Upload folder using huggingface_hub
Browse files- gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/config.json +372 -0
- gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/model.safetensors +3 -0
- gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/special_tokens_map.json +7 -0
- gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/tokenizer.json +0 -0
- gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/tokenizer_config.json +57 -0
- gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/training_args.bin +3 -0
- gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/training_meta.json +180 -0
- gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/vocab.txt +0 -0
gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/config.json
ADDED
|
@@ -0,0 +1,372 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForTokenClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"directionality": "bidi",
|
| 8 |
+
"dtype": "float32",
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "O",
|
| 14 |
+
"1": "B-ACCOUNT_NUMBER",
|
| 15 |
+
"2": "I-ACCOUNT_NUMBER",
|
| 16 |
+
"3": "E-ACCOUNT_NUMBER",
|
| 17 |
+
"4": "S-ACCOUNT_NUMBER",
|
| 18 |
+
"5": "B-ADDRESS",
|
| 19 |
+
"6": "I-ADDRESS",
|
| 20 |
+
"7": "E-ADDRESS",
|
| 21 |
+
"8": "S-ADDRESS",
|
| 22 |
+
"9": "B-API_KEY",
|
| 23 |
+
"10": "I-API_KEY",
|
| 24 |
+
"11": "E-API_KEY",
|
| 25 |
+
"12": "S-API_KEY",
|
| 26 |
+
"13": "B-BANK_ROUTING_NUMBER",
|
| 27 |
+
"14": "I-BANK_ROUTING_NUMBER",
|
| 28 |
+
"15": "E-BANK_ROUTING_NUMBER",
|
| 29 |
+
"16": "S-BANK_ROUTING_NUMBER",
|
| 30 |
+
"17": "B-BIOMETRIC_IDENTIFIER",
|
| 31 |
+
"18": "I-BIOMETRIC_IDENTIFIER",
|
| 32 |
+
"19": "E-BIOMETRIC_IDENTIFIER",
|
| 33 |
+
"20": "S-BIOMETRIC_IDENTIFIER",
|
| 34 |
+
"21": "B-CERTIFICATE_LICENSE_NUMBER",
|
| 35 |
+
"22": "I-CERTIFICATE_LICENSE_NUMBER",
|
| 36 |
+
"23": "E-CERTIFICATE_LICENSE_NUMBER",
|
| 37 |
+
"24": "S-CERTIFICATE_LICENSE_NUMBER",
|
| 38 |
+
"25": "B-CITY",
|
| 39 |
+
"26": "I-CITY",
|
| 40 |
+
"27": "E-CITY",
|
| 41 |
+
"28": "S-CITY",
|
| 42 |
+
"29": "B-COMPANY_NAME",
|
| 43 |
+
"30": "I-COMPANY_NAME",
|
| 44 |
+
"31": "E-COMPANY_NAME",
|
| 45 |
+
"32": "S-COMPANY_NAME",
|
| 46 |
+
"33": "B-COORDINATE",
|
| 47 |
+
"34": "I-COORDINATE",
|
| 48 |
+
"35": "E-COORDINATE",
|
| 49 |
+
"36": "S-COORDINATE",
|
| 50 |
+
"37": "B-COUNTRY",
|
| 51 |
+
"38": "I-COUNTRY",
|
| 52 |
+
"39": "E-COUNTRY",
|
| 53 |
+
"40": "S-COUNTRY",
|
| 54 |
+
"41": "B-CREDIT_CARD_NUMBER",
|
| 55 |
+
"42": "I-CREDIT_CARD_NUMBER",
|
| 56 |
+
"43": "E-CREDIT_CARD_NUMBER",
|
| 57 |
+
"44": "S-CREDIT_CARD_NUMBER",
|
| 58 |
+
"45": "B-CUSTOMER_ID",
|
| 59 |
+
"46": "I-CUSTOMER_ID",
|
| 60 |
+
"47": "E-CUSTOMER_ID",
|
| 61 |
+
"48": "S-CUSTOMER_ID",
|
| 62 |
+
"49": "B-CVV",
|
| 63 |
+
"50": "I-CVV",
|
| 64 |
+
"51": "E-CVV",
|
| 65 |
+
"52": "S-CVV",
|
| 66 |
+
"53": "B-DATE",
|
| 67 |
+
"54": "I-DATE",
|
| 68 |
+
"55": "E-DATE",
|
| 69 |
+
"56": "S-DATE",
|
| 70 |
+
"57": "B-DATE_OF_BIRTH",
|
| 71 |
+
"58": "I-DATE_OF_BIRTH",
|
| 72 |
+
"59": "E-DATE_OF_BIRTH",
|
| 73 |
+
"60": "S-DATE_OF_BIRTH",
|
| 74 |
+
"61": "B-DATE_TIME",
|
| 75 |
+
"62": "I-DATE_TIME",
|
| 76 |
+
"63": "E-DATE_TIME",
|
| 77 |
+
"64": "S-DATE_TIME",
|
| 78 |
+
"65": "B-DEVICE_IDENTIFIER",
|
| 79 |
+
"66": "I-DEVICE_IDENTIFIER",
|
| 80 |
+
"67": "E-DEVICE_IDENTIFIER",
|
| 81 |
+
"68": "S-DEVICE_IDENTIFIER",
|
| 82 |
+
"69": "B-EMAIL",
|
| 83 |
+
"70": "I-EMAIL",
|
| 84 |
+
"71": "E-EMAIL",
|
| 85 |
+
"72": "S-EMAIL",
|
| 86 |
+
"73": "B-EMPLOYEE_ID",
|
| 87 |
+
"74": "I-EMPLOYEE_ID",
|
| 88 |
+
"75": "E-EMPLOYEE_ID",
|
| 89 |
+
"76": "S-EMPLOYEE_ID",
|
| 90 |
+
"77": "B-FIRST_NAME",
|
| 91 |
+
"78": "I-FIRST_NAME",
|
| 92 |
+
"79": "E-FIRST_NAME",
|
| 93 |
+
"80": "S-FIRST_NAME",
|
| 94 |
+
"81": "B-HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 95 |
+
"82": "I-HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 96 |
+
"83": "E-HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 97 |
+
"84": "S-HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 98 |
+
"85": "B-IPV4",
|
| 99 |
+
"86": "I-IPV4",
|
| 100 |
+
"87": "E-IPV4",
|
| 101 |
+
"88": "S-IPV4",
|
| 102 |
+
"89": "B-IPV6",
|
| 103 |
+
"90": "I-IPV6",
|
| 104 |
+
"91": "E-IPV6",
|
| 105 |
+
"92": "S-IPV6",
|
| 106 |
+
"93": "B-LAST_NAME",
|
| 107 |
+
"94": "I-LAST_NAME",
|
| 108 |
+
"95": "E-LAST_NAME",
|
| 109 |
+
"96": "S-LAST_NAME",
|
| 110 |
+
"97": "B-LICENSE_PLATE",
|
| 111 |
+
"98": "I-LICENSE_PLATE",
|
| 112 |
+
"99": "E-LICENSE_PLATE",
|
| 113 |
+
"100": "S-LICENSE_PLATE",
|
| 114 |
+
"101": "B-MEDICAL_RECORD_NUMBER",
|
| 115 |
+
"102": "I-MEDICAL_RECORD_NUMBER",
|
| 116 |
+
"103": "E-MEDICAL_RECORD_NUMBER",
|
| 117 |
+
"104": "S-MEDICAL_RECORD_NUMBER",
|
| 118 |
+
"105": "B-NAME",
|
| 119 |
+
"106": "I-NAME",
|
| 120 |
+
"107": "E-NAME",
|
| 121 |
+
"108": "S-NAME",
|
| 122 |
+
"109": "B-NATIONAL_ID",
|
| 123 |
+
"110": "I-NATIONAL_ID",
|
| 124 |
+
"111": "E-NATIONAL_ID",
|
| 125 |
+
"112": "S-NATIONAL_ID",
|
| 126 |
+
"113": "B-PASSWORD",
|
| 127 |
+
"114": "I-PASSWORD",
|
| 128 |
+
"115": "E-PASSWORD",
|
| 129 |
+
"116": "S-PASSWORD",
|
| 130 |
+
"117": "B-PHONE_NUMBER",
|
| 131 |
+
"118": "I-PHONE_NUMBER",
|
| 132 |
+
"119": "E-PHONE_NUMBER",
|
| 133 |
+
"120": "S-PHONE_NUMBER",
|
| 134 |
+
"121": "B-PIN",
|
| 135 |
+
"122": "I-PIN",
|
| 136 |
+
"123": "E-PIN",
|
| 137 |
+
"124": "S-PIN",
|
| 138 |
+
"125": "B-POSTCODE",
|
| 139 |
+
"126": "I-POSTCODE",
|
| 140 |
+
"127": "E-POSTCODE",
|
| 141 |
+
"128": "S-POSTCODE",
|
| 142 |
+
"129": "B-SSN",
|
| 143 |
+
"130": "I-SSN",
|
| 144 |
+
"131": "E-SSN",
|
| 145 |
+
"132": "S-SSN",
|
| 146 |
+
"133": "B-STATE",
|
| 147 |
+
"134": "I-STATE",
|
| 148 |
+
"135": "E-STATE",
|
| 149 |
+
"136": "S-STATE",
|
| 150 |
+
"137": "B-STREET_ADDRESS",
|
| 151 |
+
"138": "I-STREET_ADDRESS",
|
| 152 |
+
"139": "E-STREET_ADDRESS",
|
| 153 |
+
"140": "S-STREET_ADDRESS",
|
| 154 |
+
"141": "B-SWIFT_BIC",
|
| 155 |
+
"142": "I-SWIFT_BIC",
|
| 156 |
+
"143": "E-SWIFT_BIC",
|
| 157 |
+
"144": "S-SWIFT_BIC",
|
| 158 |
+
"145": "B-TAX_ID",
|
| 159 |
+
"146": "I-TAX_ID",
|
| 160 |
+
"147": "E-TAX_ID",
|
| 161 |
+
"148": "S-TAX_ID",
|
| 162 |
+
"149": "B-TIME",
|
| 163 |
+
"150": "I-TIME",
|
| 164 |
+
"151": "E-TIME",
|
| 165 |
+
"152": "S-TIME",
|
| 166 |
+
"153": "B-UNIQUE_IDENTIFIER",
|
| 167 |
+
"154": "I-UNIQUE_IDENTIFIER",
|
| 168 |
+
"155": "E-UNIQUE_IDENTIFIER",
|
| 169 |
+
"156": "S-UNIQUE_IDENTIFIER",
|
| 170 |
+
"157": "B-URL",
|
| 171 |
+
"158": "I-URL",
|
| 172 |
+
"159": "E-URL",
|
| 173 |
+
"160": "S-URL",
|
| 174 |
+
"161": "B-USER_NAME",
|
| 175 |
+
"162": "I-USER_NAME",
|
| 176 |
+
"163": "E-USER_NAME",
|
| 177 |
+
"164": "S-USER_NAME",
|
| 178 |
+
"165": "B-VEHICLE_IDENTIFIER",
|
| 179 |
+
"166": "I-VEHICLE_IDENTIFIER",
|
| 180 |
+
"167": "E-VEHICLE_IDENTIFIER",
|
| 181 |
+
"168": "S-VEHICLE_IDENTIFIER"
|
| 182 |
+
},
|
| 183 |
+
"initializer_range": 0.02,
|
| 184 |
+
"intermediate_size": 3072,
|
| 185 |
+
"label2id": {
|
| 186 |
+
"B-ACCOUNT_NUMBER": 1,
|
| 187 |
+
"B-ADDRESS": 5,
|
| 188 |
+
"B-API_KEY": 9,
|
| 189 |
+
"B-BANK_ROUTING_NUMBER": 13,
|
| 190 |
+
"B-BIOMETRIC_IDENTIFIER": 17,
|
| 191 |
+
"B-CERTIFICATE_LICENSE_NUMBER": 21,
|
| 192 |
+
"B-CITY": 25,
|
| 193 |
+
"B-COMPANY_NAME": 29,
|
| 194 |
+
"B-COORDINATE": 33,
|
| 195 |
+
"B-COUNTRY": 37,
|
| 196 |
+
"B-CREDIT_CARD_NUMBER": 41,
|
| 197 |
+
"B-CUSTOMER_ID": 45,
|
| 198 |
+
"B-CVV": 49,
|
| 199 |
+
"B-DATE": 53,
|
| 200 |
+
"B-DATE_OF_BIRTH": 57,
|
| 201 |
+
"B-DATE_TIME": 61,
|
| 202 |
+
"B-DEVICE_IDENTIFIER": 65,
|
| 203 |
+
"B-EMAIL": 69,
|
| 204 |
+
"B-EMPLOYEE_ID": 73,
|
| 205 |
+
"B-FIRST_NAME": 77,
|
| 206 |
+
"B-HEALTH_PLAN_BENEFICIARY_NUMBER": 81,
|
| 207 |
+
"B-IPV4": 85,
|
| 208 |
+
"B-IPV6": 89,
|
| 209 |
+
"B-LAST_NAME": 93,
|
| 210 |
+
"B-LICENSE_PLATE": 97,
|
| 211 |
+
"B-MEDICAL_RECORD_NUMBER": 101,
|
| 212 |
+
"B-NAME": 105,
|
| 213 |
+
"B-NATIONAL_ID": 109,
|
| 214 |
+
"B-PASSWORD": 113,
|
| 215 |
+
"B-PHONE_NUMBER": 117,
|
| 216 |
+
"B-PIN": 121,
|
| 217 |
+
"B-POSTCODE": 125,
|
| 218 |
+
"B-SSN": 129,
|
| 219 |
+
"B-STATE": 133,
|
| 220 |
+
"B-STREET_ADDRESS": 137,
|
| 221 |
+
"B-SWIFT_BIC": 141,
|
| 222 |
+
"B-TAX_ID": 145,
|
| 223 |
+
"B-TIME": 149,
|
| 224 |
+
"B-UNIQUE_IDENTIFIER": 153,
|
| 225 |
+
"B-URL": 157,
|
| 226 |
+
"B-USER_NAME": 161,
|
| 227 |
+
"B-VEHICLE_IDENTIFIER": 165,
|
| 228 |
+
"E-ACCOUNT_NUMBER": 3,
|
| 229 |
+
"E-ADDRESS": 7,
|
| 230 |
+
"E-API_KEY": 11,
|
| 231 |
+
"E-BANK_ROUTING_NUMBER": 15,
|
| 232 |
+
"E-BIOMETRIC_IDENTIFIER": 19,
|
| 233 |
+
"E-CERTIFICATE_LICENSE_NUMBER": 23,
|
| 234 |
+
"E-CITY": 27,
|
| 235 |
+
"E-COMPANY_NAME": 31,
|
| 236 |
+
"E-COORDINATE": 35,
|
| 237 |
+
"E-COUNTRY": 39,
|
| 238 |
+
"E-CREDIT_CARD_NUMBER": 43,
|
| 239 |
+
"E-CUSTOMER_ID": 47,
|
| 240 |
+
"E-CVV": 51,
|
| 241 |
+
"E-DATE": 55,
|
| 242 |
+
"E-DATE_OF_BIRTH": 59,
|
| 243 |
+
"E-DATE_TIME": 63,
|
| 244 |
+
"E-DEVICE_IDENTIFIER": 67,
|
| 245 |
+
"E-EMAIL": 71,
|
| 246 |
+
"E-EMPLOYEE_ID": 75,
|
| 247 |
+
"E-FIRST_NAME": 79,
|
| 248 |
+
"E-HEALTH_PLAN_BENEFICIARY_NUMBER": 83,
|
| 249 |
+
"E-IPV4": 87,
|
| 250 |
+
"E-IPV6": 91,
|
| 251 |
+
"E-LAST_NAME": 95,
|
| 252 |
+
"E-LICENSE_PLATE": 99,
|
| 253 |
+
"E-MEDICAL_RECORD_NUMBER": 103,
|
| 254 |
+
"E-NAME": 107,
|
| 255 |
+
"E-NATIONAL_ID": 111,
|
| 256 |
+
"E-PASSWORD": 115,
|
| 257 |
+
"E-PHONE_NUMBER": 119,
|
| 258 |
+
"E-PIN": 123,
|
| 259 |
+
"E-POSTCODE": 127,
|
| 260 |
+
"E-SSN": 131,
|
| 261 |
+
"E-STATE": 135,
|
| 262 |
+
"E-STREET_ADDRESS": 139,
|
| 263 |
+
"E-SWIFT_BIC": 143,
|
| 264 |
+
"E-TAX_ID": 147,
|
| 265 |
+
"E-TIME": 151,
|
| 266 |
+
"E-UNIQUE_IDENTIFIER": 155,
|
| 267 |
+
"E-URL": 159,
|
| 268 |
+
"E-USER_NAME": 163,
|
| 269 |
+
"E-VEHICLE_IDENTIFIER": 167,
|
| 270 |
+
"I-ACCOUNT_NUMBER": 2,
|
| 271 |
+
"I-ADDRESS": 6,
|
| 272 |
+
"I-API_KEY": 10,
|
| 273 |
+
"I-BANK_ROUTING_NUMBER": 14,
|
| 274 |
+
"I-BIOMETRIC_IDENTIFIER": 18,
|
| 275 |
+
"I-CERTIFICATE_LICENSE_NUMBER": 22,
|
| 276 |
+
"I-CITY": 26,
|
| 277 |
+
"I-COMPANY_NAME": 30,
|
| 278 |
+
"I-COORDINATE": 34,
|
| 279 |
+
"I-COUNTRY": 38,
|
| 280 |
+
"I-CREDIT_CARD_NUMBER": 42,
|
| 281 |
+
"I-CUSTOMER_ID": 46,
|
| 282 |
+
"I-CVV": 50,
|
| 283 |
+
"I-DATE": 54,
|
| 284 |
+
"I-DATE_OF_BIRTH": 58,
|
| 285 |
+
"I-DATE_TIME": 62,
|
| 286 |
+
"I-DEVICE_IDENTIFIER": 66,
|
| 287 |
+
"I-EMAIL": 70,
|
| 288 |
+
"I-EMPLOYEE_ID": 74,
|
| 289 |
+
"I-FIRST_NAME": 78,
|
| 290 |
+
"I-HEALTH_PLAN_BENEFICIARY_NUMBER": 82,
|
| 291 |
+
"I-IPV4": 86,
|
| 292 |
+
"I-IPV6": 90,
|
| 293 |
+
"I-LAST_NAME": 94,
|
| 294 |
+
"I-LICENSE_PLATE": 98,
|
| 295 |
+
"I-MEDICAL_RECORD_NUMBER": 102,
|
| 296 |
+
"I-NAME": 106,
|
| 297 |
+
"I-NATIONAL_ID": 110,
|
| 298 |
+
"I-PASSWORD": 114,
|
| 299 |
+
"I-PHONE_NUMBER": 118,
|
| 300 |
+
"I-PIN": 122,
|
| 301 |
+
"I-POSTCODE": 126,
|
| 302 |
+
"I-SSN": 130,
|
| 303 |
+
"I-STATE": 134,
|
| 304 |
+
"I-STREET_ADDRESS": 138,
|
| 305 |
+
"I-SWIFT_BIC": 142,
|
| 306 |
+
"I-TAX_ID": 146,
|
| 307 |
+
"I-TIME": 150,
|
| 308 |
+
"I-UNIQUE_IDENTIFIER": 154,
|
| 309 |
+
"I-URL": 158,
|
| 310 |
+
"I-USER_NAME": 162,
|
| 311 |
+
"I-VEHICLE_IDENTIFIER": 166,
|
| 312 |
+
"O": 0,
|
| 313 |
+
"S-ACCOUNT_NUMBER": 4,
|
| 314 |
+
"S-ADDRESS": 8,
|
| 315 |
+
"S-API_KEY": 12,
|
| 316 |
+
"S-BANK_ROUTING_NUMBER": 16,
|
| 317 |
+
"S-BIOMETRIC_IDENTIFIER": 20,
|
| 318 |
+
"S-CERTIFICATE_LICENSE_NUMBER": 24,
|
| 319 |
+
"S-CITY": 28,
|
| 320 |
+
"S-COMPANY_NAME": 32,
|
| 321 |
+
"S-COORDINATE": 36,
|
| 322 |
+
"S-COUNTRY": 40,
|
| 323 |
+
"S-CREDIT_CARD_NUMBER": 44,
|
| 324 |
+
"S-CUSTOMER_ID": 48,
|
| 325 |
+
"S-CVV": 52,
|
| 326 |
+
"S-DATE": 56,
|
| 327 |
+
"S-DATE_OF_BIRTH": 60,
|
| 328 |
+
"S-DATE_TIME": 64,
|
| 329 |
+
"S-DEVICE_IDENTIFIER": 68,
|
| 330 |
+
"S-EMAIL": 72,
|
| 331 |
+
"S-EMPLOYEE_ID": 76,
|
| 332 |
+
"S-FIRST_NAME": 80,
|
| 333 |
+
"S-HEALTH_PLAN_BENEFICIARY_NUMBER": 84,
|
| 334 |
+
"S-IPV4": 88,
|
| 335 |
+
"S-IPV6": 92,
|
| 336 |
+
"S-LAST_NAME": 96,
|
| 337 |
+
"S-LICENSE_PLATE": 100,
|
| 338 |
+
"S-MEDICAL_RECORD_NUMBER": 104,
|
| 339 |
+
"S-NAME": 108,
|
| 340 |
+
"S-NATIONAL_ID": 112,
|
| 341 |
+
"S-PASSWORD": 116,
|
| 342 |
+
"S-PHONE_NUMBER": 120,
|
| 343 |
+
"S-PIN": 124,
|
| 344 |
+
"S-POSTCODE": 128,
|
| 345 |
+
"S-SSN": 132,
|
| 346 |
+
"S-STATE": 136,
|
| 347 |
+
"S-STREET_ADDRESS": 140,
|
| 348 |
+
"S-SWIFT_BIC": 144,
|
| 349 |
+
"S-TAX_ID": 148,
|
| 350 |
+
"S-TIME": 152,
|
| 351 |
+
"S-UNIQUE_IDENTIFIER": 156,
|
| 352 |
+
"S-URL": 160,
|
| 353 |
+
"S-USER_NAME": 164,
|
| 354 |
+
"S-VEHICLE_IDENTIFIER": 168
|
| 355 |
+
},
|
| 356 |
+
"layer_norm_eps": 1e-12,
|
| 357 |
+
"max_position_embeddings": 512,
|
| 358 |
+
"model_type": "bert",
|
| 359 |
+
"num_attention_heads": 12,
|
| 360 |
+
"num_hidden_layers": 12,
|
| 361 |
+
"pad_token_id": 0,
|
| 362 |
+
"pooler_fc_size": 768,
|
| 363 |
+
"pooler_num_attention_heads": 12,
|
| 364 |
+
"pooler_num_fc_layers": 3,
|
| 365 |
+
"pooler_size_per_head": 128,
|
| 366 |
+
"pooler_type": "first_token_transform",
|
| 367 |
+
"position_embedding_type": "absolute",
|
| 368 |
+
"transformers_version": "4.57.6",
|
| 369 |
+
"type_vocab_size": 2,
|
| 370 |
+
"use_cache": true,
|
| 371 |
+
"vocab_size": 119547
|
| 372 |
+
}
|
gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a3fa1c511a5e71c718851361d5129a0d503ed3ab2d294521056592de4aed79a7
|
| 3 |
+
size 709594596
|
gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/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 |
+
}
|
gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/tokenizer_config.json
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"101": {
|
| 21 |
+
"content": "[CLS]",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": false,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
},
|
| 28 |
+
"102": {
|
| 29 |
+
"content": "[SEP]",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": false,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false,
|
| 34 |
+
"special": true
|
| 35 |
+
},
|
| 36 |
+
"103": {
|
| 37 |
+
"content": "[MASK]",
|
| 38 |
+
"lstrip": false,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false,
|
| 42 |
+
"special": true
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "[CLS]",
|
| 47 |
+
"do_lower_case": false,
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "[MASK]",
|
| 50 |
+
"model_max_length": 512,
|
| 51 |
+
"pad_token": "[PAD]",
|
| 52 |
+
"sep_token": "[SEP]",
|
| 53 |
+
"strip_accents": null,
|
| 54 |
+
"tokenize_chinese_chars": true,
|
| 55 |
+
"tokenizer_class": "BertTokenizer",
|
| 56 |
+
"unk_token": "[UNK]"
|
| 57 |
+
}
|
gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:26d0f0f4f714fd5cfffe9a5be51e1e334a53131e42b975b6c0de5012656f36be
|
| 3 |
+
size 5905
|
gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/training_meta.json
ADDED
|
@@ -0,0 +1,180 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architecture": "token_based",
|
| 3 |
+
"base_model": "google-bert/bert-base-multilingual-cased",
|
| 4 |
+
"tagging_scheme": "bioes",
|
| 5 |
+
"use_crf": false,
|
| 6 |
+
"dataset": "gretel-pii-ready",
|
| 7 |
+
"label_list": [
|
| 8 |
+
"O",
|
| 9 |
+
"B-ACCOUNT_NUMBER",
|
| 10 |
+
"I-ACCOUNT_NUMBER",
|
| 11 |
+
"E-ACCOUNT_NUMBER",
|
| 12 |
+
"S-ACCOUNT_NUMBER",
|
| 13 |
+
"B-ADDRESS",
|
| 14 |
+
"I-ADDRESS",
|
| 15 |
+
"E-ADDRESS",
|
| 16 |
+
"S-ADDRESS",
|
| 17 |
+
"B-API_KEY",
|
| 18 |
+
"I-API_KEY",
|
| 19 |
+
"E-API_KEY",
|
| 20 |
+
"S-API_KEY",
|
| 21 |
+
"B-BANK_ROUTING_NUMBER",
|
| 22 |
+
"I-BANK_ROUTING_NUMBER",
|
| 23 |
+
"E-BANK_ROUTING_NUMBER",
|
| 24 |
+
"S-BANK_ROUTING_NUMBER",
|
| 25 |
+
"B-BIOMETRIC_IDENTIFIER",
|
| 26 |
+
"I-BIOMETRIC_IDENTIFIER",
|
| 27 |
+
"E-BIOMETRIC_IDENTIFIER",
|
| 28 |
+
"S-BIOMETRIC_IDENTIFIER",
|
| 29 |
+
"B-CERTIFICATE_LICENSE_NUMBER",
|
| 30 |
+
"I-CERTIFICATE_LICENSE_NUMBER",
|
| 31 |
+
"E-CERTIFICATE_LICENSE_NUMBER",
|
| 32 |
+
"S-CERTIFICATE_LICENSE_NUMBER",
|
| 33 |
+
"B-CITY",
|
| 34 |
+
"I-CITY",
|
| 35 |
+
"E-CITY",
|
| 36 |
+
"S-CITY",
|
| 37 |
+
"B-COMPANY_NAME",
|
| 38 |
+
"I-COMPANY_NAME",
|
| 39 |
+
"E-COMPANY_NAME",
|
| 40 |
+
"S-COMPANY_NAME",
|
| 41 |
+
"B-COORDINATE",
|
| 42 |
+
"I-COORDINATE",
|
| 43 |
+
"E-COORDINATE",
|
| 44 |
+
"S-COORDINATE",
|
| 45 |
+
"B-COUNTRY",
|
| 46 |
+
"I-COUNTRY",
|
| 47 |
+
"E-COUNTRY",
|
| 48 |
+
"S-COUNTRY",
|
| 49 |
+
"B-CREDIT_CARD_NUMBER",
|
| 50 |
+
"I-CREDIT_CARD_NUMBER",
|
| 51 |
+
"E-CREDIT_CARD_NUMBER",
|
| 52 |
+
"S-CREDIT_CARD_NUMBER",
|
| 53 |
+
"B-CUSTOMER_ID",
|
| 54 |
+
"I-CUSTOMER_ID",
|
| 55 |
+
"E-CUSTOMER_ID",
|
| 56 |
+
"S-CUSTOMER_ID",
|
| 57 |
+
"B-CVV",
|
| 58 |
+
"I-CVV",
|
| 59 |
+
"E-CVV",
|
| 60 |
+
"S-CVV",
|
| 61 |
+
"B-DATE",
|
| 62 |
+
"I-DATE",
|
| 63 |
+
"E-DATE",
|
| 64 |
+
"S-DATE",
|
| 65 |
+
"B-DATE_OF_BIRTH",
|
| 66 |
+
"I-DATE_OF_BIRTH",
|
| 67 |
+
"E-DATE_OF_BIRTH",
|
| 68 |
+
"S-DATE_OF_BIRTH",
|
| 69 |
+
"B-DATE_TIME",
|
| 70 |
+
"I-DATE_TIME",
|
| 71 |
+
"E-DATE_TIME",
|
| 72 |
+
"S-DATE_TIME",
|
| 73 |
+
"B-DEVICE_IDENTIFIER",
|
| 74 |
+
"I-DEVICE_IDENTIFIER",
|
| 75 |
+
"E-DEVICE_IDENTIFIER",
|
| 76 |
+
"S-DEVICE_IDENTIFIER",
|
| 77 |
+
"B-EMAIL",
|
| 78 |
+
"I-EMAIL",
|
| 79 |
+
"E-EMAIL",
|
| 80 |
+
"S-EMAIL",
|
| 81 |
+
"B-EMPLOYEE_ID",
|
| 82 |
+
"I-EMPLOYEE_ID",
|
| 83 |
+
"E-EMPLOYEE_ID",
|
| 84 |
+
"S-EMPLOYEE_ID",
|
| 85 |
+
"B-FIRST_NAME",
|
| 86 |
+
"I-FIRST_NAME",
|
| 87 |
+
"E-FIRST_NAME",
|
| 88 |
+
"S-FIRST_NAME",
|
| 89 |
+
"B-HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 90 |
+
"I-HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 91 |
+
"E-HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 92 |
+
"S-HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 93 |
+
"B-IPV4",
|
| 94 |
+
"I-IPV4",
|
| 95 |
+
"E-IPV4",
|
| 96 |
+
"S-IPV4",
|
| 97 |
+
"B-IPV6",
|
| 98 |
+
"I-IPV6",
|
| 99 |
+
"E-IPV6",
|
| 100 |
+
"S-IPV6",
|
| 101 |
+
"B-LAST_NAME",
|
| 102 |
+
"I-LAST_NAME",
|
| 103 |
+
"E-LAST_NAME",
|
| 104 |
+
"S-LAST_NAME",
|
| 105 |
+
"B-LICENSE_PLATE",
|
| 106 |
+
"I-LICENSE_PLATE",
|
| 107 |
+
"E-LICENSE_PLATE",
|
| 108 |
+
"S-LICENSE_PLATE",
|
| 109 |
+
"B-MEDICAL_RECORD_NUMBER",
|
| 110 |
+
"I-MEDICAL_RECORD_NUMBER",
|
| 111 |
+
"E-MEDICAL_RECORD_NUMBER",
|
| 112 |
+
"S-MEDICAL_RECORD_NUMBER",
|
| 113 |
+
"B-NAME",
|
| 114 |
+
"I-NAME",
|
| 115 |
+
"E-NAME",
|
| 116 |
+
"S-NAME",
|
| 117 |
+
"B-NATIONAL_ID",
|
| 118 |
+
"I-NATIONAL_ID",
|
| 119 |
+
"E-NATIONAL_ID",
|
| 120 |
+
"S-NATIONAL_ID",
|
| 121 |
+
"B-PASSWORD",
|
| 122 |
+
"I-PASSWORD",
|
| 123 |
+
"E-PASSWORD",
|
| 124 |
+
"S-PASSWORD",
|
| 125 |
+
"B-PHONE_NUMBER",
|
| 126 |
+
"I-PHONE_NUMBER",
|
| 127 |
+
"E-PHONE_NUMBER",
|
| 128 |
+
"S-PHONE_NUMBER",
|
| 129 |
+
"B-PIN",
|
| 130 |
+
"I-PIN",
|
| 131 |
+
"E-PIN",
|
| 132 |
+
"S-PIN",
|
| 133 |
+
"B-POSTCODE",
|
| 134 |
+
"I-POSTCODE",
|
| 135 |
+
"E-POSTCODE",
|
| 136 |
+
"S-POSTCODE",
|
| 137 |
+
"B-SSN",
|
| 138 |
+
"I-SSN",
|
| 139 |
+
"E-SSN",
|
| 140 |
+
"S-SSN",
|
| 141 |
+
"B-STATE",
|
| 142 |
+
"I-STATE",
|
| 143 |
+
"E-STATE",
|
| 144 |
+
"S-STATE",
|
| 145 |
+
"B-STREET_ADDRESS",
|
| 146 |
+
"I-STREET_ADDRESS",
|
| 147 |
+
"E-STREET_ADDRESS",
|
| 148 |
+
"S-STREET_ADDRESS",
|
| 149 |
+
"B-SWIFT_BIC",
|
| 150 |
+
"I-SWIFT_BIC",
|
| 151 |
+
"E-SWIFT_BIC",
|
| 152 |
+
"S-SWIFT_BIC",
|
| 153 |
+
"B-TAX_ID",
|
| 154 |
+
"I-TAX_ID",
|
| 155 |
+
"E-TAX_ID",
|
| 156 |
+
"S-TAX_ID",
|
| 157 |
+
"B-TIME",
|
| 158 |
+
"I-TIME",
|
| 159 |
+
"E-TIME",
|
| 160 |
+
"S-TIME",
|
| 161 |
+
"B-UNIQUE_IDENTIFIER",
|
| 162 |
+
"I-UNIQUE_IDENTIFIER",
|
| 163 |
+
"E-UNIQUE_IDENTIFIER",
|
| 164 |
+
"S-UNIQUE_IDENTIFIER",
|
| 165 |
+
"B-URL",
|
| 166 |
+
"I-URL",
|
| 167 |
+
"E-URL",
|
| 168 |
+
"S-URL",
|
| 169 |
+
"B-USER_NAME",
|
| 170 |
+
"I-USER_NAME",
|
| 171 |
+
"E-USER_NAME",
|
| 172 |
+
"S-USER_NAME",
|
| 173 |
+
"B-VEHICLE_IDENTIFIER",
|
| 174 |
+
"I-VEHICLE_IDENTIFIER",
|
| 175 |
+
"E-VEHICLE_IDENTIFIER",
|
| 176 |
+
"S-VEHICLE_IDENTIFIER"
|
| 177 |
+
],
|
| 178 |
+
"max_seq_length": 512,
|
| 179 |
+
"timestamp": "2026-04-08T04:03:00.111772"
|
| 180 |
+
}
|
gretel-pii-ready/bert-bioes-bert-base-multilingual-cased/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|