Hugging Face's logo

danitamayo
/
roberta_language_detection

Text Classification
Transformers
PyTorch
xlm-roberta
text-embeddings-inference
Model card Files
xet
Community
1
danitamayo commited on
Commit
aeaf71a
·
1 Parent(s): edb1e2b

Upload RobertaForSequenceClassification

Browse files
Files changed (2) hide show
  1. config.json +470 -40
  2. pytorch_model.bin +1 -1
config.json CHANGED
@@ -11,50 +11,480 @@
11
  "hidden_dropout_prob": 0.1,
12
  "hidden_size": 768,
13
  "id2label": {
14
- "0": "ja",
15
- "1": "nl",
16
- "2": "ar",
17
- "3": "pl",
18
- "4": "de",
19
- "5": "it",
20
- "6": "pt",
21
- "7": "tr",
22
- "8": "es",
23
- "9": "hi",
24
- "10": "el",
25
- "11": "ur",
26
- "12": "bg",
27
- "13": "en",
28
- "14": "fr",
29
- "15": "zh",
30
- "16": "ru",
31
- "17": "th",
32
- "18": "sw",
33
- "19": "vi"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
34
  },
35
  "initializer_range": 0.02,
36
  "intermediate_size": 3072,
37
  "label2id": {
38
- "ar": 2,
39
- "bg": 12,
40
- "de": 4,
41
- "el": 10,
42
- "en": 13,
43
- "es": 8,
44
- "fr": 14,
45
- "hi": 9,
46
- "it": 5,
47
- "ja": 0,
48
- "nl": 1,
49
- "pl": 3,
50
- "pt": 6,
51
- "ru": 16,
52
- "sw": 18,
53
- "th": 17,
54
- "tr": 7,
55
- "ur": 11,
56
- "vi": 19,
57
- "zh": 15
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
58
  },
59
  "layer_norm_eps": 1e-05,
60
  "max_position_embeddings": 514,
 
11
  "hidden_dropout_prob": 0.1,
12
  "hidden_size": 768,
13
  "id2label": {
14
+ "0": "ace",
15
+ "1": "afr",
16
+ "2": "als",
17
+ "3": "amh",
18
+ "4": "ang",
19
+ "5": "ara",
20
+ "6": "arg",
21
+ "7": "arz",
22
+ "8": "asm",
23
+ "9": "ast",
24
+ "10": "ava",
25
+ "11": "aym",
26
+ "12": "azb",
27
+ "13": "aze",
28
+ "14": "bak",
29
+ "15": "bar",
30
+ "16": "bcl",
31
+ "17": "be-tarask",
32
+ "18": "bel",
33
+ "19": "ben",
34
+ "20": "bho",
35
+ "21": "bjn",
36
+ "22": "bod",
37
+ "23": "bos",
38
+ "24": "bpy",
39
+ "25": "bre",
40
+ "26": "bul",
41
+ "27": "bxr",
42
+ "28": "cat",
43
+ "29": "cbk",
44
+ "30": "cdo",
45
+ "31": "ceb",
46
+ "32": "ces",
47
+ "33": "che",
48
+ "34": "chr",
49
+ "35": "chv",
50
+ "36": "ckb",
51
+ "37": "cor",
52
+ "38": "cos",
53
+ "39": "crh",
54
+ "40": "csb",
55
+ "41": "cym",
56
+ "42": "dan",
57
+ "43": "deu",
58
+ "44": "diq",
59
+ "45": "div",
60
+ "46": "dsb",
61
+ "47": "dty",
62
+ "48": "egl",
63
+ "49": "ell",
64
+ "50": "eng",
65
+ "51": "epo",
66
+ "52": "est",
67
+ "53": "eus",
68
+ "54": "ext",
69
+ "55": "fao",
70
+ "56": "fas",
71
+ "57": "fin",
72
+ "58": "fra",
73
+ "59": "frp",
74
+ "60": "fry",
75
+ "61": "fur",
76
+ "62": "gag",
77
+ "63": "gla",
78
+ "64": "gle",
79
+ "65": "glg",
80
+ "66": "glk",
81
+ "67": "glv",
82
+ "68": "grn",
83
+ "69": "guj",
84
+ "70": "hak",
85
+ "71": "hat",
86
+ "72": "hau",
87
+ "73": "hbs",
88
+ "74": "heb",
89
+ "75": "hif",
90
+ "76": "hin",
91
+ "77": "hrv",
92
+ "78": "hsb",
93
+ "79": "hun",
94
+ "80": "hye",
95
+ "81": "ibo",
96
+ "82": "ido",
97
+ "83": "ile",
98
+ "84": "ilo",
99
+ "85": "ina",
100
+ "86": "ind",
101
+ "87": "isl",
102
+ "88": "ita",
103
+ "89": "jam",
104
+ "90": "jav",
105
+ "91": "jbo",
106
+ "92": "jpn",
107
+ "93": "kaa",
108
+ "94": "kab",
109
+ "95": "kan",
110
+ "96": "kat",
111
+ "97": "kaz",
112
+ "98": "kbd",
113
+ "99": "khm",
114
+ "100": "kin",
115
+ "101": "kir",
116
+ "102": "koi",
117
+ "103": "kok",
118
+ "104": "kom",
119
+ "105": "kor",
120
+ "106": "krc",
121
+ "107": "ksh",
122
+ "108": "kur",
123
+ "109": "lad",
124
+ "110": "lao",
125
+ "111": "lat",
126
+ "112": "lav",
127
+ "113": "lez",
128
+ "114": "lij",
129
+ "115": "lim",
130
+ "116": "lin",
131
+ "117": "lit",
132
+ "118": "lmo",
133
+ "119": "lrc",
134
+ "120": "ltg",
135
+ "121": "ltz",
136
+ "122": "lug",
137
+ "123": "lzh",
138
+ "124": "mai",
139
+ "125": "mal",
140
+ "126": "map-bms",
141
+ "127": "mar",
142
+ "128": "mdf",
143
+ "129": "mhr",
144
+ "130": "min",
145
+ "131": "mkd",
146
+ "132": "mlg",
147
+ "133": "mlt",
148
+ "134": "mon",
149
+ "135": "mri",
150
+ "136": "mrj",
151
+ "137": "msa",
152
+ "138": "mwl",
153
+ "139": "mya",
154
+ "140": "myv",
155
+ "141": "mzn",
156
+ "142": "nan",
157
+ "143": "nap",
158
+ "144": "nav",
159
+ "145": "nci",
160
+ "146": "nds",
161
+ "147": "nds-nl",
162
+ "148": "nep",
163
+ "149": "new",
164
+ "150": "nld",
165
+ "151": "nno",
166
+ "152": "nob",
167
+ "153": "nrm",
168
+ "154": "nso",
169
+ "155": "oci",
170
+ "156": "olo",
171
+ "157": "ori",
172
+ "158": "orm",
173
+ "159": "oss",
174
+ "160": "pag",
175
+ "161": "pam",
176
+ "162": "pan",
177
+ "163": "pap",
178
+ "164": "pcd",
179
+ "165": "pdc",
180
+ "166": "pfl",
181
+ "167": "pnb",
182
+ "168": "pol",
183
+ "169": "por",
184
+ "170": "pus",
185
+ "171": "que",
186
+ "172": "roa-tara",
187
+ "173": "roh",
188
+ "174": "ron",
189
+ "175": "rue",
190
+ "176": "rup",
191
+ "177": "rus",
192
+ "178": "sah",
193
+ "179": "san",
194
+ "180": "scn",
195
+ "181": "sco",
196
+ "182": "sgs",
197
+ "183": "sin",
198
+ "184": "slk",
199
+ "185": "slv",
200
+ "186": "sme",
201
+ "187": "sna",
202
+ "188": "snd",
203
+ "189": "som",
204
+ "190": "spa",
205
+ "191": "sqi",
206
+ "192": "srd",
207
+ "193": "srn",
208
+ "194": "srp",
209
+ "195": "stq",
210
+ "196": "sun",
211
+ "197": "swa",
212
+ "198": "swe",
213
+ "199": "szl",
214
+ "200": "tam",
215
+ "201": "tat",
216
+ "202": "tcy",
217
+ "203": "tel",
218
+ "204": "tet",
219
+ "205": "tgk",
220
+ "206": "tgl",
221
+ "207": "tha",
222
+ "208": "ton",
223
+ "209": "tsn",
224
+ "210": "tuk",
225
+ "211": "tur",
226
+ "212": "tyv",
227
+ "213": "udm",
228
+ "214": "uig",
229
+ "215": "ukr",
230
+ "216": "urd",
231
+ "217": "uzb",
232
+ "218": "vec",
233
+ "219": "vep",
234
+ "220": "vie",
235
+ "221": "vls",
236
+ "222": "vol",
237
+ "223": "vro",
238
+ "224": "war",
239
+ "225": "wln",
240
+ "226": "wol",
241
+ "227": "wuu",
242
+ "228": "xho",
243
+ "229": "xmf",
244
+ "230": "yid",
245
+ "231": "yor",
246
+ "232": "zea",
247
+ "233": "zh-yue",
248
+ "234": "zho"
249
  },
250
  "initializer_range": 0.02,
251
  "intermediate_size": 3072,
252
  "label2id": {
253
+ "ace": 0,
254
+ "afr": 1,
255
+ "als": 2,
256
+ "amh": 3,
257
+ "ang": 4,
258
+ "ara": 5,
259
+ "arg": 6,
260
+ "arz": 7,
261
+ "asm": 8,
262
+ "ast": 9,
263
+ "ava": 10,
264
+ "aym": 11,
265
+ "azb": 12,
266
+ "aze": 13,
267
+ "bak": 14,
268
+ "bar": 15,
269
+ "bcl": 16,
270
+ "be-tarask": 17,
271
+ "bel": 18,
272
+ "ben": 19,
273
+ "bho": 20,
274
+ "bjn": 21,
275
+ "bod": 22,
276
+ "bos": 23,
277
+ "bpy": 24,
278
+ "bre": 25,
279
+ "bul": 26,
280
+ "bxr": 27,
281
+ "cat": 28,
282
+ "cbk": 29,
283
+ "cdo": 30,
284
+ "ceb": 31,
285
+ "ces": 32,
286
+ "che": 33,
287
+ "chr": 34,
288
+ "chv": 35,
289
+ "ckb": 36,
290
+ "cor": 37,
291
+ "cos": 38,
292
+ "crh": 39,
293
+ "csb": 40,
294
+ "cym": 41,
295
+ "dan": 42,
296
+ "deu": 43,
297
+ "diq": 44,
298
+ "div": 45,
299
+ "dsb": 46,
300
+ "dty": 47,
301
+ "egl": 48,
302
+ "ell": 49,
303
+ "eng": 50,
304
+ "epo": 51,
305
+ "est": 52,
306
+ "eus": 53,
307
+ "ext": 54,
308
+ "fao": 55,
309
+ "fas": 56,
310
+ "fin": 57,
311
+ "fra": 58,
312
+ "frp": 59,
313
+ "fry": 60,
314
+ "fur": 61,
315
+ "gag": 62,
316
+ "gla": 63,
317
+ "gle": 64,
318
+ "glg": 65,
319
+ "glk": 66,
320
+ "glv": 67,
321
+ "grn": 68,
322
+ "guj": 69,
323
+ "hak": 70,
324
+ "hat": 71,
325
+ "hau": 72,
326
+ "hbs": 73,
327
+ "heb": 74,
328
+ "hif": 75,
329
+ "hin": 76,
330
+ "hrv": 77,
331
+ "hsb": 78,
332
+ "hun": 79,
333
+ "hye": 80,
334
+ "ibo": 81,
335
+ "ido": 82,
336
+ "ile": 83,
337
+ "ilo": 84,
338
+ "ina": 85,
339
+ "ind": 86,
340
+ "isl": 87,
341
+ "ita": 88,
342
+ "jam": 89,
343
+ "jav": 90,
344
+ "jbo": 91,
345
+ "jpn": 92,
346
+ "kaa": 93,
347
+ "kab": 94,
348
+ "kan": 95,
349
+ "kat": 96,
350
+ "kaz": 97,
351
+ "kbd": 98,
352
+ "khm": 99,
353
+ "kin": 100,
354
+ "kir": 101,
355
+ "koi": 102,
356
+ "kok": 103,
357
+ "kom": 104,
358
+ "kor": 105,
359
+ "krc": 106,
360
+ "ksh": 107,
361
+ "kur": 108,
362
+ "lad": 109,
363
+ "lao": 110,
364
+ "lat": 111,
365
+ "lav": 112,
366
+ "lez": 113,
367
+ "lij": 114,
368
+ "lim": 115,
369
+ "lin": 116,
370
+ "lit": 117,
371
+ "lmo": 118,
372
+ "lrc": 119,
373
+ "ltg": 120,
374
+ "ltz": 121,
375
+ "lug": 122,
376
+ "lzh": 123,
377
+ "mai": 124,
378
+ "mal": 125,
379
+ "map-bms": 126,
380
+ "mar": 127,
381
+ "mdf": 128,
382
+ "mhr": 129,
383
+ "min": 130,
384
+ "mkd": 131,
385
+ "mlg": 132,
386
+ "mlt": 133,
387
+ "mon": 134,
388
+ "mri": 135,
389
+ "mrj": 136,
390
+ "msa": 137,
391
+ "mwl": 138,
392
+ "mya": 139,
393
+ "myv": 140,
394
+ "mzn": 141,
395
+ "nan": 142,
396
+ "nap": 143,
397
+ "nav": 144,
398
+ "nci": 145,
399
+ "nds": 146,
400
+ "nds-nl": 147,
401
+ "nep": 148,
402
+ "new": 149,
403
+ "nld": 150,
404
+ "nno": 151,
405
+ "nob": 152,
406
+ "nrm": 153,
407
+ "nso": 154,
408
+ "oci": 155,
409
+ "olo": 156,
410
+ "ori": 157,
411
+ "orm": 158,
412
+ "oss": 159,
413
+ "pag": 160,
414
+ "pam": 161,
415
+ "pan": 162,
416
+ "pap": 163,
417
+ "pcd": 164,
418
+ "pdc": 165,
419
+ "pfl": 166,
420
+ "pnb": 167,
421
+ "pol": 168,
422
+ "por": 169,
423
+ "pus": 170,
424
+ "que": 171,
425
+ "roa-tara": 172,
426
+ "roh": 173,
427
+ "ron": 174,
428
+ "rue": 175,
429
+ "rup": 176,
430
+ "rus": 177,
431
+ "sah": 178,
432
+ "san": 179,
433
+ "scn": 180,
434
+ "sco": 181,
435
+ "sgs": 182,
436
+ "sin": 183,
437
+ "slk": 184,
438
+ "slv": 185,
439
+ "sme": 186,
440
+ "sna": 187,
441
+ "snd": 188,
442
+ "som": 189,
443
+ "spa": 190,
444
+ "sqi": 191,
445
+ "srd": 192,
446
+ "srn": 193,
447
+ "srp": 194,
448
+ "stq": 195,
449
+ "sun": 196,
450
+ "swa": 197,
451
+ "swe": 198,
452
+ "szl": 199,
453
+ "tam": 200,
454
+ "tat": 201,
455
+ "tcy": 202,
456
+ "tel": 203,
457
+ "tet": 204,
458
+ "tgk": 205,
459
+ "tgl": 206,
460
+ "tha": 207,
461
+ "ton": 208,
462
+ "tsn": 209,
463
+ "tuk": 210,
464
+ "tur": 211,
465
+ "tyv": 212,
466
+ "udm": 213,
467
+ "uig": 214,
468
+ "ukr": 215,
469
+ "urd": 216,
470
+ "uzb": 217,
471
+ "vec": 218,
472
+ "vep": 219,
473
+ "vie": 220,
474
+ "vls": 221,
475
+ "vol": 222,
476
+ "vro": 223,
477
+ "war": 224,
478
+ "wln": 225,
479
+ "wol": 226,
480
+ "wuu": 227,
481
+ "xho": 228,
482
+ "xmf": 229,
483
+ "yid": 230,
484
+ "yor": 231,
485
+ "zea": 232,
486
+ "zh-yue": 233,
487
+ "zho": 234
488
  },
489
  "layer_norm_eps": 1e-05,
490
  "max_position_embeddings": 514,
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0d5291f635aa0c4a12f30c96c7f7467c0d558deb4bea6a0c53d0a23318d4d877
3
  size 1112970869
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af1a1335f7da050bb23658e9c09fa186b2a8f103372dac5dd6fff946eaa84bea
3
  size 1112970869