 248798,
59
+ "🆕": 248655,
60
+ "🆖": 248837,
61
+ "🆗": 248917,
62
+ "🆙": 248978,
63
+ "🆚": 248972,
64
+ "🈁": 248495,
65
+ "🈂": 248844,
66
+ "🈚": 249053,
67
+ "🈯": 248568,
68
+ "🈲": 248512,
69
+ "🈳": 248785,
70
+ "🈴": 249003,
71
+ "🈵": 248752,
72
+ "🈶": 248502,
73
+ "🈷": 249153,
74
+ "🈸": 249079,
75
+ "🈹": 248690,
76
+ "🈺": 248472,
77
+ "🉐": 249070,
78
+ "🉑": 249224,
79
+ "🌁": 248694,
80
+ "🌂": 248584,
81
+ "🌃": 248739,
82
+ "🌇": 248528,
83
+ "🌉": 248964,
84
+ "🌌": 248587,
85
+ "🌑": 249069,
86
+ "🌒": 249187,
87
+ "🌓": 248598,
88
+ "🌔": 248514,
89
+ "🌕": 248452,
90
+ "🌖": 249242,
91
+ "🌗": 248478,
92
+ "🌘": 248782,
93
+ "🌚": 249190,
94
+ "🌛": 248984,
95
+ "🌜": 248635,
96
+ "🌠": 248623,
97
+ "🌡": 248904,
98
+ "🌤": 248812,
99
+ "🌥": 248881,
100
+ "🌦": 249172,
101
+ "🌧": 248666,
102
+ "🌨": 248730,
103
+ "🌩": 248795,
104
+ "🌪": 249022,
105
+ "🌫": 248734,
106
+ "🌬": 248530,
107
+ "🌭": 248868,
108
+ "🌮": 248976,
109
+ "🌯": 248716,
110
+ "🌰": 248872,
111
+ "🌵": 249081,
112
+ "🍈": 249173,
113
+ "🍐": 248933,
114
+ "🍖": 248742,
115
+ "🍗": 248433,
116
+ "🍘": 248647,
117
+ "🍙": 249057,
118
+ "🍛": 248869,
119
+ "🍜": 248979,
120
+ "🍝": 249087,
121
+ "🍞": 248697,
122
+ "🍟": 248748,
123
+ "🍠": 248735,
124
+ "🍡": 248591,
125
+ "🍢": 248833,
126
+ "🍣": 248744,
127
+ "🍤": 248777,
128
+ "🍥": 248722,
129
+ "🍧": 248692,
130
+ "🍩": 248935,
131
+ "🍭": 248923,
132
+ "🍮": 249067,
133
+ "🍱": 248715,
134
+ "🍵": 248873,
135
+ "🍶": 248814,
136
+ "🍸": 249251,
137
+ "🍼": 248650,
138
+ "🎃": 248736,
139
+ "🎆": 248678,
140
+ "🎇": 248589,
141
+ "🎌": 248805,
142
+ "🎍": 248892,
143
+ "🎎": 249231,
144
+ "🎏": 248670,
145
+ "🎐": 249014,
146
+ "🎑": 248954,
147
+ "🎒": 249072,
148
+ "🎚": 249103,
149
+ "🎛": 248704,
150
+ "🎟": 248802,
151
+ "🎠": 249005,
152
+ "🎡": 248932,
153
+ "🎣": 249137,
154
+ "🎦": 248858,
155
+ "🎪": 249086,
156
+ "🎫": 249258,
157
+ "🎰": 248683,
158
+ "🎱": 248961,
159
+ "🎲": 248455,
160
+ "🎳": 249073,
161
+ "🎴": 248769,
162
+ "🎷": 248451,
163
+ "🎹": 248503,
164
+ "🎺": 248586,
165
+ "🎻": 249148,
166
+ "🎽": 248577,
167
+ "🎾": 249108,
168
+ "🎿": 248915,
169
+ "🏁": 249031,
170
+ "🏂": 248792,
171
+ "🏅": 249063,
172
+ "🏇": 248612,
173
+ "🏈": 248891,
174
+ "🏉": 248434,
175
+ "🏊": 249052,
176
+ "🏍": 248525,
177
+ "🏎": 248992,
178
+ "🏏": 248515,
179
+ "🏐": 248660,
180
+ "🏑": 249144,
181
+ "🏒": 248866,
182
+ "🏓": 248459,
183
+ "🏔": 248588,
184
+ "🏕": 248916,
185
+ "🏗": 249098,
186
+ "🏘": 249199,
187
+ "🏙": 248567,
188
+ "🏚": 248888,
189
+ "🏛": 248914,
190
+ "🏜": 248685,
191
+ "🏞": 249200,
192
+ "🏟": 248815,
193
+ "🏣": 248828,
194
+ "🏤": 248773,
195
+ "🏥": 248835,
196
+ "🏦": 248553,
197
+ "🏧": 248642,
198
+ "🏨": 248561,
199
+ "🏩": 249182,
200
+ "🏪": 248999,
201
+ "🏬": 248996,
202
+ "🏭": 248934,
203
+ "🏮": 248644,
204
+ "🏯": 249175,
205
+ "🏰": 248774,
206
+ "🏴": 248501,
207
+ "🏷": 248427,
208
+ "🏸": 248804,
209
+ "🏹": 249130,
210
+ "🏺": 249250,
211
+ "🐀": 249257,
212
+ "🐁": 249252,
213
+ "🐂": 248621,
214
+ "🐃": 248695,
215
+ "🐄": 249194,
216
+ "🐅": 249117,
217
+ "🐆": 249026,
218
+ "🐇": 249015,
219
+ "🐈": 248527,
220
+ "🐉": 248462,
221
+ "🐊": 248574,
222
+ "🐋": 249208,
223
+ "🐌": 248880,
224
+ "🐎": 248479,
225
+ "🐏": 248865,
226
+ "🐐": 249198,
227
+ "🐑": 249122,
228
+ "🐓": 248680,
229
+ "🐔": 249068,
230
+ "🐕": 248756,
231
+ "🐖": 248603,
232
+ "🐗": 248750,
233
+ "🐙": 248701,
234
+ "🐚": 248886,
235
+ "🐛": 248540,
236
+ "🐜": 249249,
237
+ "🐟": 248912,
238
+ "🐠": 249253,
239
+ "🐡": 248480,
240
+ "🐢": 249210,
241
+ "🐧": 249254,
242
+ "🐨": 249019,
243
+ "🐩": 248580,
244
+ "🐪": 248775,
245
+ "🐫": 248505,
246
+ "🐬": 249002,
247
+ "🐭": 248615,
248
+ "🐮": 249074,
249
+ "🐵": 248533,
250
+ "🐹": 248441,
251
+ "🐺": 249055,
252
+ "🐽": 248602,
253
+ "🐿": 248823,
254
+ "👂": 249140,
255
+ "👓": 249170,
256
+ "👘": 248542,
257
+ "👛": 249217,
258
+ "👝": 248709,
259
+ "👞": 248728,
260
+ "👡": 248549,
261
+ "👢": 248910,
262
+ "👲": 249001,
263
+ "👴": 248820,
264
+ "👵": 248960,
265
+ "👷": 249146,
266
+ "👹": 249099,
267
+ "👺": 249106,
268
+ "👽": 248883,
269
+ "👾": 249059,
270
+ "💂": 248778,
271
+ "💈": 248554,
272
+ "💒": 248594,
273
+ "💤": 248466,
274
+ "💨": 248913,
275
+ "💩": 248508,
276
+ "💱": 249100,
277
+ "💴": 248931,
278
+ "💷": 248711,
279
+ "💹": 249158,
280
+ "💺": 248831,
281
+ "💽": 249021,
282
+ "💾": 249215,
283
+ "💿": 248604,
284
+ "📀": 248518,
285
+ "📁": 248444,
286
+ "📂": 248496,
287
+ "📄": 248801,
288
+ "📆": 248487,
289
+ "📇": 249007,
290
+ "📈": 248438,
291
+ "📉": 248617,
292
+ "📊": 248473,
293
+ "📏": 248556,
294
+ "📐": 248679,
295
+ "📑": 249256,
296
+ "📒": 248710,
297
+ "📓": 248632,
298
+ "📔": 248909,
299
+ "📗": 248807,
300
+ "📘": 248708,
301
+ "📙": 249160,
302
+ "📛": 248484,
303
+ "📟": 248727,
304
+ "📠": 249065,
305
+ "📡": 248532,
306
+ "📤": 248718,
307
+ "📨": 249196,
308
+ "📪": 248565,
309
+ "📫": 248469,
310
+ "📭": 248998,
311
+ "📯": 249220,
312
+ "📳": 248471,
313
+ "📴": 249004,
314
+ "📵": 249013,
315
+ "📶": 248969,
316
+ "📻": 249084,
317
+ "📼": 249230,
318
+ "📿": 249181,
319
+ "🔀": 248946,
320
+ "🔂": 248863,
321
+ "🔃": 249050,
322
+ "🔄": 248995,
323
+ "🔆": 249046,
324
+ "🔇": 249009,
325
+ "🔈": 248559,
326
+ "🔉": 248590,
327
+ "🔋": 248907,
328
+ "🔌": 248842,
329
+ "🔏": 248622,
330
+ "🔐": 248546,
331
+ "🔑": 248937,
332
+ "🔒": 248498,
333
+ "🔓": 249058,
334
+ "🔕": 248780,
335
+ "🔙": 249206,
336
+ "🔚": 248903,
337
+ "🔛": 248671,
338
+ "🔟": 249189,
339
+ "🔠": 249075,
340
+ "🔡": 248600,
341
+ "🔢": 249142,
342
+ "🔣": 248579,
343
+ "🔤": 248997,
344
+ "🔦": 248813,
345
+ "🔧": 248618,
346
+ "🔨": 248849,
347
+ "🔩": 248781,
348
+ "🔪": 248906,
349
+ "🔭": 249261,
350
+ "🔯": 248674,
351
+ "🔲": 248988,
352
+ "🔳": 248985,
353
+ "🔼": 248754,
354
+ "🔽": 248581,
355
+ "🕋": 249027,
356
+ "🕌": 249241,
357
+ "🕍": 248719,
358
+ "🕎": 248626,
359
+ "🕐": 248855,
360
+ "🕑": 248818,
361
+ "🕒": 248485,
362
+ "🕓": 248465,
363
+ "🕕": 248494,
364
+ "🕖": 248847,
365
+ "🕗": 249221,
366
+ "🕘": 248552,
367
+ "🕙": 248539,
368
+ "🕚": 248607,
369
+ "🕛": 248889,
370
+ "🕜": 248513,
371
+ "🕝": 248763,
372
+ "🕞": 248601,
373
+ "🕟": 248884,
374
+ "🕠": 249109,
375
+ "🕡": 248982,
376
+ "🕢": 248896,
377
+ "🕣": 248854,
378
+ "🕤": 248682,
379
+ "🕥": 248627,
380
+ "🕦": 248938,
381
+ "🕧": 248482,
382
+ "🕰": 248810,
383
+ "🕳": 248700,
384
+ "🕴": 248483,
385
+ "🕶": 248688,
386
+ "🕷": 248437,
387
+ "🕸": 248749,
388
+ "🕹": 248871,
389
+ "🖇": 248563,
390
+ "🖊": 248760,
391
+ "🖌": 249209,
392
+ "🖍": 248560,
393
+ "🖕": 249105,
394
+ "🖖": 248864,
395
+ "🖨": 248947,
396
+ "🖱": 248784,
397
+ "🖲": 248766,
398
+ "🖼": 248506,
399
+ "🗂": 248631,
400
+ "🗃": 248497,
401
+ "🗄": 248980,
402
+ "🗑": 249088,
403
+ "🗒": 249048,
404
+ "🗜": 248764,
405
+ "🗝": 248629,
406
+ "🗞": 248867,
407
+ "🗡": 249134,
408
+ "🗨": 249150,
409
+ "🗯": 248585,
410
+ "🗳": 248461,
411
+ "🗺": 248493,
412
+ "🗻": 249133,
413
+ "🗾": 248464,
414
+ "🗿": 249135,
415
+ "😧": 249227,
416
+ "😸": 248908,
417
+ "😺": 248851,
418
+ "😼": 248614,
419
+ "😽": 249121,
420
+ "😾": 248675,
421
+ "😿": 248796,
422
+ "🙉": 248633,
423
+ "🙍": 249066,
424
+ "🙎": 248808,
425
+ "🚁": 248677,
426
+ "🚂": 248504,
427
+ "🚃": 249184,
428
+ "🚄": 248648,
429
+ "🚅": 249102,
430
+ "🚆": 248974,
431
+ "🚈": 248901,
432
+ "🚉": 248856,
433
+ "🚊": 249178,
434
+ "🚋": 248878,
435
+ "🚍": 249016,
436
+ "🚎": 249163,
437
+ "🚏": 249119,
438
+ "🚐": 248731,
439
+ "🚑": 249188,
440
+ "🚒": 249225,
441
+ "🚓": 248817,
442
+ "🚔": 249259,
443
+ "🚕": 248608,
444
+ "🚖": 248667,
445
+ "🚜": 249076,
446
+ "🚝": 248882,
447
+ "🚞": 248596,
448
+ "🚟": 249154,
449
+ "🚠": 248640,
450
+ "🚡": 249020,
451
+ "🚢": 248547,
452
+ "🚣": 249186,
453
+ "🚤": 249204,
454
+ "🚥": 249164,
455
+ "🚦": 248717,
456
+ "🚧": 249143,
457
+ "🚪": 248861,
458
+ "🚬": 248918,
459
+ "🚭": 248836,
460
+ "🚮": 248583,
461
+ "🚯": 248659,
462
+ "🚰": 249202,
463
+ "🚱": 248786,
464
+ "🚳": 248758,
465
+ "🚷": 249149,
466
+ "🚸": 248550,
467
+ "🚹": 249024,
468
+ "🚺": 249044,
469
+ "🚻": 248636,
470
+ "🚼": 249214,
471
+ "🚽": 248726,
472
+ "🚾": 249205,
473
+ "🚿": 248941,
474
+ "🛀": 249097,
475
+ "🛁": 248430,
476
+ "🛂": 248770,
477
+ "🛃": 248732,
478
+ "🛄": 248879,
479
+ "🛅": 249041,
480
+ "🛋": 249203,
481
+ "🛎": 249113,
482
+ "🛏": 249078,
483
+ "🛐": 248830,
484
+ "🛕": 248859,
485
+ "🛖": 248435,
486
+ "🛗": 248551,
487
+ "🛜": 248519,
488
+ "🛝": 248510,
489
+ "🛞": 248897,
490
+ "🛟": 248981,
491
+ "🛠": 248966,
492
+ "🛡": 248811,
493
+ "🛢": 249240,
494
+ "🛣": 249040,
495
+ "🛤": 248973,
496
+ "🛥": 248436,
497
+ "🛩": 249156,
498
+ "🛬": 248951,
499
+ "🛰": 248656,
500
+ "🛳": 249123,
501
+ "🛴": 249111,
502
+ "🛵": 248857,
503
+ "🛶": 248940,
504
+ "🛷": 248824,
505
+ "🛸": 248562,
506
+ "🛹": 248669,
507
+ "🛺": 248569,
508
+ "🛻": 249017,
509
+ "🛼": 248991,
510
+ "🟠": 249195,
511
+ "🟡": 248460,
512
+ "🟢": 248489,
513
+ "🟣": 248737,
514
+ "🟤": 248936,
515
+ "🟥": 249239,
516
+ "🟦": 248491,
517
+ "🟧": 249166,
518
+ "🟨": 248765,
519
+ "🟩": 248439,
520
+ "🟪": 248787,
521
+ "🟫": 249255,
522
+ "🟰": 248874,
523
+ "🤌": 248620,
524
+ "🤍": 248967,
525
+ "🤎": 249116,
526
+ "🤏": 248699,
527
+ "🤐": 248806,
528
+ "🤒": 248610,
529
+ "🤕": 248475,
530
+ "🤚": 249011,
531
+ "🤛": 249107,
532
+ "🤜": 249152,
533
+ "🤠": 248862,
534
+ "🤢": 249023,
535
+ "🤥": 248628,
536
+ "🤧": 249064,
537
+ "🤫": 248706,
538
+ "🤬": 249219,
539
+ "🤮": 248819,
540
+ "🤯": 249096,
541
+ "🤰": 249207,
542
+ "🤱": 248925,
543
+ "🤳": 248541,
544
+ "🤴": 248573,
545
+ "🤶": 249094,
546
+ "🤸": 248809,
547
+ "🤹": 248570,
548
+ "🤺": 248443,
549
+ "🤼": 249115,
550
+ "🤽": 248794,
551
+ "🤾": 248721,
552
+ "🤿": 249191,
553
+ "🥁": 249118,
554
+ "🥃": 248788,
555
+ "🥄": 248944,
556
+ "🥅": 249101,
557
+ "🥈": 248899,
558
+ "🥉": 248762,
559
+ "🥊": 248768,
560
+ "🥋": 248772,
561
+ "🥌": 249177,
562
+ "🥍": 249051,
563
+ "🥎": 248645,
564
+ "🥏": 248557,
565
+ "🥐": 249131,
566
+ "🥑": 248511,
567
+ "🥒": 249229,
568
+ "🥓": 248751,
569
+ "🥔": 248905,
570
+ "🥕": 248838,
571
+ "🥖": 248470,
572
+ "🥗": 248745,
573
+ "🥘": 248445,
574
+ "🥙": 248463,
575
+ "🥚": 248520,
576
+ "🥛": 248582,
577
+ "🥜": 249083,
578
+ "🥝": 248848,
579
+ "🥞": 248453,
580
+ "🥟": 249033,
581
+ "🥠": 248942,
582
+ "🥡": 248720,
583
+ "🥢": 249168,
584
+ "🥤": 248949,
585
+ "🥥": 248956,
586
+ "🥦": 248432,
587
+ "🥧": 248595,
588
+ "🥨": 248799,
589
+ "🥩": 248928,
590
+ "🥪": 248963,
591
+ "🥫": 248509,
592
+ "🥬": 249237,
593
+ "🥭": 248634,
594
+ "🥮": 248446,
595
+ "🥯": 248757,
596
+ "🥱": 249244,
597
+ "🥲": 248429,
598
+ "🥳": 249128,
599
+ "🥴": 248860,
600
+ "🥵": 249029,
601
+ "🥶": 249147,
602
+ "🥷": 248702,
603
+ "🥸": 248486,
604
+ "🥹": 248431,
605
+ "🥻": 249039,
606
+ "🥼": 248832,
607
+ "🥽": 248875,
608
+ "🥾": 249056,
609
+ "🥿": 249138,
610
+ "🦀": 249201,
611
+ "🦂": 249034,
612
+ "🦃": 248834,
613
+ "🦆": 249042,
614
+ "🦇": 248543,
615
+ "🦈": 249139,
616
+ "🦉": 249008,
617
+ "🦌": 248625,
618
+ "🦍": 248713,
619
+ "🦎": 248911,
620
+ "🦏": 248555,
621
+ "🦐": 248950,
622
+ "🦑": 248746,
623
+ "🦒": 248843,
624
+ "🦓": 248789,
625
+ "🦔": 248641,
626
+ "🦕": 248654,
627
+ "🦖": 248686,
628
+ "🦗": 249235,
629
+ "🦘": 248825,
630
+ "🦙": 248523,
631
+ "🦚": 248797,
632
+ "🦛": 249080,
633
+ "🦜": 248816,
634
+ "🦝": 248661,
635
+ "🦞": 249045,
636
+ "🦟": 248753,
637
+ "🦠": 249085,
638
+ "🦡": 249246,
639
+ "🦢": 248965,
640
+ "🦣": 248578,
641
+ "🦤": 248662,
642
+ "🦥": 248939,
643
+ "🦦": 249161,
644
+ "🦧": 249030,
645
+ "🦨": 248776,
646
+ "🦩": 248624,
647
+ "🦪": 248738,
648
+ "🦫": 248566,
649
+ "🦬": 249245,
650
+ "🦭": 249159,
651
+ "🦮": 248517,
652
+ "🦯": 249089,
653
+ "🦰": 249132,
654
+ "🦱": 248613,
655
+ "🦲": 249047,
656
+ "🦳": 248927,
657
+ "🦴": 248691,
658
+ "🦵": 248885,
659
+ "🦶": 248876,
660
+ "🦷": 248771,
661
+ "🦸": 248538,
662
+ "🦹": 248958,
663
+ "🦺": 248948,
664
+ "🦻": 248729,
665
+ "🦼": 248639,
666
+ "🦽": 248959,
667
+ "🦾": 248507,
668
+ "🦿": 249222,
669
+ "🧀": 248712,
670
+ "🧁": 248840,
671
+ "🧂": 249018,
672
+ "🧃": 249125,
673
+ "🧄": 248723,
674
+ "🧅": 248895,
675
+ "🧆": 248571,
676
+ "🧇": 248767,
677
+ "🧈": 249211,
678
+ "🧉": 248684,
679
+ "🧊": 249092,
680
+ "🧋": 248696,
681
+ "🧌": 249061,
682
+ "🧍": 249218,
683
+ "🧎": 248537,
684
+ "🧏": 248544,
685
+ "🧐": 249095,
686
+ "🧑": 248672,
687
+ "🧒": 248536,
688
+ "🧓": 249145,
689
+ "🧔": 249110,
690
+ "🧕": 249176,
691
+ "🧖": 249025,
692
+ "🧗": 248853,
693
+ "🧘": 248663,
694
+ "🧙": 248605,
695
+ "🧛": 248887,
696
+ "🧝": 248524,
697
+ "🧞": 248558,
698
+ "🧟": 249197,
699
+ "🧠": 248893,
700
+ "🧢": 249035,
701
+ "🧣": 249155,
702
+ "🧤": 248740,
703
+ "🧥": 248657,
704
+ "🧦": 248900,
705
+ "🧧": 248428,
706
+ "🧨": 248793,
707
+ "🧩": 249234,
708
+ "🧪": 249192,
709
+ "🧫": 248975,
710
+ "🧬": 248681,
711
+ "🧭": 249180,
712
+ "🧮": 249049,
713
+ "🧯": 249167,
714
+ "🧰": 248593,
715
+ "🧱": 249112,
716
+ "🧲": 249082,
717
+ "🧳": 249179,
718
+ "🧴": 249104,
719
+ "🧵": 249169,
720
+ "🧶": 248651,
721
+ "🧷": 248977,
722
+ "🧸": 248488,
723
+ "🧹": 249233,
724
+ "🧺": 248921,
725
+ "🧻": 249236,
726
+ "🧼": 248993,
727
+ "🧽": 248545,
728
+ "🧾": 248827,
729
+ "🧿": 249114,
730
+ "🩰": 248852,
731
+ "🩱": 248529,
732
+ "🩲": 249171,
733
+ "🩳": 248970,
734
+ "🩴": 248705,
735
+ "🩵": 248989,
736
+ "🩶": 249243,
737
+ "🩷": 248945,
738
+ "🩸": 248839,
739
+ "🩹": 248477,
740
+ "🩺": 248983,
741
+ "🩻": 248474,
742
+ "🩼": 249127,
743
+ "🪀": 249216,
744
+ "🪁": 249183,
745
+ "🪂": 248448,
746
+ "🪃": 248689,
747
+ "🪄": 248653,
748
+ "🪅": 248440,
749
+ "🪆": 248616,
750
+ "🪇": 249238,
751
+ "🪈": 249038,
752
+ "🪐": 249032,
753
+ "🪑": 249185,
754
+ "🪒": 248467,
755
+ "🪓": 249028,
756
+ "🪔": 248919,
757
+ "🪕": 248759,
758
+ "🪖": 248447,
759
+ "🪗": 249006,
760
+ "🪘": 248643,
761
+ "🪙": 248476,
762
+ "🪚": 249165,
763
+ "🪛": 249077,
764
+ "🪜": 248597,
765
+ "🪝": 248714,
766
+ "🪞": 248687,
767
+ "🪟": 249226,
768
+ "🪠": 248499,
769
+ "🪡": 248841,
770
+ "🪢": 249124,
771
+ "🪣": 248955,
772
+ "🪤": 248668,
773
+ "🪥": 248599,
774
+ "🪦": 248902,
775
+ "🪧": 248535,
776
+ "🪨": 248673,
777
+ "🪩": 248962,
778
+ "🪪": 249223,
779
+ "🪫": 248990,
780
+ "🪬": 248564,
781
+ "🪭": 248926,
782
+ "🪮": 248630,
783
+ "🪯": 248822,
784
+ "🪰": 248829,
785
+ "🪱": 249162,
786
+ "🪲": 248522,
787
+ "🪳": 249037,
788
+ "🪴": 248457,
789
+ "🪵": 248877,
790
+ "🪶": 249151,
791
+ "🪷": 249036,
792
+ "🪸": 248492,
793
+ "🪹": 248592,
794
+ "🪺": 248646,
795
+ "🪻": 248898,
796
+ "🪼": 248952,
797
+ "🪽": 249136,
798
+ "🪿": 249071,
799
+ "🫀": 248442,
800
+ "🫁": 248449,
801
+ "🫂": 248606,
802
+ "🫃": 248637,
803
+ "🫄": 248658,
804
+ "🫅": 248846,
805
+ "🫎": 248894,
806
+ "🫏": 248826,
807
+ "🫐": 248500,
808
+ "🫑": 248456,
809
+ "🫒": 248703,
810
+ "🫓": 248743,
811
+ "🫔": 249126,
812
+ "🫕": 249174,
813
+ "🫖": 248724,
814
+ "🫗": 248652,
815
+ "🫘": 249093,
816
+ "🫙": 249054,
817
+ "🫚": 248800,
818
+ "🫛": 248698,
819
+ "🫠": 249247,
820
+ "🫡": 248968,
821
+ "🫢": 248783,
822
+ "🫣": 248481,
823
+ "🫤": 248986,
824
+ "🫥": 249043,
825
+ "🫦": 248526,
826
+ "🫧": 248548,
827
+ "🫨": 248747,
828
+ "🫰": 249012,
829
+ "🫱": 249090,
830
+ "🫲": 248924,
831
+ "🫳": 249228,
832
+ "🫴": 248987,
833
+ "🫵": 248531,
834
+ "🫶": 249157,
835
+ "🫷": 249091,
836
+ "🫸": 248870
837
+ }
config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "CamembertForMaskedLM"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.1,
6
+ "bos_token_id": 0,
7
+ "classifier_dropout": null,
8
+ "eos_token_id": 2,
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 768,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 3072,
14
+ "layer_norm_eps": 1e-12,
15
+ "max_position_embeddings": 512,
16
+ "model_type": "camembert",
17
+ "num_attention_heads": 12,
18
+ "num_hidden_layers": 12,
19
+ "pad_token_id": 1,
20
+ "position_embedding_type": "absolute",
21
+ "torch_dtype": "float32",
22
+ "transformers_version": "4.30.2",
23
+ "type_vocab_size": 1,
24
+ "use_cache": true,
25
+ "vocab_size": 249262
26
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ddb1786f3b22021ff636938d898203fcfe33381b15a63f72b65bdfd375520e4
3
+ size 1110969209
sentencepiece.bpe.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e295c936bc0d8b6669ae769a2f8a0363e6d3abcfd8d0869134aa1e903a447d26
3
+ size 5261686
special_tokens_map.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<s>NOTUSED",
4
+ "</s>NOTUSED",
5
+ "<_>"
6
+ ],
7
+ "bos_token": "<s>",
8
+ "cls_token": "<s>",
9
+ "eos_token": "</s>",
10
+ "mask_token": {
11
+ "content": "<mask>",
12
+ "lstrip": true,
13
+ "normalized": true,
14
+ "rstrip": false,
15
+ "single_word": false
16
+ },
17
+ "pad_token": "<pad>",
18
+ "sep_token": "</s>",
19
+ "unk_token": "<unk>"
20
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d8c7c1ec376502023be87abc9440ef5e643f70e569225799ad2d261c488c1083
3
+ size 17364638
tokenizer_config.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<s>NOTUSED",
4
+ "</s>NOTUSED",
5
+ "<_>"
6
+ ],
7
+ "bos_token": "<s>",
8
+ "clean_up_tokenization_spaces": true,
9
+ "cls_token": "<s>",
10
+ "eos_token": "</s>",
11
+ "mask_token": {
12
+ "__type": "AddedToken",
13
+ "content": "<mask>",
14
+ "lstrip": true,
15
+ "normalized": true,
16
+ "rstrip": false,
17
+ "single_word": false
18
+ },
19
+ "model_max_length": 1000000000000000019884624838656,
20
+ "pad_token": "<pad>",
21
+ "sep_token": "</s>",
22
+ "sp_model_kwargs": {},
23
+ "tokenizer_class": "CamembertTokenizer",
24
+ "unk_token": "<unk>"
25
+ }