Upload 5 files
Browse files- config.json +296 -0
- model.safetensors +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +60 -0
config.json
ADDED
|
@@ -0,0 +1,296 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/muril-large-cased",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForTokenClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"embedding_size": 1024,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 1024,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "B-art-broadcastprogram",
|
| 14 |
+
"1": "B-art-film",
|
| 15 |
+
"2": "B-art-music",
|
| 16 |
+
"3": "B-art-other",
|
| 17 |
+
"4": "B-art-painting",
|
| 18 |
+
"5": "B-art-writtenart",
|
| 19 |
+
"6": "B-building-airport",
|
| 20 |
+
"7": "B-building-hospital",
|
| 21 |
+
"8": "B-building-hotel",
|
| 22 |
+
"9": "B-building-library",
|
| 23 |
+
"10": "B-building-other",
|
| 24 |
+
"11": "B-building-restaurant",
|
| 25 |
+
"12": "B-building-sportsfacility",
|
| 26 |
+
"13": "B-building-theater",
|
| 27 |
+
"14": "B-event-attack/battle/war/militaryconflict",
|
| 28 |
+
"15": "B-event-disaster",
|
| 29 |
+
"16": "B-event-election",
|
| 30 |
+
"17": "B-event-other",
|
| 31 |
+
"18": "B-event-protest",
|
| 32 |
+
"19": "B-event-sportsevent",
|
| 33 |
+
"20": "B-location-GPE",
|
| 34 |
+
"21": "B-location-bodiesofwater",
|
| 35 |
+
"22": "B-location-island",
|
| 36 |
+
"23": "B-location-mountain",
|
| 37 |
+
"24": "B-location-other",
|
| 38 |
+
"25": "B-location-park",
|
| 39 |
+
"26": "B-location-road/railway/highway/transit",
|
| 40 |
+
"27": "B-organization-company",
|
| 41 |
+
"28": "B-organization-education",
|
| 42 |
+
"29": "B-organization-government/governmentagency",
|
| 43 |
+
"30": "B-organization-media/newspaper",
|
| 44 |
+
"31": "B-organization-other",
|
| 45 |
+
"32": "B-organization-politicalparty",
|
| 46 |
+
"33": "B-organization-religion",
|
| 47 |
+
"34": "B-organization-showorganization",
|
| 48 |
+
"35": "B-organization-sportsleague",
|
| 49 |
+
"36": "B-organization-sportsteam",
|
| 50 |
+
"37": "B-other-astronomything",
|
| 51 |
+
"38": "B-other-award",
|
| 52 |
+
"39": "B-other-biologything",
|
| 53 |
+
"40": "B-other-chemicalthing",
|
| 54 |
+
"41": "B-other-currency",
|
| 55 |
+
"42": "B-other-disease",
|
| 56 |
+
"43": "B-other-educationaldegree",
|
| 57 |
+
"44": "B-other-god",
|
| 58 |
+
"45": "B-other-language",
|
| 59 |
+
"46": "B-other-law",
|
| 60 |
+
"47": "B-other-livingthing",
|
| 61 |
+
"48": "B-other-medical",
|
| 62 |
+
"49": "B-person-actor",
|
| 63 |
+
"50": "B-person-artist/author",
|
| 64 |
+
"51": "B-person-athlete",
|
| 65 |
+
"52": "B-person-director",
|
| 66 |
+
"53": "B-person-other",
|
| 67 |
+
"54": "B-person-politician",
|
| 68 |
+
"55": "B-person-scholar",
|
| 69 |
+
"56": "B-person-soldier",
|
| 70 |
+
"57": "B-product-airplane",
|
| 71 |
+
"58": "B-product-car",
|
| 72 |
+
"59": "B-product-food",
|
| 73 |
+
"60": "B-product-game",
|
| 74 |
+
"61": "B-product-other",
|
| 75 |
+
"62": "B-product-ship",
|
| 76 |
+
"63": "B-product-software",
|
| 77 |
+
"64": "B-product-train",
|
| 78 |
+
"65": "B-product-weapon",
|
| 79 |
+
"66": "I-art-broadcastprogram",
|
| 80 |
+
"67": "I-art-film",
|
| 81 |
+
"68": "I-art-music",
|
| 82 |
+
"69": "I-art-other",
|
| 83 |
+
"70": "I-art-painting",
|
| 84 |
+
"71": "I-art-writtenart",
|
| 85 |
+
"72": "I-building-airport",
|
| 86 |
+
"73": "I-building-hospital",
|
| 87 |
+
"74": "I-building-hotel",
|
| 88 |
+
"75": "I-building-library",
|
| 89 |
+
"76": "I-building-other",
|
| 90 |
+
"77": "I-building-restaurant",
|
| 91 |
+
"78": "I-building-sportsfacility",
|
| 92 |
+
"79": "I-building-theater",
|
| 93 |
+
"80": "I-event-attack/battle/war/militaryconflict",
|
| 94 |
+
"81": "I-event-disaster",
|
| 95 |
+
"82": "I-event-election",
|
| 96 |
+
"83": "I-event-other",
|
| 97 |
+
"84": "I-event-protest",
|
| 98 |
+
"85": "I-event-sportsevent",
|
| 99 |
+
"86": "I-location-GPE",
|
| 100 |
+
"87": "I-location-bodiesofwater",
|
| 101 |
+
"88": "I-location-island",
|
| 102 |
+
"89": "I-location-mountain",
|
| 103 |
+
"90": "I-location-other",
|
| 104 |
+
"91": "I-location-park",
|
| 105 |
+
"92": "I-location-road/railway/highway/transit",
|
| 106 |
+
"93": "I-organization-company",
|
| 107 |
+
"94": "I-organization-education",
|
| 108 |
+
"95": "I-organization-government/governmentagency",
|
| 109 |
+
"96": "I-organization-media/newspaper",
|
| 110 |
+
"97": "I-organization-other",
|
| 111 |
+
"98": "I-organization-politicalparty",
|
| 112 |
+
"99": "I-organization-religion",
|
| 113 |
+
"100": "I-organization-showorganization",
|
| 114 |
+
"101": "I-organization-sportsleague",
|
| 115 |
+
"102": "I-organization-sportsteam",
|
| 116 |
+
"103": "I-other-astronomything",
|
| 117 |
+
"104": "I-other-award",
|
| 118 |
+
"105": "I-other-biologything",
|
| 119 |
+
"106": "I-other-chemicalthing",
|
| 120 |
+
"107": "I-other-currency",
|
| 121 |
+
"108": "I-other-disease",
|
| 122 |
+
"109": "I-other-educationaldegree",
|
| 123 |
+
"110": "I-other-god",
|
| 124 |
+
"111": "I-other-language",
|
| 125 |
+
"112": "I-other-law",
|
| 126 |
+
"113": "I-other-livingthing",
|
| 127 |
+
"114": "I-other-medical",
|
| 128 |
+
"115": "I-person-actor",
|
| 129 |
+
"116": "I-person-artist/author",
|
| 130 |
+
"117": "I-person-athlete",
|
| 131 |
+
"118": "I-person-director",
|
| 132 |
+
"119": "I-person-other",
|
| 133 |
+
"120": "I-person-politician",
|
| 134 |
+
"121": "I-person-scholar",
|
| 135 |
+
"122": "I-person-soldier",
|
| 136 |
+
"123": "I-product-airplane",
|
| 137 |
+
"124": "I-product-car",
|
| 138 |
+
"125": "I-product-food",
|
| 139 |
+
"126": "I-product-game",
|
| 140 |
+
"127": "I-product-other",
|
| 141 |
+
"128": "I-product-ship",
|
| 142 |
+
"129": "I-product-software",
|
| 143 |
+
"130": "I-product-train",
|
| 144 |
+
"131": "I-product-weapon",
|
| 145 |
+
"132": "O"
|
| 146 |
+
},
|
| 147 |
+
"initializer_range": 0.02,
|
| 148 |
+
"intermediate_size": 4096,
|
| 149 |
+
"label2id": {
|
| 150 |
+
"B-art-broadcastprogram": 0,
|
| 151 |
+
"B-art-film": 1,
|
| 152 |
+
"B-art-music": 2,
|
| 153 |
+
"B-art-other": 3,
|
| 154 |
+
"B-art-painting": 4,
|
| 155 |
+
"B-art-writtenart": 5,
|
| 156 |
+
"B-building-airport": 6,
|
| 157 |
+
"B-building-hospital": 7,
|
| 158 |
+
"B-building-hotel": 8,
|
| 159 |
+
"B-building-library": 9,
|
| 160 |
+
"B-building-other": 10,
|
| 161 |
+
"B-building-restaurant": 11,
|
| 162 |
+
"B-building-sportsfacility": 12,
|
| 163 |
+
"B-building-theater": 13,
|
| 164 |
+
"B-event-attack/battle/war/militaryconflict": 14,
|
| 165 |
+
"B-event-disaster": 15,
|
| 166 |
+
"B-event-election": 16,
|
| 167 |
+
"B-event-other": 17,
|
| 168 |
+
"B-event-protest": 18,
|
| 169 |
+
"B-event-sportsevent": 19,
|
| 170 |
+
"B-location-GPE": 20,
|
| 171 |
+
"B-location-bodiesofwater": 21,
|
| 172 |
+
"B-location-island": 22,
|
| 173 |
+
"B-location-mountain": 23,
|
| 174 |
+
"B-location-other": 24,
|
| 175 |
+
"B-location-park": 25,
|
| 176 |
+
"B-location-road/railway/highway/transit": 26,
|
| 177 |
+
"B-organization-company": 27,
|
| 178 |
+
"B-organization-education": 28,
|
| 179 |
+
"B-organization-government/governmentagency": 29,
|
| 180 |
+
"B-organization-media/newspaper": 30,
|
| 181 |
+
"B-organization-other": 31,
|
| 182 |
+
"B-organization-politicalparty": 32,
|
| 183 |
+
"B-organization-religion": 33,
|
| 184 |
+
"B-organization-showorganization": 34,
|
| 185 |
+
"B-organization-sportsleague": 35,
|
| 186 |
+
"B-organization-sportsteam": 36,
|
| 187 |
+
"B-other-astronomything": 37,
|
| 188 |
+
"B-other-award": 38,
|
| 189 |
+
"B-other-biologything": 39,
|
| 190 |
+
"B-other-chemicalthing": 40,
|
| 191 |
+
"B-other-currency": 41,
|
| 192 |
+
"B-other-disease": 42,
|
| 193 |
+
"B-other-educationaldegree": 43,
|
| 194 |
+
"B-other-god": 44,
|
| 195 |
+
"B-other-language": 45,
|
| 196 |
+
"B-other-law": 46,
|
| 197 |
+
"B-other-livingthing": 47,
|
| 198 |
+
"B-other-medical": 48,
|
| 199 |
+
"B-person-actor": 49,
|
| 200 |
+
"B-person-artist/author": 50,
|
| 201 |
+
"B-person-athlete": 51,
|
| 202 |
+
"B-person-director": 52,
|
| 203 |
+
"B-person-other": 53,
|
| 204 |
+
"B-person-politician": 54,
|
| 205 |
+
"B-person-scholar": 55,
|
| 206 |
+
"B-person-soldier": 56,
|
| 207 |
+
"B-product-airplane": 57,
|
| 208 |
+
"B-product-car": 58,
|
| 209 |
+
"B-product-food": 59,
|
| 210 |
+
"B-product-game": 60,
|
| 211 |
+
"B-product-other": 61,
|
| 212 |
+
"B-product-ship": 62,
|
| 213 |
+
"B-product-software": 63,
|
| 214 |
+
"B-product-train": 64,
|
| 215 |
+
"B-product-weapon": 65,
|
| 216 |
+
"I-art-broadcastprogram": 66,
|
| 217 |
+
"I-art-film": 67,
|
| 218 |
+
"I-art-music": 68,
|
| 219 |
+
"I-art-other": 69,
|
| 220 |
+
"I-art-painting": 70,
|
| 221 |
+
"I-art-writtenart": 71,
|
| 222 |
+
"I-building-airport": 72,
|
| 223 |
+
"I-building-hospital": 73,
|
| 224 |
+
"I-building-hotel": 74,
|
| 225 |
+
"I-building-library": 75,
|
| 226 |
+
"I-building-other": 76,
|
| 227 |
+
"I-building-restaurant": 77,
|
| 228 |
+
"I-building-sportsfacility": 78,
|
| 229 |
+
"I-building-theater": 79,
|
| 230 |
+
"I-event-attack/battle/war/militaryconflict": 80,
|
| 231 |
+
"I-event-disaster": 81,
|
| 232 |
+
"I-event-election": 82,
|
| 233 |
+
"I-event-other": 83,
|
| 234 |
+
"I-event-protest": 84,
|
| 235 |
+
"I-event-sportsevent": 85,
|
| 236 |
+
"I-location-GPE": 86,
|
| 237 |
+
"I-location-bodiesofwater": 87,
|
| 238 |
+
"I-location-island": 88,
|
| 239 |
+
"I-location-mountain": 89,
|
| 240 |
+
"I-location-other": 90,
|
| 241 |
+
"I-location-park": 91,
|
| 242 |
+
"I-location-road/railway/highway/transit": 92,
|
| 243 |
+
"I-organization-company": 93,
|
| 244 |
+
"I-organization-education": 94,
|
| 245 |
+
"I-organization-government/governmentagency": 95,
|
| 246 |
+
"I-organization-media/newspaper": 96,
|
| 247 |
+
"I-organization-other": 97,
|
| 248 |
+
"I-organization-politicalparty": 98,
|
| 249 |
+
"I-organization-religion": 99,
|
| 250 |
+
"I-organization-showorganization": 100,
|
| 251 |
+
"I-organization-sportsleague": 101,
|
| 252 |
+
"I-organization-sportsteam": 102,
|
| 253 |
+
"I-other-astronomything": 103,
|
| 254 |
+
"I-other-award": 104,
|
| 255 |
+
"I-other-biologything": 105,
|
| 256 |
+
"I-other-chemicalthing": 106,
|
| 257 |
+
"I-other-currency": 107,
|
| 258 |
+
"I-other-disease": 108,
|
| 259 |
+
"I-other-educationaldegree": 109,
|
| 260 |
+
"I-other-god": 110,
|
| 261 |
+
"I-other-language": 111,
|
| 262 |
+
"I-other-law": 112,
|
| 263 |
+
"I-other-livingthing": 113,
|
| 264 |
+
"I-other-medical": 114,
|
| 265 |
+
"I-person-actor": 115,
|
| 266 |
+
"I-person-artist/author": 116,
|
| 267 |
+
"I-person-athlete": 117,
|
| 268 |
+
"I-person-director": 118,
|
| 269 |
+
"I-person-other": 119,
|
| 270 |
+
"I-person-politician": 120,
|
| 271 |
+
"I-person-scholar": 121,
|
| 272 |
+
"I-person-soldier": 122,
|
| 273 |
+
"I-product-airplane": 123,
|
| 274 |
+
"I-product-car": 124,
|
| 275 |
+
"I-product-food": 125,
|
| 276 |
+
"I-product-game": 126,
|
| 277 |
+
"I-product-other": 127,
|
| 278 |
+
"I-product-ship": 128,
|
| 279 |
+
"I-product-software": 129,
|
| 280 |
+
"I-product-train": 130,
|
| 281 |
+
"I-product-weapon": 131,
|
| 282 |
+
"O": 132
|
| 283 |
+
},
|
| 284 |
+
"layer_norm_eps": 1e-12,
|
| 285 |
+
"max_position_embeddings": 512,
|
| 286 |
+
"model_type": "bert",
|
| 287 |
+
"num_attention_heads": 16,
|
| 288 |
+
"num_hidden_layers": 24,
|
| 289 |
+
"pad_token_id": 0,
|
| 290 |
+
"position_embedding_type": "absolute",
|
| 291 |
+
"torch_dtype": "float32",
|
| 292 |
+
"transformers_version": "4.48.2",
|
| 293 |
+
"type_vocab_size": 2,
|
| 294 |
+
"use_cache": true,
|
| 295 |
+
"vocab_size": 197285
|
| 296 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:efb8cec58ef655e8bb1f597a2a32cc59efd25b58098e951c1d883fa3d6ad7e21
|
| 3 |
+
size 2020022956
|
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.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": true,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"0": {
|
| 5 |
+
"content": "[PAD]",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": false,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
},
|
| 12 |
+
"100": {
|
| 13 |
+
"content": "[UNK]",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": false,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false,
|
| 18 |
+
"special": true
|
| 19 |
+
},
|
| 20 |
+
"103": {
|
| 21 |
+
"content": "[MASK]",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": false,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
},
|
| 28 |
+
"104": {
|
| 29 |
+
"content": "[CLS]",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": false,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false,
|
| 34 |
+
"special": true
|
| 35 |
+
},
|
| 36 |
+
"105": {
|
| 37 |
+
"content": "[SEP]",
|
| 38 |
+
"lstrip": false,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false,
|
| 42 |
+
"special": true
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
"clean_up_tokenization_spaces": true,
|
| 46 |
+
"cls_token": "[CLS]",
|
| 47 |
+
"do_basic_tokenize": true,
|
| 48 |
+
"do_lower_case": false,
|
| 49 |
+
"extra_special_tokens": {},
|
| 50 |
+
"lowercase": false,
|
| 51 |
+
"mask_token": "[MASK]",
|
| 52 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 53 |
+
"never_split": null,
|
| 54 |
+
"pad_token": "[PAD]",
|
| 55 |
+
"sep_token": "[SEP]",
|
| 56 |
+
"strip_accents": false,
|
| 57 |
+
"tokenize_chinese_chars": true,
|
| 58 |
+
"tokenizer_class": "BertTokenizer",
|
| 59 |
+
"unk_token": "[UNK]"
|
| 60 |
+
}
|