Commit
•
90d7728
1
Parent(s):
805464a
New model version
Browse files- added_tokens.json +17 -17
- config.json +11 -11
- pytorch_model.bin +2 -2
- tokenizer.json +0 -0
- tokenizer_config.json +70 -70
added_tokens.json
CHANGED
|
@@ -1,19 +1,19 @@
|
|
| 1 |
{
|
| 2 |
-
"<cel>":
|
| 3 |
-
"<concept>":
|
| 4 |
-
"<date>":
|
| 5 |
-
"<dis>":
|
| 6 |
-
"<eve>":
|
| 7 |
-
"<loc>":
|
| 8 |
-
"<media>":
|
| 9 |
-
"<misc>":
|
| 10 |
-
"<num>":
|
| 11 |
-
"<org>":
|
| 12 |
-
"<per>":
|
| 13 |
-
"<relation>":
|
| 14 |
-
"<time>":
|
| 15 |
-
"<triplet>":
|
| 16 |
-
"ca_XX":
|
| 17 |
-
"el_EL":
|
| 18 |
-
"tp_XX":
|
| 19 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"<cel>": 60065,
|
| 3 |
+
"<concept>": 60068,
|
| 4 |
+
"<date>": 60063,
|
| 5 |
+
"<dis>": 60067,
|
| 6 |
+
"<eve>": 60064,
|
| 7 |
+
"<loc>": 60057,
|
| 8 |
+
"<media>": 60066,
|
| 9 |
+
"<misc>": 60058,
|
| 10 |
+
"<num>": 60060,
|
| 11 |
+
"<org>": 60062,
|
| 12 |
+
"<per>": 60059,
|
| 13 |
+
"<relation>": 60053,
|
| 14 |
+
"<time>": 60061,
|
| 15 |
+
"<triplet>": 60052,
|
| 16 |
+
"ca_XX": 60055,
|
| 17 |
+
"el_EL": 60054,
|
| 18 |
+
"tp_XX": 60056
|
| 19 |
}
|
config.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "
|
| 3 |
"_num_labels": 3,
|
| 4 |
"activation_dropout": 0.0,
|
| 5 |
"activation_function": "gelu",
|
|
@@ -53,26 +53,26 @@
|
|
| 53 |
"torch_dtype": "float32",
|
| 54 |
"transformers_version": "4.34.1",
|
| 55 |
"use_cache": true,
|
| 56 |
-
"vocab_size":
|
| 57 |
"vocabtrimmer": {
|
| 58 |
"mining_config": {
|
| 59 |
-
"dataset": "
|
| 60 |
-
"dataset_column": "
|
| 61 |
"dataset_name": "zh",
|
| 62 |
-
"dataset_split": "
|
| 63 |
"language": "zh",
|
| 64 |
"min_frequency": 2,
|
| 65 |
-
"target_vocab_size":
|
| 66 |
},
|
| 67 |
"stats": {
|
| 68 |
-
"compression_rate_embedding":
|
| 69 |
-
"compression_rate_full":
|
| 70 |
"parameter_size_embedding/raw": 256072704,
|
| 71 |
-
"parameter_size_embedding/trimmed":
|
| 72 |
"parameter_size_full/raw": 610896896,
|
| 73 |
-
"parameter_size_full/trimmed":
|
| 74 |
"vocab_size/raw": 250071,
|
| 75 |
-
"vocab_size/trimmed":
|
| 76 |
}
|
| 77 |
}
|
| 78 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "DarrenChensformer/mrebel-large-trim-zh",
|
| 3 |
"_num_labels": 3,
|
| 4 |
"activation_dropout": 0.0,
|
| 5 |
"activation_function": "gelu",
|
|
|
|
| 53 |
"torch_dtype": "float32",
|
| 54 |
"transformers_version": "4.34.1",
|
| 55 |
"use_cache": true,
|
| 56 |
+
"vocab_size": 60069,
|
| 57 |
"vocabtrimmer": {
|
| 58 |
"mining_config": {
|
| 59 |
+
"dataset": "ikala/datahub-2023Q3",
|
| 60 |
+
"dataset_column": "content",
|
| 61 |
"dataset_name": "zh",
|
| 62 |
+
"dataset_split": "train",
|
| 63 |
"language": "zh",
|
| 64 |
"min_frequency": 2,
|
| 65 |
+
"target_vocab_size": 60000
|
| 66 |
},
|
| 67 |
"stats": {
|
| 68 |
+
"compression_rate_embedding": 24.02077809901988,
|
| 69 |
+
"compression_rate_full": 68.15140995576445,
|
| 70 |
"parameter_size_embedding/raw": 256072704,
|
| 71 |
+
"parameter_size_embedding/trimmed": 61510656,
|
| 72 |
"parameter_size_full/raw": 610896896,
|
| 73 |
+
"parameter_size_full/trimmed": 416334848,
|
| 74 |
"vocab_size/raw": 250071,
|
| 75 |
+
"vocab_size/trimmed": 60069
|
| 76 |
}
|
| 77 |
}
|
| 78 |
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fce28a7bfe49af25440f26db7cedce4ea080d3308012f2e013f713733867e2dd
|
| 3 |
+
size 1665748698
|
tokenizer.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
CHANGED
|
@@ -32,7 +32,7 @@
|
|
| 32 |
"single_word": false,
|
| 33 |
"special": true
|
| 34 |
},
|
| 35 |
-
"
|
| 36 |
"content": "ar_AR",
|
| 37 |
"lstrip": false,
|
| 38 |
"normalized": false,
|
|
@@ -40,7 +40,7 @@
|
|
| 40 |
"single_word": false,
|
| 41 |
"special": true
|
| 42 |
},
|
| 43 |
-
"
|
| 44 |
"content": "cs_CZ",
|
| 45 |
"lstrip": false,
|
| 46 |
"normalized": false,
|
|
@@ -48,7 +48,7 @@
|
|
| 48 |
"single_word": false,
|
| 49 |
"special": true
|
| 50 |
},
|
| 51 |
-
"
|
| 52 |
"content": "de_DE",
|
| 53 |
"lstrip": false,
|
| 54 |
"normalized": false,
|
|
@@ -56,7 +56,7 @@
|
|
| 56 |
"single_word": false,
|
| 57 |
"special": true
|
| 58 |
},
|
| 59 |
-
"
|
| 60 |
"content": "en_XX",
|
| 61 |
"lstrip": false,
|
| 62 |
"normalized": false,
|
|
@@ -64,7 +64,7 @@
|
|
| 64 |
"single_word": false,
|
| 65 |
"special": true
|
| 66 |
},
|
| 67 |
-
"
|
| 68 |
"content": "es_XX",
|
| 69 |
"lstrip": false,
|
| 70 |
"normalized": false,
|
|
@@ -72,7 +72,7 @@
|
|
| 72 |
"single_word": false,
|
| 73 |
"special": true
|
| 74 |
},
|
| 75 |
-
"
|
| 76 |
"content": "et_EE",
|
| 77 |
"lstrip": false,
|
| 78 |
"normalized": false,
|
|
@@ -80,7 +80,7 @@
|
|
| 80 |
"single_word": false,
|
| 81 |
"special": true
|
| 82 |
},
|
| 83 |
-
"
|
| 84 |
"content": "fi_FI",
|
| 85 |
"lstrip": false,
|
| 86 |
"normalized": false,
|
|
@@ -88,7 +88,7 @@
|
|
| 88 |
"single_word": false,
|
| 89 |
"special": true
|
| 90 |
},
|
| 91 |
-
"
|
| 92 |
"content": "fr_XX",
|
| 93 |
"lstrip": false,
|
| 94 |
"normalized": false,
|
|
@@ -96,7 +96,7 @@
|
|
| 96 |
"single_word": false,
|
| 97 |
"special": true
|
| 98 |
},
|
| 99 |
-
"
|
| 100 |
"content": "gu_IN",
|
| 101 |
"lstrip": false,
|
| 102 |
"normalized": false,
|
|
@@ -104,7 +104,7 @@
|
|
| 104 |
"single_word": false,
|
| 105 |
"special": true
|
| 106 |
},
|
| 107 |
-
"
|
| 108 |
"content": "hi_IN",
|
| 109 |
"lstrip": false,
|
| 110 |
"normalized": false,
|
|
@@ -112,7 +112,7 @@
|
|
| 112 |
"single_word": false,
|
| 113 |
"special": true
|
| 114 |
},
|
| 115 |
-
"
|
| 116 |
"content": "it_IT",
|
| 117 |
"lstrip": false,
|
| 118 |
"normalized": false,
|
|
@@ -120,7 +120,7 @@
|
|
| 120 |
"single_word": false,
|
| 121 |
"special": true
|
| 122 |
},
|
| 123 |
-
"
|
| 124 |
"content": "ja_XX",
|
| 125 |
"lstrip": false,
|
| 126 |
"normalized": false,
|
|
@@ -128,7 +128,7 @@
|
|
| 128 |
"single_word": false,
|
| 129 |
"special": true
|
| 130 |
},
|
| 131 |
-
"
|
| 132 |
"content": "kk_KZ",
|
| 133 |
"lstrip": false,
|
| 134 |
"normalized": false,
|
|
@@ -136,7 +136,7 @@
|
|
| 136 |
"single_word": false,
|
| 137 |
"special": true
|
| 138 |
},
|
| 139 |
-
"
|
| 140 |
"content": "ko_KR",
|
| 141 |
"lstrip": false,
|
| 142 |
"normalized": false,
|
|
@@ -144,7 +144,7 @@
|
|
| 144 |
"single_word": false,
|
| 145 |
"special": true
|
| 146 |
},
|
| 147 |
-
"
|
| 148 |
"content": "lt_LT",
|
| 149 |
"lstrip": false,
|
| 150 |
"normalized": false,
|
|
@@ -152,7 +152,7 @@
|
|
| 152 |
"single_word": false,
|
| 153 |
"special": true
|
| 154 |
},
|
| 155 |
-
"
|
| 156 |
"content": "lv_LV",
|
| 157 |
"lstrip": false,
|
| 158 |
"normalized": false,
|
|
@@ -160,7 +160,7 @@
|
|
| 160 |
"single_word": false,
|
| 161 |
"special": true
|
| 162 |
},
|
| 163 |
-
"
|
| 164 |
"content": "my_MM",
|
| 165 |
"lstrip": false,
|
| 166 |
"normalized": false,
|
|
@@ -168,7 +168,7 @@
|
|
| 168 |
"single_word": false,
|
| 169 |
"special": true
|
| 170 |
},
|
| 171 |
-
"
|
| 172 |
"content": "ne_NP",
|
| 173 |
"lstrip": false,
|
| 174 |
"normalized": false,
|
|
@@ -176,7 +176,7 @@
|
|
| 176 |
"single_word": false,
|
| 177 |
"special": true
|
| 178 |
},
|
| 179 |
-
"
|
| 180 |
"content": "nl_XX",
|
| 181 |
"lstrip": false,
|
| 182 |
"normalized": false,
|
|
@@ -184,7 +184,7 @@
|
|
| 184 |
"single_word": false,
|
| 185 |
"special": true
|
| 186 |
},
|
| 187 |
-
"
|
| 188 |
"content": "ro_RO",
|
| 189 |
"lstrip": false,
|
| 190 |
"normalized": false,
|
|
@@ -192,7 +192,7 @@
|
|
| 192 |
"single_word": false,
|
| 193 |
"special": true
|
| 194 |
},
|
| 195 |
-
"
|
| 196 |
"content": "ru_RU",
|
| 197 |
"lstrip": false,
|
| 198 |
"normalized": false,
|
|
@@ -200,7 +200,7 @@
|
|
| 200 |
"single_word": false,
|
| 201 |
"special": true
|
| 202 |
},
|
| 203 |
-
"
|
| 204 |
"content": "si_LK",
|
| 205 |
"lstrip": false,
|
| 206 |
"normalized": false,
|
|
@@ -208,7 +208,7 @@
|
|
| 208 |
"single_word": false,
|
| 209 |
"special": true
|
| 210 |
},
|
| 211 |
-
"
|
| 212 |
"content": "tr_TR",
|
| 213 |
"lstrip": false,
|
| 214 |
"normalized": false,
|
|
@@ -216,7 +216,7 @@
|
|
| 216 |
"single_word": false,
|
| 217 |
"special": true
|
| 218 |
},
|
| 219 |
-
"
|
| 220 |
"content": "vi_VN",
|
| 221 |
"lstrip": false,
|
| 222 |
"normalized": false,
|
|
@@ -224,7 +224,7 @@
|
|
| 224 |
"single_word": false,
|
| 225 |
"special": true
|
| 226 |
},
|
| 227 |
-
"
|
| 228 |
"content": "zh_CN",
|
| 229 |
"lstrip": false,
|
| 230 |
"normalized": false,
|
|
@@ -232,7 +232,7 @@
|
|
| 232 |
"single_word": false,
|
| 233 |
"special": true
|
| 234 |
},
|
| 235 |
-
"
|
| 236 |
"content": "af_ZA",
|
| 237 |
"lstrip": false,
|
| 238 |
"normalized": false,
|
|
@@ -240,7 +240,7 @@
|
|
| 240 |
"single_word": false,
|
| 241 |
"special": true
|
| 242 |
},
|
| 243 |
-
"
|
| 244 |
"content": "az_AZ",
|
| 245 |
"lstrip": false,
|
| 246 |
"normalized": false,
|
|
@@ -248,7 +248,7 @@
|
|
| 248 |
"single_word": false,
|
| 249 |
"special": true
|
| 250 |
},
|
| 251 |
-
"
|
| 252 |
"content": "bn_IN",
|
| 253 |
"lstrip": false,
|
| 254 |
"normalized": false,
|
|
@@ -256,7 +256,7 @@
|
|
| 256 |
"single_word": false,
|
| 257 |
"special": true
|
| 258 |
},
|
| 259 |
-
"
|
| 260 |
"content": "fa_IR",
|
| 261 |
"lstrip": false,
|
| 262 |
"normalized": false,
|
|
@@ -264,7 +264,7 @@
|
|
| 264 |
"single_word": false,
|
| 265 |
"special": true
|
| 266 |
},
|
| 267 |
-
"
|
| 268 |
"content": "he_IL",
|
| 269 |
"lstrip": false,
|
| 270 |
"normalized": false,
|
|
@@ -272,7 +272,7 @@
|
|
| 272 |
"single_word": false,
|
| 273 |
"special": true
|
| 274 |
},
|
| 275 |
-
"
|
| 276 |
"content": "hr_HR",
|
| 277 |
"lstrip": false,
|
| 278 |
"normalized": false,
|
|
@@ -280,7 +280,7 @@
|
|
| 280 |
"single_word": false,
|
| 281 |
"special": true
|
| 282 |
},
|
| 283 |
-
"
|
| 284 |
"content": "id_ID",
|
| 285 |
"lstrip": false,
|
| 286 |
"normalized": false,
|
|
@@ -288,7 +288,7 @@
|
|
| 288 |
"single_word": false,
|
| 289 |
"special": true
|
| 290 |
},
|
| 291 |
-
"
|
| 292 |
"content": "ka_GE",
|
| 293 |
"lstrip": false,
|
| 294 |
"normalized": false,
|
|
@@ -296,7 +296,7 @@
|
|
| 296 |
"single_word": false,
|
| 297 |
"special": true
|
| 298 |
},
|
| 299 |
-
"
|
| 300 |
"content": "km_KH",
|
| 301 |
"lstrip": false,
|
| 302 |
"normalized": false,
|
|
@@ -304,7 +304,7 @@
|
|
| 304 |
"single_word": false,
|
| 305 |
"special": true
|
| 306 |
},
|
| 307 |
-
"
|
| 308 |
"content": "mk_MK",
|
| 309 |
"lstrip": false,
|
| 310 |
"normalized": false,
|
|
@@ -312,7 +312,7 @@
|
|
| 312 |
"single_word": false,
|
| 313 |
"special": true
|
| 314 |
},
|
| 315 |
-
"
|
| 316 |
"content": "ml_IN",
|
| 317 |
"lstrip": false,
|
| 318 |
"normalized": false,
|
|
@@ -320,7 +320,7 @@
|
|
| 320 |
"single_word": false,
|
| 321 |
"special": true
|
| 322 |
},
|
| 323 |
-
"
|
| 324 |
"content": "mn_MN",
|
| 325 |
"lstrip": false,
|
| 326 |
"normalized": false,
|
|
@@ -328,7 +328,7 @@
|
|
| 328 |
"single_word": false,
|
| 329 |
"special": true
|
| 330 |
},
|
| 331 |
-
"
|
| 332 |
"content": "mr_IN",
|
| 333 |
"lstrip": false,
|
| 334 |
"normalized": false,
|
|
@@ -336,7 +336,7 @@
|
|
| 336 |
"single_word": false,
|
| 337 |
"special": true
|
| 338 |
},
|
| 339 |
-
"
|
| 340 |
"content": "pl_PL",
|
| 341 |
"lstrip": false,
|
| 342 |
"normalized": false,
|
|
@@ -344,7 +344,7 @@
|
|
| 344 |
"single_word": false,
|
| 345 |
"special": true
|
| 346 |
},
|
| 347 |
-
"
|
| 348 |
"content": "ps_AF",
|
| 349 |
"lstrip": false,
|
| 350 |
"normalized": false,
|
|
@@ -352,7 +352,7 @@
|
|
| 352 |
"single_word": false,
|
| 353 |
"special": true
|
| 354 |
},
|
| 355 |
-
"
|
| 356 |
"content": "pt_XX",
|
| 357 |
"lstrip": false,
|
| 358 |
"normalized": false,
|
|
@@ -360,7 +360,7 @@
|
|
| 360 |
"single_word": false,
|
| 361 |
"special": true
|
| 362 |
},
|
| 363 |
-
"
|
| 364 |
"content": "sv_SE",
|
| 365 |
"lstrip": false,
|
| 366 |
"normalized": false,
|
|
@@ -368,7 +368,7 @@
|
|
| 368 |
"single_word": false,
|
| 369 |
"special": true
|
| 370 |
},
|
| 371 |
-
"
|
| 372 |
"content": "sw_KE",
|
| 373 |
"lstrip": false,
|
| 374 |
"normalized": false,
|
|
@@ -376,7 +376,7 @@
|
|
| 376 |
"single_word": false,
|
| 377 |
"special": true
|
| 378 |
},
|
| 379 |
-
"
|
| 380 |
"content": "ta_IN",
|
| 381 |
"lstrip": false,
|
| 382 |
"normalized": false,
|
|
@@ -384,7 +384,7 @@
|
|
| 384 |
"single_word": false,
|
| 385 |
"special": true
|
| 386 |
},
|
| 387 |
-
"
|
| 388 |
"content": "te_IN",
|
| 389 |
"lstrip": false,
|
| 390 |
"normalized": false,
|
|
@@ -392,7 +392,7 @@
|
|
| 392 |
"single_word": false,
|
| 393 |
"special": true
|
| 394 |
},
|
| 395 |
-
"
|
| 396 |
"content": "th_TH",
|
| 397 |
"lstrip": false,
|
| 398 |
"normalized": false,
|
|
@@ -400,7 +400,7 @@
|
|
| 400 |
"single_word": false,
|
| 401 |
"special": true
|
| 402 |
},
|
| 403 |
-
"
|
| 404 |
"content": "tl_XX",
|
| 405 |
"lstrip": false,
|
| 406 |
"normalized": false,
|
|
@@ -408,7 +408,7 @@
|
|
| 408 |
"single_word": false,
|
| 409 |
"special": true
|
| 410 |
},
|
| 411 |
-
"
|
| 412 |
"content": "uk_UA",
|
| 413 |
"lstrip": false,
|
| 414 |
"normalized": false,
|
|
@@ -416,7 +416,7 @@
|
|
| 416 |
"single_word": false,
|
| 417 |
"special": true
|
| 418 |
},
|
| 419 |
-
"
|
| 420 |
"content": "ur_PK",
|
| 421 |
"lstrip": false,
|
| 422 |
"normalized": false,
|
|
@@ -424,7 +424,7 @@
|
|
| 424 |
"single_word": false,
|
| 425 |
"special": true
|
| 426 |
},
|
| 427 |
-
"
|
| 428 |
"content": "xh_ZA",
|
| 429 |
"lstrip": false,
|
| 430 |
"normalized": false,
|
|
@@ -432,7 +432,7 @@
|
|
| 432 |
"single_word": false,
|
| 433 |
"special": true
|
| 434 |
},
|
| 435 |
-
"
|
| 436 |
"content": "gl_ES",
|
| 437 |
"lstrip": false,
|
| 438 |
"normalized": false,
|
|
@@ -440,7 +440,7 @@
|
|
| 440 |
"single_word": false,
|
| 441 |
"special": true
|
| 442 |
},
|
| 443 |
-
"
|
| 444 |
"content": "sl_SI",
|
| 445 |
"lstrip": false,
|
| 446 |
"normalized": false,
|
|
@@ -448,7 +448,7 @@
|
|
| 448 |
"single_word": false,
|
| 449 |
"special": true
|
| 450 |
},
|
| 451 |
-
"
|
| 452 |
"content": "<mask>",
|
| 453 |
"lstrip": true,
|
| 454 |
"normalized": false,
|
|
@@ -456,7 +456,7 @@
|
|
| 456 |
"single_word": false,
|
| 457 |
"special": true
|
| 458 |
},
|
| 459 |
-
"
|
| 460 |
"content": "<triplet>",
|
| 461 |
"lstrip": false,
|
| 462 |
"normalized": false,
|
|
@@ -464,7 +464,7 @@
|
|
| 464 |
"single_word": false,
|
| 465 |
"special": true
|
| 466 |
},
|
| 467 |
-
"
|
| 468 |
"content": "<relation>",
|
| 469 |
"lstrip": false,
|
| 470 |
"normalized": false,
|
|
@@ -472,7 +472,7 @@
|
|
| 472 |
"single_word": false,
|
| 473 |
"special": true
|
| 474 |
},
|
| 475 |
-
"
|
| 476 |
"content": "el_EL",
|
| 477 |
"lstrip": false,
|
| 478 |
"normalized": false,
|
|
@@ -480,7 +480,7 @@
|
|
| 480 |
"single_word": false,
|
| 481 |
"special": true
|
| 482 |
},
|
| 483 |
-
"
|
| 484 |
"content": "ca_XX",
|
| 485 |
"lstrip": false,
|
| 486 |
"normalized": false,
|
|
@@ -488,7 +488,7 @@
|
|
| 488 |
"single_word": false,
|
| 489 |
"special": true
|
| 490 |
},
|
| 491 |
-
"
|
| 492 |
"content": "tp_XX",
|
| 493 |
"lstrip": false,
|
| 494 |
"normalized": false,
|
|
@@ -496,7 +496,7 @@
|
|
| 496 |
"single_word": false,
|
| 497 |
"special": true
|
| 498 |
},
|
| 499 |
-
"
|
| 500 |
"content": "<loc>",
|
| 501 |
"lstrip": false,
|
| 502 |
"normalized": false,
|
|
@@ -504,7 +504,7 @@
|
|
| 504 |
"single_word": false,
|
| 505 |
"special": true
|
| 506 |
},
|
| 507 |
-
"
|
| 508 |
"content": "<misc>",
|
| 509 |
"lstrip": false,
|
| 510 |
"normalized": false,
|
|
@@ -512,7 +512,7 @@
|
|
| 512 |
"single_word": false,
|
| 513 |
"special": true
|
| 514 |
},
|
| 515 |
-
"
|
| 516 |
"content": "<per>",
|
| 517 |
"lstrip": false,
|
| 518 |
"normalized": false,
|
|
@@ -520,7 +520,7 @@
|
|
| 520 |
"single_word": false,
|
| 521 |
"special": true
|
| 522 |
},
|
| 523 |
-
"
|
| 524 |
"content": "<num>",
|
| 525 |
"lstrip": false,
|
| 526 |
"normalized": false,
|
|
@@ -528,7 +528,7 @@
|
|
| 528 |
"single_word": false,
|
| 529 |
"special": true
|
| 530 |
},
|
| 531 |
-
"
|
| 532 |
"content": "<time>",
|
| 533 |
"lstrip": false,
|
| 534 |
"normalized": false,
|
|
@@ -536,7 +536,7 @@
|
|
| 536 |
"single_word": false,
|
| 537 |
"special": true
|
| 538 |
},
|
| 539 |
-
"
|
| 540 |
"content": "<org>",
|
| 541 |
"lstrip": false,
|
| 542 |
"normalized": false,
|
|
@@ -544,7 +544,7 @@
|
|
| 544 |
"single_word": false,
|
| 545 |
"special": true
|
| 546 |
},
|
| 547 |
-
"
|
| 548 |
"content": "<date>",
|
| 549 |
"lstrip": false,
|
| 550 |
"normalized": false,
|
|
@@ -552,7 +552,7 @@
|
|
| 552 |
"single_word": false,
|
| 553 |
"special": true
|
| 554 |
},
|
| 555 |
-
"
|
| 556 |
"content": "<eve>",
|
| 557 |
"lstrip": false,
|
| 558 |
"normalized": false,
|
|
@@ -560,7 +560,7 @@
|
|
| 560 |
"single_word": false,
|
| 561 |
"special": true
|
| 562 |
},
|
| 563 |
-
"
|
| 564 |
"content": "<cel>",
|
| 565 |
"lstrip": false,
|
| 566 |
"normalized": false,
|
|
@@ -568,7 +568,7 @@
|
|
| 568 |
"single_word": false,
|
| 569 |
"special": true
|
| 570 |
},
|
| 571 |
-
"
|
| 572 |
"content": "<media>",
|
| 573 |
"lstrip": false,
|
| 574 |
"normalized": false,
|
|
@@ -576,7 +576,7 @@
|
|
| 576 |
"single_word": false,
|
| 577 |
"special": true
|
| 578 |
},
|
| 579 |
-
"
|
| 580 |
"content": "<dis>",
|
| 581 |
"lstrip": false,
|
| 582 |
"normalized": false,
|
|
@@ -584,7 +584,7 @@
|
|
| 584 |
"single_word": false,
|
| 585 |
"special": true
|
| 586 |
},
|
| 587 |
-
"
|
| 588 |
"content": "<concept>",
|
| 589 |
"lstrip": false,
|
| 590 |
"normalized": false,
|
|
|
|
| 32 |
"single_word": false,
|
| 33 |
"special": true
|
| 34 |
},
|
| 35 |
+
"59999": {
|
| 36 |
"content": "ar_AR",
|
| 37 |
"lstrip": false,
|
| 38 |
"normalized": false,
|
|
|
|
| 40 |
"single_word": false,
|
| 41 |
"special": true
|
| 42 |
},
|
| 43 |
+
"60000": {
|
| 44 |
"content": "cs_CZ",
|
| 45 |
"lstrip": false,
|
| 46 |
"normalized": false,
|
|
|
|
| 48 |
"single_word": false,
|
| 49 |
"special": true
|
| 50 |
},
|
| 51 |
+
"60001": {
|
| 52 |
"content": "de_DE",
|
| 53 |
"lstrip": false,
|
| 54 |
"normalized": false,
|
|
|
|
| 56 |
"single_word": false,
|
| 57 |
"special": true
|
| 58 |
},
|
| 59 |
+
"60002": {
|
| 60 |
"content": "en_XX",
|
| 61 |
"lstrip": false,
|
| 62 |
"normalized": false,
|
|
|
|
| 64 |
"single_word": false,
|
| 65 |
"special": true
|
| 66 |
},
|
| 67 |
+
"60003": {
|
| 68 |
"content": "es_XX",
|
| 69 |
"lstrip": false,
|
| 70 |
"normalized": false,
|
|
|
|
| 72 |
"single_word": false,
|
| 73 |
"special": true
|
| 74 |
},
|
| 75 |
+
"60004": {
|
| 76 |
"content": "et_EE",
|
| 77 |
"lstrip": false,
|
| 78 |
"normalized": false,
|
|
|
|
| 80 |
"single_word": false,
|
| 81 |
"special": true
|
| 82 |
},
|
| 83 |
+
"60005": {
|
| 84 |
"content": "fi_FI",
|
| 85 |
"lstrip": false,
|
| 86 |
"normalized": false,
|
|
|
|
| 88 |
"single_word": false,
|
| 89 |
"special": true
|
| 90 |
},
|
| 91 |
+
"60006": {
|
| 92 |
"content": "fr_XX",
|
| 93 |
"lstrip": false,
|
| 94 |
"normalized": false,
|
|
|
|
| 96 |
"single_word": false,
|
| 97 |
"special": true
|
| 98 |
},
|
| 99 |
+
"60007": {
|
| 100 |
"content": "gu_IN",
|
| 101 |
"lstrip": false,
|
| 102 |
"normalized": false,
|
|
|
|
| 104 |
"single_word": false,
|
| 105 |
"special": true
|
| 106 |
},
|
| 107 |
+
"60008": {
|
| 108 |
"content": "hi_IN",
|
| 109 |
"lstrip": false,
|
| 110 |
"normalized": false,
|
|
|
|
| 112 |
"single_word": false,
|
| 113 |
"special": true
|
| 114 |
},
|
| 115 |
+
"60009": {
|
| 116 |
"content": "it_IT",
|
| 117 |
"lstrip": false,
|
| 118 |
"normalized": false,
|
|
|
|
| 120 |
"single_word": false,
|
| 121 |
"special": true
|
| 122 |
},
|
| 123 |
+
"60010": {
|
| 124 |
"content": "ja_XX",
|
| 125 |
"lstrip": false,
|
| 126 |
"normalized": false,
|
|
|
|
| 128 |
"single_word": false,
|
| 129 |
"special": true
|
| 130 |
},
|
| 131 |
+
"60011": {
|
| 132 |
"content": "kk_KZ",
|
| 133 |
"lstrip": false,
|
| 134 |
"normalized": false,
|
|
|
|
| 136 |
"single_word": false,
|
| 137 |
"special": true
|
| 138 |
},
|
| 139 |
+
"60012": {
|
| 140 |
"content": "ko_KR",
|
| 141 |
"lstrip": false,
|
| 142 |
"normalized": false,
|
|
|
|
| 144 |
"single_word": false,
|
| 145 |
"special": true
|
| 146 |
},
|
| 147 |
+
"60013": {
|
| 148 |
"content": "lt_LT",
|
| 149 |
"lstrip": false,
|
| 150 |
"normalized": false,
|
|
|
|
| 152 |
"single_word": false,
|
| 153 |
"special": true
|
| 154 |
},
|
| 155 |
+
"60014": {
|
| 156 |
"content": "lv_LV",
|
| 157 |
"lstrip": false,
|
| 158 |
"normalized": false,
|
|
|
|
| 160 |
"single_word": false,
|
| 161 |
"special": true
|
| 162 |
},
|
| 163 |
+
"60015": {
|
| 164 |
"content": "my_MM",
|
| 165 |
"lstrip": false,
|
| 166 |
"normalized": false,
|
|
|
|
| 168 |
"single_word": false,
|
| 169 |
"special": true
|
| 170 |
},
|
| 171 |
+
"60016": {
|
| 172 |
"content": "ne_NP",
|
| 173 |
"lstrip": false,
|
| 174 |
"normalized": false,
|
|
|
|
| 176 |
"single_word": false,
|
| 177 |
"special": true
|
| 178 |
},
|
| 179 |
+
"60017": {
|
| 180 |
"content": "nl_XX",
|
| 181 |
"lstrip": false,
|
| 182 |
"normalized": false,
|
|
|
|
| 184 |
"single_word": false,
|
| 185 |
"special": true
|
| 186 |
},
|
| 187 |
+
"60018": {
|
| 188 |
"content": "ro_RO",
|
| 189 |
"lstrip": false,
|
| 190 |
"normalized": false,
|
|
|
|
| 192 |
"single_word": false,
|
| 193 |
"special": true
|
| 194 |
},
|
| 195 |
+
"60019": {
|
| 196 |
"content": "ru_RU",
|
| 197 |
"lstrip": false,
|
| 198 |
"normalized": false,
|
|
|
|
| 200 |
"single_word": false,
|
| 201 |
"special": true
|
| 202 |
},
|
| 203 |
+
"60020": {
|
| 204 |
"content": "si_LK",
|
| 205 |
"lstrip": false,
|
| 206 |
"normalized": false,
|
|
|
|
| 208 |
"single_word": false,
|
| 209 |
"special": true
|
| 210 |
},
|
| 211 |
+
"60021": {
|
| 212 |
"content": "tr_TR",
|
| 213 |
"lstrip": false,
|
| 214 |
"normalized": false,
|
|
|
|
| 216 |
"single_word": false,
|
| 217 |
"special": true
|
| 218 |
},
|
| 219 |
+
"60022": {
|
| 220 |
"content": "vi_VN",
|
| 221 |
"lstrip": false,
|
| 222 |
"normalized": false,
|
|
|
|
| 224 |
"single_word": false,
|
| 225 |
"special": true
|
| 226 |
},
|
| 227 |
+
"60023": {
|
| 228 |
"content": "zh_CN",
|
| 229 |
"lstrip": false,
|
| 230 |
"normalized": false,
|
|
|
|
| 232 |
"single_word": false,
|
| 233 |
"special": true
|
| 234 |
},
|
| 235 |
+
"60024": {
|
| 236 |
"content": "af_ZA",
|
| 237 |
"lstrip": false,
|
| 238 |
"normalized": false,
|
|
|
|
| 240 |
"single_word": false,
|
| 241 |
"special": true
|
| 242 |
},
|
| 243 |
+
"60025": {
|
| 244 |
"content": "az_AZ",
|
| 245 |
"lstrip": false,
|
| 246 |
"normalized": false,
|
|
|
|
| 248 |
"single_word": false,
|
| 249 |
"special": true
|
| 250 |
},
|
| 251 |
+
"60026": {
|
| 252 |
"content": "bn_IN",
|
| 253 |
"lstrip": false,
|
| 254 |
"normalized": false,
|
|
|
|
| 256 |
"single_word": false,
|
| 257 |
"special": true
|
| 258 |
},
|
| 259 |
+
"60027": {
|
| 260 |
"content": "fa_IR",
|
| 261 |
"lstrip": false,
|
| 262 |
"normalized": false,
|
|
|
|
| 264 |
"single_word": false,
|
| 265 |
"special": true
|
| 266 |
},
|
| 267 |
+
"60028": {
|
| 268 |
"content": "he_IL",
|
| 269 |
"lstrip": false,
|
| 270 |
"normalized": false,
|
|
|
|
| 272 |
"single_word": false,
|
| 273 |
"special": true
|
| 274 |
},
|
| 275 |
+
"60029": {
|
| 276 |
"content": "hr_HR",
|
| 277 |
"lstrip": false,
|
| 278 |
"normalized": false,
|
|
|
|
| 280 |
"single_word": false,
|
| 281 |
"special": true
|
| 282 |
},
|
| 283 |
+
"60030": {
|
| 284 |
"content": "id_ID",
|
| 285 |
"lstrip": false,
|
| 286 |
"normalized": false,
|
|
|
|
| 288 |
"single_word": false,
|
| 289 |
"special": true
|
| 290 |
},
|
| 291 |
+
"60031": {
|
| 292 |
"content": "ka_GE",
|
| 293 |
"lstrip": false,
|
| 294 |
"normalized": false,
|
|
|
|
| 296 |
"single_word": false,
|
| 297 |
"special": true
|
| 298 |
},
|
| 299 |
+
"60032": {
|
| 300 |
"content": "km_KH",
|
| 301 |
"lstrip": false,
|
| 302 |
"normalized": false,
|
|
|
|
| 304 |
"single_word": false,
|
| 305 |
"special": true
|
| 306 |
},
|
| 307 |
+
"60033": {
|
| 308 |
"content": "mk_MK",
|
| 309 |
"lstrip": false,
|
| 310 |
"normalized": false,
|
|
|
|
| 312 |
"single_word": false,
|
| 313 |
"special": true
|
| 314 |
},
|
| 315 |
+
"60034": {
|
| 316 |
"content": "ml_IN",
|
| 317 |
"lstrip": false,
|
| 318 |
"normalized": false,
|
|
|
|
| 320 |
"single_word": false,
|
| 321 |
"special": true
|
| 322 |
},
|
| 323 |
+
"60035": {
|
| 324 |
"content": "mn_MN",
|
| 325 |
"lstrip": false,
|
| 326 |
"normalized": false,
|
|
|
|
| 328 |
"single_word": false,
|
| 329 |
"special": true
|
| 330 |
},
|
| 331 |
+
"60036": {
|
| 332 |
"content": "mr_IN",
|
| 333 |
"lstrip": false,
|
| 334 |
"normalized": false,
|
|
|
|
| 336 |
"single_word": false,
|
| 337 |
"special": true
|
| 338 |
},
|
| 339 |
+
"60037": {
|
| 340 |
"content": "pl_PL",
|
| 341 |
"lstrip": false,
|
| 342 |
"normalized": false,
|
|
|
|
| 344 |
"single_word": false,
|
| 345 |
"special": true
|
| 346 |
},
|
| 347 |
+
"60038": {
|
| 348 |
"content": "ps_AF",
|
| 349 |
"lstrip": false,
|
| 350 |
"normalized": false,
|
|
|
|
| 352 |
"single_word": false,
|
| 353 |
"special": true
|
| 354 |
},
|
| 355 |
+
"60039": {
|
| 356 |
"content": "pt_XX",
|
| 357 |
"lstrip": false,
|
| 358 |
"normalized": false,
|
|
|
|
| 360 |
"single_word": false,
|
| 361 |
"special": true
|
| 362 |
},
|
| 363 |
+
"60040": {
|
| 364 |
"content": "sv_SE",
|
| 365 |
"lstrip": false,
|
| 366 |
"normalized": false,
|
|
|
|
| 368 |
"single_word": false,
|
| 369 |
"special": true
|
| 370 |
},
|
| 371 |
+
"60041": {
|
| 372 |
"content": "sw_KE",
|
| 373 |
"lstrip": false,
|
| 374 |
"normalized": false,
|
|
|
|
| 376 |
"single_word": false,
|
| 377 |
"special": true
|
| 378 |
},
|
| 379 |
+
"60042": {
|
| 380 |
"content": "ta_IN",
|
| 381 |
"lstrip": false,
|
| 382 |
"normalized": false,
|
|
|
|
| 384 |
"single_word": false,
|
| 385 |
"special": true
|
| 386 |
},
|
| 387 |
+
"60043": {
|
| 388 |
"content": "te_IN",
|
| 389 |
"lstrip": false,
|
| 390 |
"normalized": false,
|
|
|
|
| 392 |
"single_word": false,
|
| 393 |
"special": true
|
| 394 |
},
|
| 395 |
+
"60044": {
|
| 396 |
"content": "th_TH",
|
| 397 |
"lstrip": false,
|
| 398 |
"normalized": false,
|
|
|
|
| 400 |
"single_word": false,
|
| 401 |
"special": true
|
| 402 |
},
|
| 403 |
+
"60045": {
|
| 404 |
"content": "tl_XX",
|
| 405 |
"lstrip": false,
|
| 406 |
"normalized": false,
|
|
|
|
| 408 |
"single_word": false,
|
| 409 |
"special": true
|
| 410 |
},
|
| 411 |
+
"60046": {
|
| 412 |
"content": "uk_UA",
|
| 413 |
"lstrip": false,
|
| 414 |
"normalized": false,
|
|
|
|
| 416 |
"single_word": false,
|
| 417 |
"special": true
|
| 418 |
},
|
| 419 |
+
"60047": {
|
| 420 |
"content": "ur_PK",
|
| 421 |
"lstrip": false,
|
| 422 |
"normalized": false,
|
|
|
|
| 424 |
"single_word": false,
|
| 425 |
"special": true
|
| 426 |
},
|
| 427 |
+
"60048": {
|
| 428 |
"content": "xh_ZA",
|
| 429 |
"lstrip": false,
|
| 430 |
"normalized": false,
|
|
|
|
| 432 |
"single_word": false,
|
| 433 |
"special": true
|
| 434 |
},
|
| 435 |
+
"60049": {
|
| 436 |
"content": "gl_ES",
|
| 437 |
"lstrip": false,
|
| 438 |
"normalized": false,
|
|
|
|
| 440 |
"single_word": false,
|
| 441 |
"special": true
|
| 442 |
},
|
| 443 |
+
"60050": {
|
| 444 |
"content": "sl_SI",
|
| 445 |
"lstrip": false,
|
| 446 |
"normalized": false,
|
|
|
|
| 448 |
"single_word": false,
|
| 449 |
"special": true
|
| 450 |
},
|
| 451 |
+
"60051": {
|
| 452 |
"content": "<mask>",
|
| 453 |
"lstrip": true,
|
| 454 |
"normalized": false,
|
|
|
|
| 456 |
"single_word": false,
|
| 457 |
"special": true
|
| 458 |
},
|
| 459 |
+
"60052": {
|
| 460 |
"content": "<triplet>",
|
| 461 |
"lstrip": false,
|
| 462 |
"normalized": false,
|
|
|
|
| 464 |
"single_word": false,
|
| 465 |
"special": true
|
| 466 |
},
|
| 467 |
+
"60053": {
|
| 468 |
"content": "<relation>",
|
| 469 |
"lstrip": false,
|
| 470 |
"normalized": false,
|
|
|
|
| 472 |
"single_word": false,
|
| 473 |
"special": true
|
| 474 |
},
|
| 475 |
+
"60054": {
|
| 476 |
"content": "el_EL",
|
| 477 |
"lstrip": false,
|
| 478 |
"normalized": false,
|
|
|
|
| 480 |
"single_word": false,
|
| 481 |
"special": true
|
| 482 |
},
|
| 483 |
+
"60055": {
|
| 484 |
"content": "ca_XX",
|
| 485 |
"lstrip": false,
|
| 486 |
"normalized": false,
|
|
|
|
| 488 |
"single_word": false,
|
| 489 |
"special": true
|
| 490 |
},
|
| 491 |
+
"60056": {
|
| 492 |
"content": "tp_XX",
|
| 493 |
"lstrip": false,
|
| 494 |
"normalized": false,
|
|
|
|
| 496 |
"single_word": false,
|
| 497 |
"special": true
|
| 498 |
},
|
| 499 |
+
"60057": {
|
| 500 |
"content": "<loc>",
|
| 501 |
"lstrip": false,
|
| 502 |
"normalized": false,
|
|
|
|
| 504 |
"single_word": false,
|
| 505 |
"special": true
|
| 506 |
},
|
| 507 |
+
"60058": {
|
| 508 |
"content": "<misc>",
|
| 509 |
"lstrip": false,
|
| 510 |
"normalized": false,
|
|
|
|
| 512 |
"single_word": false,
|
| 513 |
"special": true
|
| 514 |
},
|
| 515 |
+
"60059": {
|
| 516 |
"content": "<per>",
|
| 517 |
"lstrip": false,
|
| 518 |
"normalized": false,
|
|
|
|
| 520 |
"single_word": false,
|
| 521 |
"special": true
|
| 522 |
},
|
| 523 |
+
"60060": {
|
| 524 |
"content": "<num>",
|
| 525 |
"lstrip": false,
|
| 526 |
"normalized": false,
|
|
|
|
| 528 |
"single_word": false,
|
| 529 |
"special": true
|
| 530 |
},
|
| 531 |
+
"60061": {
|
| 532 |
"content": "<time>",
|
| 533 |
"lstrip": false,
|
| 534 |
"normalized": false,
|
|
|
|
| 536 |
"single_word": false,
|
| 537 |
"special": true
|
| 538 |
},
|
| 539 |
+
"60062": {
|
| 540 |
"content": "<org>",
|
| 541 |
"lstrip": false,
|
| 542 |
"normalized": false,
|
|
|
|
| 544 |
"single_word": false,
|
| 545 |
"special": true
|
| 546 |
},
|
| 547 |
+
"60063": {
|
| 548 |
"content": "<date>",
|
| 549 |
"lstrip": false,
|
| 550 |
"normalized": false,
|
|
|
|
| 552 |
"single_word": false,
|
| 553 |
"special": true
|
| 554 |
},
|
| 555 |
+
"60064": {
|
| 556 |
"content": "<eve>",
|
| 557 |
"lstrip": false,
|
| 558 |
"normalized": false,
|
|
|
|
| 560 |
"single_word": false,
|
| 561 |
"special": true
|
| 562 |
},
|
| 563 |
+
"60065": {
|
| 564 |
"content": "<cel>",
|
| 565 |
"lstrip": false,
|
| 566 |
"normalized": false,
|
|
|
|
| 568 |
"single_word": false,
|
| 569 |
"special": true
|
| 570 |
},
|
| 571 |
+
"60066": {
|
| 572 |
"content": "<media>",
|
| 573 |
"lstrip": false,
|
| 574 |
"normalized": false,
|
|
|
|
| 576 |
"single_word": false,
|
| 577 |
"special": true
|
| 578 |
},
|
| 579 |
+
"60067": {
|
| 580 |
"content": "<dis>",
|
| 581 |
"lstrip": false,
|
| 582 |
"normalized": false,
|
|
|
|
| 584 |
"single_word": false,
|
| 585 |
"special": true
|
| 586 |
},
|
| 587 |
+
"60068": {
|
| 588 |
"content": "<concept>",
|
| 589 |
"lstrip": false,
|
| 590 |
"normalized": false,
|