Upload folder using huggingface_hub
Browse files- gretel-pii-ready/TokenBased-CRF/checkpoint-63/config.json +204 -0
- gretel-pii-ready/TokenBased-CRF/checkpoint-63/model.safetensors +3 -0
- gretel-pii-ready/TokenBased-CRF/checkpoint-63/optimizer.pt +3 -0
- gretel-pii-ready/TokenBased-CRF/checkpoint-63/rng_state.pth +3 -0
- gretel-pii-ready/TokenBased-CRF/checkpoint-63/scheduler.pt +3 -0
- gretel-pii-ready/TokenBased-CRF/checkpoint-63/tokenizer.json +0 -0
- gretel-pii-ready/TokenBased-CRF/checkpoint-63/tokenizer_config.json +15 -0
- gretel-pii-ready/TokenBased-CRF/checkpoint-63/trainer_state.json +33 -0
- gretel-pii-ready/TokenBased-CRF/checkpoint-63/training_args.bin +3 -0
- gretel-pii-ready/TokenBased-CRF/config.json +204 -0
- gretel-pii-ready/TokenBased-CRF/pytorch_model.bin +3 -0
- gretel-pii-ready/TokenBased-CRF/tokenizer.json +0 -0
- gretel-pii-ready/TokenBased-CRF/tokenizer_config.json +15 -0
gretel-pii-ready/TokenBased-CRF/checkpoint-63/config.json
ADDED
|
@@ -0,0 +1,204 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_cross_attention": false,
|
| 3 |
+
"architectures": [
|
| 4 |
+
"TransformerCrfForTokenClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": null,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"directionality": "bidi",
|
| 10 |
+
"dtype": "float32",
|
| 11 |
+
"eos_token_id": null,
|
| 12 |
+
"hidden_act": "gelu",
|
| 13 |
+
"hidden_dropout_prob": 0.1,
|
| 14 |
+
"hidden_size": 768,
|
| 15 |
+
"id2label": {
|
| 16 |
+
"0": "O",
|
| 17 |
+
"1": "B-ACCOUNT_NUMBER",
|
| 18 |
+
"2": "I-ACCOUNT_NUMBER",
|
| 19 |
+
"3": "B-ADDRESS",
|
| 20 |
+
"4": "I-ADDRESS",
|
| 21 |
+
"5": "B-API_KEY",
|
| 22 |
+
"6": "I-API_KEY",
|
| 23 |
+
"7": "B-BANK_ROUTING_NUMBER",
|
| 24 |
+
"8": "I-BANK_ROUTING_NUMBER",
|
| 25 |
+
"9": "B-BIOMETRIC_IDENTIFIER",
|
| 26 |
+
"10": "I-BIOMETRIC_IDENTIFIER",
|
| 27 |
+
"11": "B-CERTIFICATE_LICENSE_NUMBER",
|
| 28 |
+
"12": "I-CERTIFICATE_LICENSE_NUMBER",
|
| 29 |
+
"13": "B-CITY",
|
| 30 |
+
"14": "I-CITY",
|
| 31 |
+
"15": "B-COMPANY_NAME",
|
| 32 |
+
"16": "I-COMPANY_NAME",
|
| 33 |
+
"17": "B-COORDINATE",
|
| 34 |
+
"18": "I-COORDINATE",
|
| 35 |
+
"19": "B-COUNTRY",
|
| 36 |
+
"20": "I-COUNTRY",
|
| 37 |
+
"21": "B-CREDIT_CARD_NUMBER",
|
| 38 |
+
"22": "I-CREDIT_CARD_NUMBER",
|
| 39 |
+
"23": "B-CUSTOMER_ID",
|
| 40 |
+
"24": "I-CUSTOMER_ID",
|
| 41 |
+
"25": "B-CVV",
|
| 42 |
+
"26": "I-CVV",
|
| 43 |
+
"27": "B-DATE",
|
| 44 |
+
"28": "I-DATE",
|
| 45 |
+
"29": "B-DATE_OF_BIRTH",
|
| 46 |
+
"30": "I-DATE_OF_BIRTH",
|
| 47 |
+
"31": "B-DATE_TIME",
|
| 48 |
+
"32": "I-DATE_TIME",
|
| 49 |
+
"33": "B-DEVICE_IDENTIFIER",
|
| 50 |
+
"34": "I-DEVICE_IDENTIFIER",
|
| 51 |
+
"35": "B-EMAIL",
|
| 52 |
+
"36": "I-EMAIL",
|
| 53 |
+
"37": "B-EMPLOYEE_ID",
|
| 54 |
+
"38": "I-EMPLOYEE_ID",
|
| 55 |
+
"39": "B-FIRST_NAME",
|
| 56 |
+
"40": "I-FIRST_NAME",
|
| 57 |
+
"41": "B-HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 58 |
+
"42": "I-HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 59 |
+
"43": "B-IPV4",
|
| 60 |
+
"44": "I-IPV4",
|
| 61 |
+
"45": "B-IPV6",
|
| 62 |
+
"46": "I-IPV6",
|
| 63 |
+
"47": "B-LAST_NAME",
|
| 64 |
+
"48": "I-LAST_NAME",
|
| 65 |
+
"49": "B-LICENSE_PLATE",
|
| 66 |
+
"50": "I-LICENSE_PLATE",
|
| 67 |
+
"51": "B-MEDICAL_RECORD_NUMBER",
|
| 68 |
+
"52": "I-MEDICAL_RECORD_NUMBER",
|
| 69 |
+
"53": "B-NAME",
|
| 70 |
+
"54": "I-NAME",
|
| 71 |
+
"55": "B-NATIONAL_ID",
|
| 72 |
+
"56": "I-NATIONAL_ID",
|
| 73 |
+
"57": "B-PASSWORD",
|
| 74 |
+
"58": "I-PASSWORD",
|
| 75 |
+
"59": "B-PHONE_NUMBER",
|
| 76 |
+
"60": "I-PHONE_NUMBER",
|
| 77 |
+
"61": "B-POSTCODE",
|
| 78 |
+
"62": "I-POSTCODE",
|
| 79 |
+
"63": "B-SSN",
|
| 80 |
+
"64": "I-SSN",
|
| 81 |
+
"65": "B-STATE",
|
| 82 |
+
"66": "I-STATE",
|
| 83 |
+
"67": "B-STREET_ADDRESS",
|
| 84 |
+
"68": "I-STREET_ADDRESS",
|
| 85 |
+
"69": "B-SWIFT_BIC",
|
| 86 |
+
"70": "I-SWIFT_BIC",
|
| 87 |
+
"71": "B-TAX_ID",
|
| 88 |
+
"72": "I-TAX_ID",
|
| 89 |
+
"73": "B-TIME",
|
| 90 |
+
"74": "I-TIME",
|
| 91 |
+
"75": "B-UNIQUE_IDENTIFIER",
|
| 92 |
+
"76": "I-UNIQUE_IDENTIFIER",
|
| 93 |
+
"77": "B-URL",
|
| 94 |
+
"78": "I-URL",
|
| 95 |
+
"79": "B-USER_NAME",
|
| 96 |
+
"80": "I-USER_NAME",
|
| 97 |
+
"81": "B-VEHICLE_IDENTIFIER",
|
| 98 |
+
"82": "I-VEHICLE_IDENTIFIER"
|
| 99 |
+
},
|
| 100 |
+
"initializer_range": 0.02,
|
| 101 |
+
"intermediate_size": 3072,
|
| 102 |
+
"is_decoder": false,
|
| 103 |
+
"label2id": {
|
| 104 |
+
"B-ACCOUNT_NUMBER": 1,
|
| 105 |
+
"B-ADDRESS": 3,
|
| 106 |
+
"B-API_KEY": 5,
|
| 107 |
+
"B-BANK_ROUTING_NUMBER": 7,
|
| 108 |
+
"B-BIOMETRIC_IDENTIFIER": 9,
|
| 109 |
+
"B-CERTIFICATE_LICENSE_NUMBER": 11,
|
| 110 |
+
"B-CITY": 13,
|
| 111 |
+
"B-COMPANY_NAME": 15,
|
| 112 |
+
"B-COORDINATE": 17,
|
| 113 |
+
"B-COUNTRY": 19,
|
| 114 |
+
"B-CREDIT_CARD_NUMBER": 21,
|
| 115 |
+
"B-CUSTOMER_ID": 23,
|
| 116 |
+
"B-CVV": 25,
|
| 117 |
+
"B-DATE": 27,
|
| 118 |
+
"B-DATE_OF_BIRTH": 29,
|
| 119 |
+
"B-DATE_TIME": 31,
|
| 120 |
+
"B-DEVICE_IDENTIFIER": 33,
|
| 121 |
+
"B-EMAIL": 35,
|
| 122 |
+
"B-EMPLOYEE_ID": 37,
|
| 123 |
+
"B-FIRST_NAME": 39,
|
| 124 |
+
"B-HEALTH_PLAN_BENEFICIARY_NUMBER": 41,
|
| 125 |
+
"B-IPV4": 43,
|
| 126 |
+
"B-IPV6": 45,
|
| 127 |
+
"B-LAST_NAME": 47,
|
| 128 |
+
"B-LICENSE_PLATE": 49,
|
| 129 |
+
"B-MEDICAL_RECORD_NUMBER": 51,
|
| 130 |
+
"B-NAME": 53,
|
| 131 |
+
"B-NATIONAL_ID": 55,
|
| 132 |
+
"B-PASSWORD": 57,
|
| 133 |
+
"B-PHONE_NUMBER": 59,
|
| 134 |
+
"B-POSTCODE": 61,
|
| 135 |
+
"B-SSN": 63,
|
| 136 |
+
"B-STATE": 65,
|
| 137 |
+
"B-STREET_ADDRESS": 67,
|
| 138 |
+
"B-SWIFT_BIC": 69,
|
| 139 |
+
"B-TAX_ID": 71,
|
| 140 |
+
"B-TIME": 73,
|
| 141 |
+
"B-UNIQUE_IDENTIFIER": 75,
|
| 142 |
+
"B-URL": 77,
|
| 143 |
+
"B-USER_NAME": 79,
|
| 144 |
+
"B-VEHICLE_IDENTIFIER": 81,
|
| 145 |
+
"I-ACCOUNT_NUMBER": 2,
|
| 146 |
+
"I-ADDRESS": 4,
|
| 147 |
+
"I-API_KEY": 6,
|
| 148 |
+
"I-BANK_ROUTING_NUMBER": 8,
|
| 149 |
+
"I-BIOMETRIC_IDENTIFIER": 10,
|
| 150 |
+
"I-CERTIFICATE_LICENSE_NUMBER": 12,
|
| 151 |
+
"I-CITY": 14,
|
| 152 |
+
"I-COMPANY_NAME": 16,
|
| 153 |
+
"I-COORDINATE": 18,
|
| 154 |
+
"I-COUNTRY": 20,
|
| 155 |
+
"I-CREDIT_CARD_NUMBER": 22,
|
| 156 |
+
"I-CUSTOMER_ID": 24,
|
| 157 |
+
"I-CVV": 26,
|
| 158 |
+
"I-DATE": 28,
|
| 159 |
+
"I-DATE_OF_BIRTH": 30,
|
| 160 |
+
"I-DATE_TIME": 32,
|
| 161 |
+
"I-DEVICE_IDENTIFIER": 34,
|
| 162 |
+
"I-EMAIL": 36,
|
| 163 |
+
"I-EMPLOYEE_ID": 38,
|
| 164 |
+
"I-FIRST_NAME": 40,
|
| 165 |
+
"I-HEALTH_PLAN_BENEFICIARY_NUMBER": 42,
|
| 166 |
+
"I-IPV4": 44,
|
| 167 |
+
"I-IPV6": 46,
|
| 168 |
+
"I-LAST_NAME": 48,
|
| 169 |
+
"I-LICENSE_PLATE": 50,
|
| 170 |
+
"I-MEDICAL_RECORD_NUMBER": 52,
|
| 171 |
+
"I-NAME": 54,
|
| 172 |
+
"I-NATIONAL_ID": 56,
|
| 173 |
+
"I-PASSWORD": 58,
|
| 174 |
+
"I-PHONE_NUMBER": 60,
|
| 175 |
+
"I-POSTCODE": 62,
|
| 176 |
+
"I-SSN": 64,
|
| 177 |
+
"I-STATE": 66,
|
| 178 |
+
"I-STREET_ADDRESS": 68,
|
| 179 |
+
"I-SWIFT_BIC": 70,
|
| 180 |
+
"I-TAX_ID": 72,
|
| 181 |
+
"I-TIME": 74,
|
| 182 |
+
"I-UNIQUE_IDENTIFIER": 76,
|
| 183 |
+
"I-URL": 78,
|
| 184 |
+
"I-USER_NAME": 80,
|
| 185 |
+
"I-VEHICLE_IDENTIFIER": 82,
|
| 186 |
+
"O": 0
|
| 187 |
+
},
|
| 188 |
+
"layer_norm_eps": 1e-12,
|
| 189 |
+
"max_position_embeddings": 512,
|
| 190 |
+
"model_type": "bert",
|
| 191 |
+
"num_attention_heads": 12,
|
| 192 |
+
"num_hidden_layers": 12,
|
| 193 |
+
"pad_token_id": 0,
|
| 194 |
+
"pooler_fc_size": 768,
|
| 195 |
+
"pooler_num_attention_heads": 12,
|
| 196 |
+
"pooler_num_fc_layers": 3,
|
| 197 |
+
"pooler_size_per_head": 128,
|
| 198 |
+
"pooler_type": "first_token_transform",
|
| 199 |
+
"tie_word_embeddings": true,
|
| 200 |
+
"transformers_version": "5.3.0",
|
| 201 |
+
"type_vocab_size": 2,
|
| 202 |
+
"use_cache": false,
|
| 203 |
+
"vocab_size": 119547
|
| 204 |
+
}
|
gretel-pii-ready/TokenBased-CRF/checkpoint-63/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:439798edc50763ad5fa5410fa6de9d93ec8113fae5b5c2af87888e17ec9d762a
|
| 3 |
+
size 711722464
|
gretel-pii-ready/TokenBased-CRF/checkpoint-63/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ddab49756ad4f6d019794d9313059edf9260a9dd99a2353fa29783859753da60
|
| 3 |
+
size 1418841931
|
gretel-pii-ready/TokenBased-CRF/checkpoint-63/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:77e1bc0cc29528179cf7713b093fffbd0754c58c2875616aa74e1b6cd67bfe7d
|
| 3 |
+
size 14645
|
gretel-pii-ready/TokenBased-CRF/checkpoint-63/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8baf2b5a46fac85f95c0a15caf2414a92c036d41b6ccf016dc0542ed26e499c2
|
| 3 |
+
size 1465
|
gretel-pii-ready/TokenBased-CRF/checkpoint-63/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gretel-pii-ready/TokenBased-CRF/checkpoint-63/tokenizer_config.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": true,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"cls_token": "[CLS]",
|
| 5 |
+
"do_lower_case": false,
|
| 6 |
+
"is_local": false,
|
| 7 |
+
"mask_token": "[MASK]",
|
| 8 |
+
"model_max_length": 512,
|
| 9 |
+
"pad_token": "[PAD]",
|
| 10 |
+
"sep_token": "[SEP]",
|
| 11 |
+
"strip_accents": null,
|
| 12 |
+
"tokenize_chinese_chars": true,
|
| 13 |
+
"tokenizer_class": "BertTokenizer",
|
| 14 |
+
"unk_token": "[UNK]"
|
| 15 |
+
}
|
gretel-pii-ready/TokenBased-CRF/checkpoint-63/trainer_state.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 63,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [],
|
| 12 |
+
"logging_steps": 500,
|
| 13 |
+
"max_steps": 63,
|
| 14 |
+
"num_input_tokens_seen": 0,
|
| 15 |
+
"num_train_epochs": 1,
|
| 16 |
+
"save_steps": 500,
|
| 17 |
+
"stateful_callbacks": {
|
| 18 |
+
"TrainerControl": {
|
| 19 |
+
"args": {
|
| 20 |
+
"should_epoch_stop": false,
|
| 21 |
+
"should_evaluate": false,
|
| 22 |
+
"should_log": false,
|
| 23 |
+
"should_save": true,
|
| 24 |
+
"should_training_stop": true
|
| 25 |
+
},
|
| 26 |
+
"attributes": {}
|
| 27 |
+
}
|
| 28 |
+
},
|
| 29 |
+
"total_flos": 131662040064000.0,
|
| 30 |
+
"train_batch_size": 8,
|
| 31 |
+
"trial_name": null,
|
| 32 |
+
"trial_params": null
|
| 33 |
+
}
|
gretel-pii-ready/TokenBased-CRF/checkpoint-63/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d7adfd43ab830d4bbdcf45c20dc46d1b41e035d973db9875a684066dc6dd0ec9
|
| 3 |
+
size 5201
|
gretel-pii-ready/TokenBased-CRF/config.json
ADDED
|
@@ -0,0 +1,204 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_cross_attention": false,
|
| 3 |
+
"architectures": [
|
| 4 |
+
"TransformerCrfForTokenClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": null,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"directionality": "bidi",
|
| 10 |
+
"dtype": "float32",
|
| 11 |
+
"eos_token_id": null,
|
| 12 |
+
"hidden_act": "gelu",
|
| 13 |
+
"hidden_dropout_prob": 0.1,
|
| 14 |
+
"hidden_size": 768,
|
| 15 |
+
"id2label": {
|
| 16 |
+
"0": "O",
|
| 17 |
+
"1": "B-ACCOUNT_NUMBER",
|
| 18 |
+
"2": "I-ACCOUNT_NUMBER",
|
| 19 |
+
"3": "B-ADDRESS",
|
| 20 |
+
"4": "I-ADDRESS",
|
| 21 |
+
"5": "B-API_KEY",
|
| 22 |
+
"6": "I-API_KEY",
|
| 23 |
+
"7": "B-BANK_ROUTING_NUMBER",
|
| 24 |
+
"8": "I-BANK_ROUTING_NUMBER",
|
| 25 |
+
"9": "B-BIOMETRIC_IDENTIFIER",
|
| 26 |
+
"10": "I-BIOMETRIC_IDENTIFIER",
|
| 27 |
+
"11": "B-CERTIFICATE_LICENSE_NUMBER",
|
| 28 |
+
"12": "I-CERTIFICATE_LICENSE_NUMBER",
|
| 29 |
+
"13": "B-CITY",
|
| 30 |
+
"14": "I-CITY",
|
| 31 |
+
"15": "B-COMPANY_NAME",
|
| 32 |
+
"16": "I-COMPANY_NAME",
|
| 33 |
+
"17": "B-COORDINATE",
|
| 34 |
+
"18": "I-COORDINATE",
|
| 35 |
+
"19": "B-COUNTRY",
|
| 36 |
+
"20": "I-COUNTRY",
|
| 37 |
+
"21": "B-CREDIT_CARD_NUMBER",
|
| 38 |
+
"22": "I-CREDIT_CARD_NUMBER",
|
| 39 |
+
"23": "B-CUSTOMER_ID",
|
| 40 |
+
"24": "I-CUSTOMER_ID",
|
| 41 |
+
"25": "B-CVV",
|
| 42 |
+
"26": "I-CVV",
|
| 43 |
+
"27": "B-DATE",
|
| 44 |
+
"28": "I-DATE",
|
| 45 |
+
"29": "B-DATE_OF_BIRTH",
|
| 46 |
+
"30": "I-DATE_OF_BIRTH",
|
| 47 |
+
"31": "B-DATE_TIME",
|
| 48 |
+
"32": "I-DATE_TIME",
|
| 49 |
+
"33": "B-DEVICE_IDENTIFIER",
|
| 50 |
+
"34": "I-DEVICE_IDENTIFIER",
|
| 51 |
+
"35": "B-EMAIL",
|
| 52 |
+
"36": "I-EMAIL",
|
| 53 |
+
"37": "B-EMPLOYEE_ID",
|
| 54 |
+
"38": "I-EMPLOYEE_ID",
|
| 55 |
+
"39": "B-FIRST_NAME",
|
| 56 |
+
"40": "I-FIRST_NAME",
|
| 57 |
+
"41": "B-HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 58 |
+
"42": "I-HEALTH_PLAN_BENEFICIARY_NUMBER",
|
| 59 |
+
"43": "B-IPV4",
|
| 60 |
+
"44": "I-IPV4",
|
| 61 |
+
"45": "B-IPV6",
|
| 62 |
+
"46": "I-IPV6",
|
| 63 |
+
"47": "B-LAST_NAME",
|
| 64 |
+
"48": "I-LAST_NAME",
|
| 65 |
+
"49": "B-LICENSE_PLATE",
|
| 66 |
+
"50": "I-LICENSE_PLATE",
|
| 67 |
+
"51": "B-MEDICAL_RECORD_NUMBER",
|
| 68 |
+
"52": "I-MEDICAL_RECORD_NUMBER",
|
| 69 |
+
"53": "B-NAME",
|
| 70 |
+
"54": "I-NAME",
|
| 71 |
+
"55": "B-NATIONAL_ID",
|
| 72 |
+
"56": "I-NATIONAL_ID",
|
| 73 |
+
"57": "B-PASSWORD",
|
| 74 |
+
"58": "I-PASSWORD",
|
| 75 |
+
"59": "B-PHONE_NUMBER",
|
| 76 |
+
"60": "I-PHONE_NUMBER",
|
| 77 |
+
"61": "B-POSTCODE",
|
| 78 |
+
"62": "I-POSTCODE",
|
| 79 |
+
"63": "B-SSN",
|
| 80 |
+
"64": "I-SSN",
|
| 81 |
+
"65": "B-STATE",
|
| 82 |
+
"66": "I-STATE",
|
| 83 |
+
"67": "B-STREET_ADDRESS",
|
| 84 |
+
"68": "I-STREET_ADDRESS",
|
| 85 |
+
"69": "B-SWIFT_BIC",
|
| 86 |
+
"70": "I-SWIFT_BIC",
|
| 87 |
+
"71": "B-TAX_ID",
|
| 88 |
+
"72": "I-TAX_ID",
|
| 89 |
+
"73": "B-TIME",
|
| 90 |
+
"74": "I-TIME",
|
| 91 |
+
"75": "B-UNIQUE_IDENTIFIER",
|
| 92 |
+
"76": "I-UNIQUE_IDENTIFIER",
|
| 93 |
+
"77": "B-URL",
|
| 94 |
+
"78": "I-URL",
|
| 95 |
+
"79": "B-USER_NAME",
|
| 96 |
+
"80": "I-USER_NAME",
|
| 97 |
+
"81": "B-VEHICLE_IDENTIFIER",
|
| 98 |
+
"82": "I-VEHICLE_IDENTIFIER"
|
| 99 |
+
},
|
| 100 |
+
"initializer_range": 0.02,
|
| 101 |
+
"intermediate_size": 3072,
|
| 102 |
+
"is_decoder": false,
|
| 103 |
+
"label2id": {
|
| 104 |
+
"B-ACCOUNT_NUMBER": 1,
|
| 105 |
+
"B-ADDRESS": 3,
|
| 106 |
+
"B-API_KEY": 5,
|
| 107 |
+
"B-BANK_ROUTING_NUMBER": 7,
|
| 108 |
+
"B-BIOMETRIC_IDENTIFIER": 9,
|
| 109 |
+
"B-CERTIFICATE_LICENSE_NUMBER": 11,
|
| 110 |
+
"B-CITY": 13,
|
| 111 |
+
"B-COMPANY_NAME": 15,
|
| 112 |
+
"B-COORDINATE": 17,
|
| 113 |
+
"B-COUNTRY": 19,
|
| 114 |
+
"B-CREDIT_CARD_NUMBER": 21,
|
| 115 |
+
"B-CUSTOMER_ID": 23,
|
| 116 |
+
"B-CVV": 25,
|
| 117 |
+
"B-DATE": 27,
|
| 118 |
+
"B-DATE_OF_BIRTH": 29,
|
| 119 |
+
"B-DATE_TIME": 31,
|
| 120 |
+
"B-DEVICE_IDENTIFIER": 33,
|
| 121 |
+
"B-EMAIL": 35,
|
| 122 |
+
"B-EMPLOYEE_ID": 37,
|
| 123 |
+
"B-FIRST_NAME": 39,
|
| 124 |
+
"B-HEALTH_PLAN_BENEFICIARY_NUMBER": 41,
|
| 125 |
+
"B-IPV4": 43,
|
| 126 |
+
"B-IPV6": 45,
|
| 127 |
+
"B-LAST_NAME": 47,
|
| 128 |
+
"B-LICENSE_PLATE": 49,
|
| 129 |
+
"B-MEDICAL_RECORD_NUMBER": 51,
|
| 130 |
+
"B-NAME": 53,
|
| 131 |
+
"B-NATIONAL_ID": 55,
|
| 132 |
+
"B-PASSWORD": 57,
|
| 133 |
+
"B-PHONE_NUMBER": 59,
|
| 134 |
+
"B-POSTCODE": 61,
|
| 135 |
+
"B-SSN": 63,
|
| 136 |
+
"B-STATE": 65,
|
| 137 |
+
"B-STREET_ADDRESS": 67,
|
| 138 |
+
"B-SWIFT_BIC": 69,
|
| 139 |
+
"B-TAX_ID": 71,
|
| 140 |
+
"B-TIME": 73,
|
| 141 |
+
"B-UNIQUE_IDENTIFIER": 75,
|
| 142 |
+
"B-URL": 77,
|
| 143 |
+
"B-USER_NAME": 79,
|
| 144 |
+
"B-VEHICLE_IDENTIFIER": 81,
|
| 145 |
+
"I-ACCOUNT_NUMBER": 2,
|
| 146 |
+
"I-ADDRESS": 4,
|
| 147 |
+
"I-API_KEY": 6,
|
| 148 |
+
"I-BANK_ROUTING_NUMBER": 8,
|
| 149 |
+
"I-BIOMETRIC_IDENTIFIER": 10,
|
| 150 |
+
"I-CERTIFICATE_LICENSE_NUMBER": 12,
|
| 151 |
+
"I-CITY": 14,
|
| 152 |
+
"I-COMPANY_NAME": 16,
|
| 153 |
+
"I-COORDINATE": 18,
|
| 154 |
+
"I-COUNTRY": 20,
|
| 155 |
+
"I-CREDIT_CARD_NUMBER": 22,
|
| 156 |
+
"I-CUSTOMER_ID": 24,
|
| 157 |
+
"I-CVV": 26,
|
| 158 |
+
"I-DATE": 28,
|
| 159 |
+
"I-DATE_OF_BIRTH": 30,
|
| 160 |
+
"I-DATE_TIME": 32,
|
| 161 |
+
"I-DEVICE_IDENTIFIER": 34,
|
| 162 |
+
"I-EMAIL": 36,
|
| 163 |
+
"I-EMPLOYEE_ID": 38,
|
| 164 |
+
"I-FIRST_NAME": 40,
|
| 165 |
+
"I-HEALTH_PLAN_BENEFICIARY_NUMBER": 42,
|
| 166 |
+
"I-IPV4": 44,
|
| 167 |
+
"I-IPV6": 46,
|
| 168 |
+
"I-LAST_NAME": 48,
|
| 169 |
+
"I-LICENSE_PLATE": 50,
|
| 170 |
+
"I-MEDICAL_RECORD_NUMBER": 52,
|
| 171 |
+
"I-NAME": 54,
|
| 172 |
+
"I-NATIONAL_ID": 56,
|
| 173 |
+
"I-PASSWORD": 58,
|
| 174 |
+
"I-PHONE_NUMBER": 60,
|
| 175 |
+
"I-POSTCODE": 62,
|
| 176 |
+
"I-SSN": 64,
|
| 177 |
+
"I-STATE": 66,
|
| 178 |
+
"I-STREET_ADDRESS": 68,
|
| 179 |
+
"I-SWIFT_BIC": 70,
|
| 180 |
+
"I-TAX_ID": 72,
|
| 181 |
+
"I-TIME": 74,
|
| 182 |
+
"I-UNIQUE_IDENTIFIER": 76,
|
| 183 |
+
"I-URL": 78,
|
| 184 |
+
"I-USER_NAME": 80,
|
| 185 |
+
"I-VEHICLE_IDENTIFIER": 82,
|
| 186 |
+
"O": 0
|
| 187 |
+
},
|
| 188 |
+
"layer_norm_eps": 1e-12,
|
| 189 |
+
"max_position_embeddings": 512,
|
| 190 |
+
"model_type": "bert",
|
| 191 |
+
"num_attention_heads": 12,
|
| 192 |
+
"num_hidden_layers": 12,
|
| 193 |
+
"pad_token_id": 0,
|
| 194 |
+
"pooler_fc_size": 768,
|
| 195 |
+
"pooler_num_attention_heads": 12,
|
| 196 |
+
"pooler_num_fc_layers": 3,
|
| 197 |
+
"pooler_size_per_head": 128,
|
| 198 |
+
"pooler_type": "first_token_transform",
|
| 199 |
+
"tie_word_embeddings": true,
|
| 200 |
+
"transformers_version": "5.3.0",
|
| 201 |
+
"type_vocab_size": 2,
|
| 202 |
+
"use_cache": false,
|
| 203 |
+
"vocab_size": 119547
|
| 204 |
+
}
|
gretel-pii-ready/TokenBased-CRF/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ebf6fa05e9ef79aacd8d79dd484d2d564ec9be20273c94f7c8cc167ea383afdf
|
| 3 |
+
size 711784467
|
gretel-pii-ready/TokenBased-CRF/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gretel-pii-ready/TokenBased-CRF/tokenizer_config.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": true,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"cls_token": "[CLS]",
|
| 5 |
+
"do_lower_case": false,
|
| 6 |
+
"is_local": false,
|
| 7 |
+
"mask_token": "[MASK]",
|
| 8 |
+
"model_max_length": 512,
|
| 9 |
+
"pad_token": "[PAD]",
|
| 10 |
+
"sep_token": "[SEP]",
|
| 11 |
+
"strip_accents": null,
|
| 12 |
+
"tokenize_chinese_chars": true,
|
| 13 |
+
"tokenizer_class": "BertTokenizer",
|
| 14 |
+
"unk_token": "[UNK]"
|
| 15 |
+
}
|