PuxAI commited on
Commit
bf22252
·
verified ·
1 Parent(s): 5f91590

Upload folder using huggingface_hub

Browse files
nemotron-pii-ready/TokenBased-CRF/checkpoint-63/config.json ADDED
@@ -0,0 +1,256 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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-AGE",
20
+ "4": "I-AGE",
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-BLOOD_TYPE",
28
+ "12": "I-BLOOD_TYPE",
29
+ "13": "B-CERTIFICATE_LICENSE_NUMBER",
30
+ "14": "I-CERTIFICATE_LICENSE_NUMBER",
31
+ "15": "B-CITY",
32
+ "16": "I-CITY",
33
+ "17": "B-COMPANY_NAME",
34
+ "18": "I-COMPANY_NAME",
35
+ "19": "B-COORDINATE",
36
+ "20": "I-COORDINATE",
37
+ "21": "B-COUNTRY",
38
+ "22": "I-COUNTRY",
39
+ "23": "B-COUNTY",
40
+ "24": "I-COUNTY",
41
+ "25": "B-CREDIT_DEBIT_CARD",
42
+ "26": "I-CREDIT_DEBIT_CARD",
43
+ "27": "B-CUSTOMER_ID",
44
+ "28": "I-CUSTOMER_ID",
45
+ "29": "B-CVV",
46
+ "30": "I-CVV",
47
+ "31": "B-DATE",
48
+ "32": "I-DATE",
49
+ "33": "B-DATE_OF_BIRTH",
50
+ "34": "I-DATE_OF_BIRTH",
51
+ "35": "B-DATE_TIME",
52
+ "36": "I-DATE_TIME",
53
+ "37": "B-DEVICE_IDENTIFIER",
54
+ "38": "I-DEVICE_IDENTIFIER",
55
+ "39": "B-EDUCATION_LEVEL",
56
+ "40": "I-EDUCATION_LEVEL",
57
+ "41": "B-EMAIL",
58
+ "42": "I-EMAIL",
59
+ "43": "B-EMPLOYEE_ID",
60
+ "44": "I-EMPLOYEE_ID",
61
+ "45": "B-EMPLOYMENT_STATUS",
62
+ "46": "I-EMPLOYMENT_STATUS",
63
+ "47": "B-FAX_NUMBER",
64
+ "48": "I-FAX_NUMBER",
65
+ "49": "B-FIRST_NAME",
66
+ "50": "I-FIRST_NAME",
67
+ "51": "B-GENDER",
68
+ "52": "I-GENDER",
69
+ "53": "B-HEALTH_PLAN_BENEFICIARY_NUMBER",
70
+ "54": "I-HEALTH_PLAN_BENEFICIARY_NUMBER",
71
+ "55": "B-HTTP_COOKIE",
72
+ "56": "I-HTTP_COOKIE",
73
+ "57": "B-IPV4",
74
+ "58": "I-IPV4",
75
+ "59": "B-IPV6",
76
+ "60": "I-IPV6",
77
+ "61": "B-LANGUAGE",
78
+ "62": "I-LANGUAGE",
79
+ "63": "B-LAST_NAME",
80
+ "64": "I-LAST_NAME",
81
+ "65": "B-LICENSE_PLATE",
82
+ "66": "I-LICENSE_PLATE",
83
+ "67": "B-MAC_ADDRESS",
84
+ "68": "I-MAC_ADDRESS",
85
+ "69": "B-MEDICAL_RECORD_NUMBER",
86
+ "70": "I-MEDICAL_RECORD_NUMBER",
87
+ "71": "B-OCCUPATION",
88
+ "72": "I-OCCUPATION",
89
+ "73": "B-PASSWORD",
90
+ "74": "I-PASSWORD",
91
+ "75": "B-PHONE_NUMBER",
92
+ "76": "I-PHONE_NUMBER",
93
+ "77": "B-PIN",
94
+ "78": "I-PIN",
95
+ "79": "B-POLITICAL_VIEW",
96
+ "80": "I-POLITICAL_VIEW",
97
+ "81": "B-POSTCODE",
98
+ "82": "I-POSTCODE",
99
+ "83": "B-RACE_ETHNICITY",
100
+ "84": "I-RACE_ETHNICITY",
101
+ "85": "B-RELIGIOUS_BELIEF",
102
+ "86": "I-RELIGIOUS_BELIEF",
103
+ "87": "B-SEXUALITY",
104
+ "88": "I-SEXUALITY",
105
+ "89": "B-SSN",
106
+ "90": "I-SSN",
107
+ "91": "B-STATE",
108
+ "92": "I-STATE",
109
+ "93": "B-STREET_ADDRESS",
110
+ "94": "I-STREET_ADDRESS",
111
+ "95": "B-SWIFT_BIC",
112
+ "96": "I-SWIFT_BIC",
113
+ "97": "B-TAX_ID",
114
+ "98": "I-TAX_ID",
115
+ "99": "B-TIME",
116
+ "100": "I-TIME",
117
+ "101": "B-UNIQUE_ID",
118
+ "102": "I-UNIQUE_ID",
119
+ "103": "B-URL",
120
+ "104": "I-URL",
121
+ "105": "B-USER_NAME",
122
+ "106": "I-USER_NAME",
123
+ "107": "B-VEHICLE_IDENTIFIER",
124
+ "108": "I-VEHICLE_IDENTIFIER"
125
+ },
126
+ "initializer_range": 0.02,
127
+ "intermediate_size": 3072,
128
+ "is_decoder": false,
129
+ "label2id": {
130
+ "B-ACCOUNT_NUMBER": 1,
131
+ "B-AGE": 3,
132
+ "B-API_KEY": 5,
133
+ "B-BANK_ROUTING_NUMBER": 7,
134
+ "B-BIOMETRIC_IDENTIFIER": 9,
135
+ "B-BLOOD_TYPE": 11,
136
+ "B-CERTIFICATE_LICENSE_NUMBER": 13,
137
+ "B-CITY": 15,
138
+ "B-COMPANY_NAME": 17,
139
+ "B-COORDINATE": 19,
140
+ "B-COUNTRY": 21,
141
+ "B-COUNTY": 23,
142
+ "B-CREDIT_DEBIT_CARD": 25,
143
+ "B-CUSTOMER_ID": 27,
144
+ "B-CVV": 29,
145
+ "B-DATE": 31,
146
+ "B-DATE_OF_BIRTH": 33,
147
+ "B-DATE_TIME": 35,
148
+ "B-DEVICE_IDENTIFIER": 37,
149
+ "B-EDUCATION_LEVEL": 39,
150
+ "B-EMAIL": 41,
151
+ "B-EMPLOYEE_ID": 43,
152
+ "B-EMPLOYMENT_STATUS": 45,
153
+ "B-FAX_NUMBER": 47,
154
+ "B-FIRST_NAME": 49,
155
+ "B-GENDER": 51,
156
+ "B-HEALTH_PLAN_BENEFICIARY_NUMBER": 53,
157
+ "B-HTTP_COOKIE": 55,
158
+ "B-IPV4": 57,
159
+ "B-IPV6": 59,
160
+ "B-LANGUAGE": 61,
161
+ "B-LAST_NAME": 63,
162
+ "B-LICENSE_PLATE": 65,
163
+ "B-MAC_ADDRESS": 67,
164
+ "B-MEDICAL_RECORD_NUMBER": 69,
165
+ "B-OCCUPATION": 71,
166
+ "B-PASSWORD": 73,
167
+ "B-PHONE_NUMBER": 75,
168
+ "B-PIN": 77,
169
+ "B-POLITICAL_VIEW": 79,
170
+ "B-POSTCODE": 81,
171
+ "B-RACE_ETHNICITY": 83,
172
+ "B-RELIGIOUS_BELIEF": 85,
173
+ "B-SEXUALITY": 87,
174
+ "B-SSN": 89,
175
+ "B-STATE": 91,
176
+ "B-STREET_ADDRESS": 93,
177
+ "B-SWIFT_BIC": 95,
178
+ "B-TAX_ID": 97,
179
+ "B-TIME": 99,
180
+ "B-UNIQUE_ID": 101,
181
+ "B-URL": 103,
182
+ "B-USER_NAME": 105,
183
+ "B-VEHICLE_IDENTIFIER": 107,
184
+ "I-ACCOUNT_NUMBER": 2,
185
+ "I-AGE": 4,
186
+ "I-API_KEY": 6,
187
+ "I-BANK_ROUTING_NUMBER": 8,
188
+ "I-BIOMETRIC_IDENTIFIER": 10,
189
+ "I-BLOOD_TYPE": 12,
190
+ "I-CERTIFICATE_LICENSE_NUMBER": 14,
191
+ "I-CITY": 16,
192
+ "I-COMPANY_NAME": 18,
193
+ "I-COORDINATE": 20,
194
+ "I-COUNTRY": 22,
195
+ "I-COUNTY": 24,
196
+ "I-CREDIT_DEBIT_CARD": 26,
197
+ "I-CUSTOMER_ID": 28,
198
+ "I-CVV": 30,
199
+ "I-DATE": 32,
200
+ "I-DATE_OF_BIRTH": 34,
201
+ "I-DATE_TIME": 36,
202
+ "I-DEVICE_IDENTIFIER": 38,
203
+ "I-EDUCATION_LEVEL": 40,
204
+ "I-EMAIL": 42,
205
+ "I-EMPLOYEE_ID": 44,
206
+ "I-EMPLOYMENT_STATUS": 46,
207
+ "I-FAX_NUMBER": 48,
208
+ "I-FIRST_NAME": 50,
209
+ "I-GENDER": 52,
210
+ "I-HEALTH_PLAN_BENEFICIARY_NUMBER": 54,
211
+ "I-HTTP_COOKIE": 56,
212
+ "I-IPV4": 58,
213
+ "I-IPV6": 60,
214
+ "I-LANGUAGE": 62,
215
+ "I-LAST_NAME": 64,
216
+ "I-LICENSE_PLATE": 66,
217
+ "I-MAC_ADDRESS": 68,
218
+ "I-MEDICAL_RECORD_NUMBER": 70,
219
+ "I-OCCUPATION": 72,
220
+ "I-PASSWORD": 74,
221
+ "I-PHONE_NUMBER": 76,
222
+ "I-PIN": 78,
223
+ "I-POLITICAL_VIEW": 80,
224
+ "I-POSTCODE": 82,
225
+ "I-RACE_ETHNICITY": 84,
226
+ "I-RELIGIOUS_BELIEF": 86,
227
+ "I-SEXUALITY": 88,
228
+ "I-SSN": 90,
229
+ "I-STATE": 92,
230
+ "I-STREET_ADDRESS": 94,
231
+ "I-SWIFT_BIC": 96,
232
+ "I-TAX_ID": 98,
233
+ "I-TIME": 100,
234
+ "I-UNIQUE_ID": 102,
235
+ "I-URL": 104,
236
+ "I-USER_NAME": 106,
237
+ "I-VEHICLE_IDENTIFIER": 108,
238
+ "O": 0
239
+ },
240
+ "layer_norm_eps": 1e-12,
241
+ "max_position_embeddings": 512,
242
+ "model_type": "bert",
243
+ "num_attention_heads": 12,
244
+ "num_hidden_layers": 12,
245
+ "pad_token_id": 0,
246
+ "pooler_fc_size": 768,
247
+ "pooler_num_attention_heads": 12,
248
+ "pooler_num_fc_layers": 3,
249
+ "pooler_size_per_head": 128,
250
+ "pooler_type": "first_token_transform",
251
+ "tie_word_embeddings": true,
252
+ "transformers_version": "5.3.0",
253
+ "type_vocab_size": 2,
254
+ "use_cache": false,
255
+ "vocab_size": 119547
256
+ }
nemotron-pii-ready/TokenBased-CRF/checkpoint-63/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:08d1ac8aa8d053477a6f88bfbc10972386877e64f69109bde11da057440e0961
3
+ size 711822624
nemotron-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:d51eb37cc352c9df7416b2fc4c276ace2c75c2dfae431da816de08f21e303939
3
+ size 1419041995
nemotron-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
nemotron-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
nemotron-pii-ready/TokenBased-CRF/checkpoint-63/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
nemotron-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
+ }
nemotron-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": 131700498432000.0,
30
+ "train_batch_size": 8,
31
+ "trial_name": null,
32
+ "trial_params": null
33
+ }
nemotron-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 sha256:0da5c7f9b3b0ea0cd38227f7eea880e1d4257af32102df957e7e8dabfbf6d7e5
3
+ size 5201
nemotron-pii-ready/TokenBased-CRF/config.json ADDED
@@ -0,0 +1,256 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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-AGE",
20
+ "4": "I-AGE",
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-BLOOD_TYPE",
28
+ "12": "I-BLOOD_TYPE",
29
+ "13": "B-CERTIFICATE_LICENSE_NUMBER",
30
+ "14": "I-CERTIFICATE_LICENSE_NUMBER",
31
+ "15": "B-CITY",
32
+ "16": "I-CITY",
33
+ "17": "B-COMPANY_NAME",
34
+ "18": "I-COMPANY_NAME",
35
+ "19": "B-COORDINATE",
36
+ "20": "I-COORDINATE",
37
+ "21": "B-COUNTRY",
38
+ "22": "I-COUNTRY",
39
+ "23": "B-COUNTY",
40
+ "24": "I-COUNTY",
41
+ "25": "B-CREDIT_DEBIT_CARD",
42
+ "26": "I-CREDIT_DEBIT_CARD",
43
+ "27": "B-CUSTOMER_ID",
44
+ "28": "I-CUSTOMER_ID",
45
+ "29": "B-CVV",
46
+ "30": "I-CVV",
47
+ "31": "B-DATE",
48
+ "32": "I-DATE",
49
+ "33": "B-DATE_OF_BIRTH",
50
+ "34": "I-DATE_OF_BIRTH",
51
+ "35": "B-DATE_TIME",
52
+ "36": "I-DATE_TIME",
53
+ "37": "B-DEVICE_IDENTIFIER",
54
+ "38": "I-DEVICE_IDENTIFIER",
55
+ "39": "B-EDUCATION_LEVEL",
56
+ "40": "I-EDUCATION_LEVEL",
57
+ "41": "B-EMAIL",
58
+ "42": "I-EMAIL",
59
+ "43": "B-EMPLOYEE_ID",
60
+ "44": "I-EMPLOYEE_ID",
61
+ "45": "B-EMPLOYMENT_STATUS",
62
+ "46": "I-EMPLOYMENT_STATUS",
63
+ "47": "B-FAX_NUMBER",
64
+ "48": "I-FAX_NUMBER",
65
+ "49": "B-FIRST_NAME",
66
+ "50": "I-FIRST_NAME",
67
+ "51": "B-GENDER",
68
+ "52": "I-GENDER",
69
+ "53": "B-HEALTH_PLAN_BENEFICIARY_NUMBER",
70
+ "54": "I-HEALTH_PLAN_BENEFICIARY_NUMBER",
71
+ "55": "B-HTTP_COOKIE",
72
+ "56": "I-HTTP_COOKIE",
73
+ "57": "B-IPV4",
74
+ "58": "I-IPV4",
75
+ "59": "B-IPV6",
76
+ "60": "I-IPV6",
77
+ "61": "B-LANGUAGE",
78
+ "62": "I-LANGUAGE",
79
+ "63": "B-LAST_NAME",
80
+ "64": "I-LAST_NAME",
81
+ "65": "B-LICENSE_PLATE",
82
+ "66": "I-LICENSE_PLATE",
83
+ "67": "B-MAC_ADDRESS",
84
+ "68": "I-MAC_ADDRESS",
85
+ "69": "B-MEDICAL_RECORD_NUMBER",
86
+ "70": "I-MEDICAL_RECORD_NUMBER",
87
+ "71": "B-OCCUPATION",
88
+ "72": "I-OCCUPATION",
89
+ "73": "B-PASSWORD",
90
+ "74": "I-PASSWORD",
91
+ "75": "B-PHONE_NUMBER",
92
+ "76": "I-PHONE_NUMBER",
93
+ "77": "B-PIN",
94
+ "78": "I-PIN",
95
+ "79": "B-POLITICAL_VIEW",
96
+ "80": "I-POLITICAL_VIEW",
97
+ "81": "B-POSTCODE",
98
+ "82": "I-POSTCODE",
99
+ "83": "B-RACE_ETHNICITY",
100
+ "84": "I-RACE_ETHNICITY",
101
+ "85": "B-RELIGIOUS_BELIEF",
102
+ "86": "I-RELIGIOUS_BELIEF",
103
+ "87": "B-SEXUALITY",
104
+ "88": "I-SEXUALITY",
105
+ "89": "B-SSN",
106
+ "90": "I-SSN",
107
+ "91": "B-STATE",
108
+ "92": "I-STATE",
109
+ "93": "B-STREET_ADDRESS",
110
+ "94": "I-STREET_ADDRESS",
111
+ "95": "B-SWIFT_BIC",
112
+ "96": "I-SWIFT_BIC",
113
+ "97": "B-TAX_ID",
114
+ "98": "I-TAX_ID",
115
+ "99": "B-TIME",
116
+ "100": "I-TIME",
117
+ "101": "B-UNIQUE_ID",
118
+ "102": "I-UNIQUE_ID",
119
+ "103": "B-URL",
120
+ "104": "I-URL",
121
+ "105": "B-USER_NAME",
122
+ "106": "I-USER_NAME",
123
+ "107": "B-VEHICLE_IDENTIFIER",
124
+ "108": "I-VEHICLE_IDENTIFIER"
125
+ },
126
+ "initializer_range": 0.02,
127
+ "intermediate_size": 3072,
128
+ "is_decoder": false,
129
+ "label2id": {
130
+ "B-ACCOUNT_NUMBER": 1,
131
+ "B-AGE": 3,
132
+ "B-API_KEY": 5,
133
+ "B-BANK_ROUTING_NUMBER": 7,
134
+ "B-BIOMETRIC_IDENTIFIER": 9,
135
+ "B-BLOOD_TYPE": 11,
136
+ "B-CERTIFICATE_LICENSE_NUMBER": 13,
137
+ "B-CITY": 15,
138
+ "B-COMPANY_NAME": 17,
139
+ "B-COORDINATE": 19,
140
+ "B-COUNTRY": 21,
141
+ "B-COUNTY": 23,
142
+ "B-CREDIT_DEBIT_CARD": 25,
143
+ "B-CUSTOMER_ID": 27,
144
+ "B-CVV": 29,
145
+ "B-DATE": 31,
146
+ "B-DATE_OF_BIRTH": 33,
147
+ "B-DATE_TIME": 35,
148
+ "B-DEVICE_IDENTIFIER": 37,
149
+ "B-EDUCATION_LEVEL": 39,
150
+ "B-EMAIL": 41,
151
+ "B-EMPLOYEE_ID": 43,
152
+ "B-EMPLOYMENT_STATUS": 45,
153
+ "B-FAX_NUMBER": 47,
154
+ "B-FIRST_NAME": 49,
155
+ "B-GENDER": 51,
156
+ "B-HEALTH_PLAN_BENEFICIARY_NUMBER": 53,
157
+ "B-HTTP_COOKIE": 55,
158
+ "B-IPV4": 57,
159
+ "B-IPV6": 59,
160
+ "B-LANGUAGE": 61,
161
+ "B-LAST_NAME": 63,
162
+ "B-LICENSE_PLATE": 65,
163
+ "B-MAC_ADDRESS": 67,
164
+ "B-MEDICAL_RECORD_NUMBER": 69,
165
+ "B-OCCUPATION": 71,
166
+ "B-PASSWORD": 73,
167
+ "B-PHONE_NUMBER": 75,
168
+ "B-PIN": 77,
169
+ "B-POLITICAL_VIEW": 79,
170
+ "B-POSTCODE": 81,
171
+ "B-RACE_ETHNICITY": 83,
172
+ "B-RELIGIOUS_BELIEF": 85,
173
+ "B-SEXUALITY": 87,
174
+ "B-SSN": 89,
175
+ "B-STATE": 91,
176
+ "B-STREET_ADDRESS": 93,
177
+ "B-SWIFT_BIC": 95,
178
+ "B-TAX_ID": 97,
179
+ "B-TIME": 99,
180
+ "B-UNIQUE_ID": 101,
181
+ "B-URL": 103,
182
+ "B-USER_NAME": 105,
183
+ "B-VEHICLE_IDENTIFIER": 107,
184
+ "I-ACCOUNT_NUMBER": 2,
185
+ "I-AGE": 4,
186
+ "I-API_KEY": 6,
187
+ "I-BANK_ROUTING_NUMBER": 8,
188
+ "I-BIOMETRIC_IDENTIFIER": 10,
189
+ "I-BLOOD_TYPE": 12,
190
+ "I-CERTIFICATE_LICENSE_NUMBER": 14,
191
+ "I-CITY": 16,
192
+ "I-COMPANY_NAME": 18,
193
+ "I-COORDINATE": 20,
194
+ "I-COUNTRY": 22,
195
+ "I-COUNTY": 24,
196
+ "I-CREDIT_DEBIT_CARD": 26,
197
+ "I-CUSTOMER_ID": 28,
198
+ "I-CVV": 30,
199
+ "I-DATE": 32,
200
+ "I-DATE_OF_BIRTH": 34,
201
+ "I-DATE_TIME": 36,
202
+ "I-DEVICE_IDENTIFIER": 38,
203
+ "I-EDUCATION_LEVEL": 40,
204
+ "I-EMAIL": 42,
205
+ "I-EMPLOYEE_ID": 44,
206
+ "I-EMPLOYMENT_STATUS": 46,
207
+ "I-FAX_NUMBER": 48,
208
+ "I-FIRST_NAME": 50,
209
+ "I-GENDER": 52,
210
+ "I-HEALTH_PLAN_BENEFICIARY_NUMBER": 54,
211
+ "I-HTTP_COOKIE": 56,
212
+ "I-IPV4": 58,
213
+ "I-IPV6": 60,
214
+ "I-LANGUAGE": 62,
215
+ "I-LAST_NAME": 64,
216
+ "I-LICENSE_PLATE": 66,
217
+ "I-MAC_ADDRESS": 68,
218
+ "I-MEDICAL_RECORD_NUMBER": 70,
219
+ "I-OCCUPATION": 72,
220
+ "I-PASSWORD": 74,
221
+ "I-PHONE_NUMBER": 76,
222
+ "I-PIN": 78,
223
+ "I-POLITICAL_VIEW": 80,
224
+ "I-POSTCODE": 82,
225
+ "I-RACE_ETHNICITY": 84,
226
+ "I-RELIGIOUS_BELIEF": 86,
227
+ "I-SEXUALITY": 88,
228
+ "I-SSN": 90,
229
+ "I-STATE": 92,
230
+ "I-STREET_ADDRESS": 94,
231
+ "I-SWIFT_BIC": 96,
232
+ "I-TAX_ID": 98,
233
+ "I-TIME": 100,
234
+ "I-UNIQUE_ID": 102,
235
+ "I-URL": 104,
236
+ "I-USER_NAME": 106,
237
+ "I-VEHICLE_IDENTIFIER": 108,
238
+ "O": 0
239
+ },
240
+ "layer_norm_eps": 1e-12,
241
+ "max_position_embeddings": 512,
242
+ "model_type": "bert",
243
+ "num_attention_heads": 12,
244
+ "num_hidden_layers": 12,
245
+ "pad_token_id": 0,
246
+ "pooler_fc_size": 768,
247
+ "pooler_num_attention_heads": 12,
248
+ "pooler_num_fc_layers": 3,
249
+ "pooler_size_per_head": 128,
250
+ "pooler_type": "first_token_transform",
251
+ "tie_word_embeddings": true,
252
+ "transformers_version": "5.3.0",
253
+ "type_vocab_size": 2,
254
+ "use_cache": false,
255
+ "vocab_size": 119547
256
+ }
nemotron-pii-ready/TokenBased-CRF/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f6bc807c5941b759066c4f6f5499c62fcb0e7471255caba1aa1b887d6a48c4c2
3
+ size 711884499
nemotron-pii-ready/TokenBased-CRF/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
nemotron-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
+ }