PuxAI commited on
Commit
bcf56bc
·
verified ·
1 Parent(s): 874a056

Upload folder using huggingface_hub

Browse files
Files changed (32) hide show
  1. nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/config.json +260 -0
  2. nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/model.safetensors +3 -0
  3. nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/optimizer.pt +3 -0
  4. nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/rng_state.pth +3 -0
  5. nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/scheduler.pt +3 -0
  6. nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/tokenizer.json +0 -0
  7. nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/tokenizer_config.json +15 -0
  8. nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/trainer_state.json +84 -0
  9. nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/training_args.bin +3 -0
  10. nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/config.json +260 -0
  11. nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/model.safetensors +3 -0
  12. nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/optimizer.pt +3 -0
  13. nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/rng_state.pth +3 -0
  14. nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/scheduler.pt +3 -0
  15. nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/tokenizer.json +0 -0
  16. nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/tokenizer_config.json +15 -0
  17. nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/trainer_state.json +134 -0
  18. nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/training_args.bin +3 -0
  19. nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/config.json +260 -0
  20. nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/model.safetensors +3 -0
  21. nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/optimizer.pt +3 -0
  22. nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/rng_state.pth +3 -0
  23. nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/scheduler.pt +3 -0
  24. nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/tokenizer.json +0 -0
  25. nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/tokenizer_config.json +15 -0
  26. nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/trainer_state.json +184 -0
  27. nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/training_args.bin +3 -0
  28. nemotron-pii-ready/TokenBased-BERT/config.json +260 -0
  29. nemotron-pii-ready/TokenBased-BERT/model.safetensors +3 -0
  30. nemotron-pii-ready/TokenBased-BERT/tokenizer.json +0 -0
  31. nemotron-pii-ready/TokenBased-BERT/tokenizer_config.json +15 -0
  32. nemotron-pii-ready/TokenBased-BERT/training_args.bin +3 -0
nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/config.json ADDED
@@ -0,0 +1,260 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_cross_attention": false,
3
+ "architectures": [
4
+ "BertForTokenClassification"
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-NATIONAL_ID",
88
+ "72": "I-NATIONAL_ID",
89
+ "73": "B-OCCUPATION",
90
+ "74": "I-OCCUPATION",
91
+ "75": "B-PASSWORD",
92
+ "76": "I-PASSWORD",
93
+ "77": "B-PHONE_NUMBER",
94
+ "78": "I-PHONE_NUMBER",
95
+ "79": "B-PIN",
96
+ "80": "I-PIN",
97
+ "81": "B-POLITICAL_VIEW",
98
+ "82": "I-POLITICAL_VIEW",
99
+ "83": "B-POSTCODE",
100
+ "84": "I-POSTCODE",
101
+ "85": "B-RACE_ETHNICITY",
102
+ "86": "I-RACE_ETHNICITY",
103
+ "87": "B-RELIGIOUS_BELIEF",
104
+ "88": "I-RELIGIOUS_BELIEF",
105
+ "89": "B-SEXUALITY",
106
+ "90": "I-SEXUALITY",
107
+ "91": "B-SSN",
108
+ "92": "I-SSN",
109
+ "93": "B-STATE",
110
+ "94": "I-STATE",
111
+ "95": "B-STREET_ADDRESS",
112
+ "96": "I-STREET_ADDRESS",
113
+ "97": "B-SWIFT_BIC",
114
+ "98": "I-SWIFT_BIC",
115
+ "99": "B-TAX_ID",
116
+ "100": "I-TAX_ID",
117
+ "101": "B-TIME",
118
+ "102": "I-TIME",
119
+ "103": "B-UNIQUE_ID",
120
+ "104": "I-UNIQUE_ID",
121
+ "105": "B-URL",
122
+ "106": "I-URL",
123
+ "107": "B-USER_NAME",
124
+ "108": "I-USER_NAME",
125
+ "109": "B-VEHICLE_IDENTIFIER",
126
+ "110": "I-VEHICLE_IDENTIFIER"
127
+ },
128
+ "initializer_range": 0.02,
129
+ "intermediate_size": 3072,
130
+ "is_decoder": false,
131
+ "label2id": {
132
+ "B-ACCOUNT_NUMBER": 1,
133
+ "B-AGE": 3,
134
+ "B-API_KEY": 5,
135
+ "B-BANK_ROUTING_NUMBER": 7,
136
+ "B-BIOMETRIC_IDENTIFIER": 9,
137
+ "B-BLOOD_TYPE": 11,
138
+ "B-CERTIFICATE_LICENSE_NUMBER": 13,
139
+ "B-CITY": 15,
140
+ "B-COMPANY_NAME": 17,
141
+ "B-COORDINATE": 19,
142
+ "B-COUNTRY": 21,
143
+ "B-COUNTY": 23,
144
+ "B-CREDIT_DEBIT_CARD": 25,
145
+ "B-CUSTOMER_ID": 27,
146
+ "B-CVV": 29,
147
+ "B-DATE": 31,
148
+ "B-DATE_OF_BIRTH": 33,
149
+ "B-DATE_TIME": 35,
150
+ "B-DEVICE_IDENTIFIER": 37,
151
+ "B-EDUCATION_LEVEL": 39,
152
+ "B-EMAIL": 41,
153
+ "B-EMPLOYEE_ID": 43,
154
+ "B-EMPLOYMENT_STATUS": 45,
155
+ "B-FAX_NUMBER": 47,
156
+ "B-FIRST_NAME": 49,
157
+ "B-GENDER": 51,
158
+ "B-HEALTH_PLAN_BENEFICIARY_NUMBER": 53,
159
+ "B-HTTP_COOKIE": 55,
160
+ "B-IPV4": 57,
161
+ "B-IPV6": 59,
162
+ "B-LANGUAGE": 61,
163
+ "B-LAST_NAME": 63,
164
+ "B-LICENSE_PLATE": 65,
165
+ "B-MAC_ADDRESS": 67,
166
+ "B-MEDICAL_RECORD_NUMBER": 69,
167
+ "B-NATIONAL_ID": 71,
168
+ "B-OCCUPATION": 73,
169
+ "B-PASSWORD": 75,
170
+ "B-PHONE_NUMBER": 77,
171
+ "B-PIN": 79,
172
+ "B-POLITICAL_VIEW": 81,
173
+ "B-POSTCODE": 83,
174
+ "B-RACE_ETHNICITY": 85,
175
+ "B-RELIGIOUS_BELIEF": 87,
176
+ "B-SEXUALITY": 89,
177
+ "B-SSN": 91,
178
+ "B-STATE": 93,
179
+ "B-STREET_ADDRESS": 95,
180
+ "B-SWIFT_BIC": 97,
181
+ "B-TAX_ID": 99,
182
+ "B-TIME": 101,
183
+ "B-UNIQUE_ID": 103,
184
+ "B-URL": 105,
185
+ "B-USER_NAME": 107,
186
+ "B-VEHICLE_IDENTIFIER": 109,
187
+ "I-ACCOUNT_NUMBER": 2,
188
+ "I-AGE": 4,
189
+ "I-API_KEY": 6,
190
+ "I-BANK_ROUTING_NUMBER": 8,
191
+ "I-BIOMETRIC_IDENTIFIER": 10,
192
+ "I-BLOOD_TYPE": 12,
193
+ "I-CERTIFICATE_LICENSE_NUMBER": 14,
194
+ "I-CITY": 16,
195
+ "I-COMPANY_NAME": 18,
196
+ "I-COORDINATE": 20,
197
+ "I-COUNTRY": 22,
198
+ "I-COUNTY": 24,
199
+ "I-CREDIT_DEBIT_CARD": 26,
200
+ "I-CUSTOMER_ID": 28,
201
+ "I-CVV": 30,
202
+ "I-DATE": 32,
203
+ "I-DATE_OF_BIRTH": 34,
204
+ "I-DATE_TIME": 36,
205
+ "I-DEVICE_IDENTIFIER": 38,
206
+ "I-EDUCATION_LEVEL": 40,
207
+ "I-EMAIL": 42,
208
+ "I-EMPLOYEE_ID": 44,
209
+ "I-EMPLOYMENT_STATUS": 46,
210
+ "I-FAX_NUMBER": 48,
211
+ "I-FIRST_NAME": 50,
212
+ "I-GENDER": 52,
213
+ "I-HEALTH_PLAN_BENEFICIARY_NUMBER": 54,
214
+ "I-HTTP_COOKIE": 56,
215
+ "I-IPV4": 58,
216
+ "I-IPV6": 60,
217
+ "I-LANGUAGE": 62,
218
+ "I-LAST_NAME": 64,
219
+ "I-LICENSE_PLATE": 66,
220
+ "I-MAC_ADDRESS": 68,
221
+ "I-MEDICAL_RECORD_NUMBER": 70,
222
+ "I-NATIONAL_ID": 72,
223
+ "I-OCCUPATION": 74,
224
+ "I-PASSWORD": 76,
225
+ "I-PHONE_NUMBER": 78,
226
+ "I-PIN": 80,
227
+ "I-POLITICAL_VIEW": 82,
228
+ "I-POSTCODE": 84,
229
+ "I-RACE_ETHNICITY": 86,
230
+ "I-RELIGIOUS_BELIEF": 88,
231
+ "I-SEXUALITY": 90,
232
+ "I-SSN": 92,
233
+ "I-STATE": 94,
234
+ "I-STREET_ADDRESS": 96,
235
+ "I-SWIFT_BIC": 98,
236
+ "I-TAX_ID": 100,
237
+ "I-TIME": 102,
238
+ "I-UNIQUE_ID": 104,
239
+ "I-URL": 106,
240
+ "I-USER_NAME": 108,
241
+ "I-VEHICLE_IDENTIFIER": 110,
242
+ "O": 0
243
+ },
244
+ "layer_norm_eps": 1e-12,
245
+ "max_position_embeddings": 512,
246
+ "model_type": "bert",
247
+ "num_attention_heads": 12,
248
+ "num_hidden_layers": 12,
249
+ "pad_token_id": 0,
250
+ "pooler_fc_size": 768,
251
+ "pooler_num_attention_heads": 12,
252
+ "pooler_num_fc_layers": 3,
253
+ "pooler_size_per_head": 128,
254
+ "pooler_type": "first_token_transform",
255
+ "tie_word_embeddings": true,
256
+ "transformers_version": "5.3.0",
257
+ "type_vocab_size": 2,
258
+ "use_cache": false,
259
+ "vocab_size": 119547
260
+ }
nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8bec580239feb4fcd8a8da40f1fb5aa5881527236ac19654717da1ca0c160a00
3
+ size 709416164
nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:793b63b85ac233d2c22e0f7e113b5cec7e58537ab768004c840df554834539b0
3
+ size 1418955275
nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:20166816c78861a3cde613fc81c3f9869a6352f17b0ef6764223644675ea6822
3
+ size 14645
nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef29a928ee0235888b782d54ff175831c38ba4240becc0446a6c0fbd25ab546c
3
+ size 1465
nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/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-BERT/checkpoint-3125/trainer_state.json ADDED
@@ -0,0 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 3125,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.16,
14
+ "grad_norm": 0.8835734128952026,
15
+ "learning_rate": 1.893546666666667e-05,
16
+ "loss": 0.27433673095703126,
17
+ "step": 500
18
+ },
19
+ {
20
+ "epoch": 0.32,
21
+ "grad_norm": 0.8788529634475708,
22
+ "learning_rate": 1.78688e-05,
23
+ "loss": 0.051882022857666014,
24
+ "step": 1000
25
+ },
26
+ {
27
+ "epoch": 0.48,
28
+ "grad_norm": 0.35897135734558105,
29
+ "learning_rate": 1.6802133333333336e-05,
30
+ "loss": 0.04031004333496094,
31
+ "step": 1500
32
+ },
33
+ {
34
+ "epoch": 0.64,
35
+ "grad_norm": 0.2843490540981293,
36
+ "learning_rate": 1.573546666666667e-05,
37
+ "loss": 0.034954586029052735,
38
+ "step": 2000
39
+ },
40
+ {
41
+ "epoch": 0.8,
42
+ "grad_norm": 0.2961885333061218,
43
+ "learning_rate": 1.4668800000000001e-05,
44
+ "loss": 0.03310459899902344,
45
+ "step": 2500
46
+ },
47
+ {
48
+ "epoch": 0.96,
49
+ "grad_norm": 0.6217581033706665,
50
+ "learning_rate": 1.3602133333333333e-05,
51
+ "loss": 0.02999618911743164,
52
+ "step": 3000
53
+ },
54
+ {
55
+ "epoch": 1.0,
56
+ "eval_loss": 0.032662831246852875,
57
+ "eval_runtime": 68.1445,
58
+ "eval_samples_per_second": 293.494,
59
+ "eval_steps_per_second": 36.687,
60
+ "step": 3125
61
+ }
62
+ ],
63
+ "logging_steps": 500,
64
+ "max_steps": 9375,
65
+ "num_input_tokens_seen": 0,
66
+ "num_train_epochs": 3,
67
+ "save_steps": 500,
68
+ "stateful_callbacks": {
69
+ "TrainerControl": {
70
+ "args": {
71
+ "should_epoch_stop": false,
72
+ "should_evaluate": false,
73
+ "should_log": false,
74
+ "should_save": true,
75
+ "should_training_stop": false
76
+ },
77
+ "attributes": {}
78
+ }
79
+ },
80
+ "total_flos": 2.61554254848e+16,
81
+ "train_batch_size": 32,
82
+ "trial_name": null,
83
+ "trial_params": null
84
+ }
nemotron-pii-ready/TokenBased-BERT/checkpoint-3125/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54258d8cd6ea85d5af9ffa138c92bf8a8c9ec44c6cb90d9f1043a2f004d015d6
3
+ size 5201
nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/config.json ADDED
@@ -0,0 +1,260 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_cross_attention": false,
3
+ "architectures": [
4
+ "BertForTokenClassification"
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-NATIONAL_ID",
88
+ "72": "I-NATIONAL_ID",
89
+ "73": "B-OCCUPATION",
90
+ "74": "I-OCCUPATION",
91
+ "75": "B-PASSWORD",
92
+ "76": "I-PASSWORD",
93
+ "77": "B-PHONE_NUMBER",
94
+ "78": "I-PHONE_NUMBER",
95
+ "79": "B-PIN",
96
+ "80": "I-PIN",
97
+ "81": "B-POLITICAL_VIEW",
98
+ "82": "I-POLITICAL_VIEW",
99
+ "83": "B-POSTCODE",
100
+ "84": "I-POSTCODE",
101
+ "85": "B-RACE_ETHNICITY",
102
+ "86": "I-RACE_ETHNICITY",
103
+ "87": "B-RELIGIOUS_BELIEF",
104
+ "88": "I-RELIGIOUS_BELIEF",
105
+ "89": "B-SEXUALITY",
106
+ "90": "I-SEXUALITY",
107
+ "91": "B-SSN",
108
+ "92": "I-SSN",
109
+ "93": "B-STATE",
110
+ "94": "I-STATE",
111
+ "95": "B-STREET_ADDRESS",
112
+ "96": "I-STREET_ADDRESS",
113
+ "97": "B-SWIFT_BIC",
114
+ "98": "I-SWIFT_BIC",
115
+ "99": "B-TAX_ID",
116
+ "100": "I-TAX_ID",
117
+ "101": "B-TIME",
118
+ "102": "I-TIME",
119
+ "103": "B-UNIQUE_ID",
120
+ "104": "I-UNIQUE_ID",
121
+ "105": "B-URL",
122
+ "106": "I-URL",
123
+ "107": "B-USER_NAME",
124
+ "108": "I-USER_NAME",
125
+ "109": "B-VEHICLE_IDENTIFIER",
126
+ "110": "I-VEHICLE_IDENTIFIER"
127
+ },
128
+ "initializer_range": 0.02,
129
+ "intermediate_size": 3072,
130
+ "is_decoder": false,
131
+ "label2id": {
132
+ "B-ACCOUNT_NUMBER": 1,
133
+ "B-AGE": 3,
134
+ "B-API_KEY": 5,
135
+ "B-BANK_ROUTING_NUMBER": 7,
136
+ "B-BIOMETRIC_IDENTIFIER": 9,
137
+ "B-BLOOD_TYPE": 11,
138
+ "B-CERTIFICATE_LICENSE_NUMBER": 13,
139
+ "B-CITY": 15,
140
+ "B-COMPANY_NAME": 17,
141
+ "B-COORDINATE": 19,
142
+ "B-COUNTRY": 21,
143
+ "B-COUNTY": 23,
144
+ "B-CREDIT_DEBIT_CARD": 25,
145
+ "B-CUSTOMER_ID": 27,
146
+ "B-CVV": 29,
147
+ "B-DATE": 31,
148
+ "B-DATE_OF_BIRTH": 33,
149
+ "B-DATE_TIME": 35,
150
+ "B-DEVICE_IDENTIFIER": 37,
151
+ "B-EDUCATION_LEVEL": 39,
152
+ "B-EMAIL": 41,
153
+ "B-EMPLOYEE_ID": 43,
154
+ "B-EMPLOYMENT_STATUS": 45,
155
+ "B-FAX_NUMBER": 47,
156
+ "B-FIRST_NAME": 49,
157
+ "B-GENDER": 51,
158
+ "B-HEALTH_PLAN_BENEFICIARY_NUMBER": 53,
159
+ "B-HTTP_COOKIE": 55,
160
+ "B-IPV4": 57,
161
+ "B-IPV6": 59,
162
+ "B-LANGUAGE": 61,
163
+ "B-LAST_NAME": 63,
164
+ "B-LICENSE_PLATE": 65,
165
+ "B-MAC_ADDRESS": 67,
166
+ "B-MEDICAL_RECORD_NUMBER": 69,
167
+ "B-NATIONAL_ID": 71,
168
+ "B-OCCUPATION": 73,
169
+ "B-PASSWORD": 75,
170
+ "B-PHONE_NUMBER": 77,
171
+ "B-PIN": 79,
172
+ "B-POLITICAL_VIEW": 81,
173
+ "B-POSTCODE": 83,
174
+ "B-RACE_ETHNICITY": 85,
175
+ "B-RELIGIOUS_BELIEF": 87,
176
+ "B-SEXUALITY": 89,
177
+ "B-SSN": 91,
178
+ "B-STATE": 93,
179
+ "B-STREET_ADDRESS": 95,
180
+ "B-SWIFT_BIC": 97,
181
+ "B-TAX_ID": 99,
182
+ "B-TIME": 101,
183
+ "B-UNIQUE_ID": 103,
184
+ "B-URL": 105,
185
+ "B-USER_NAME": 107,
186
+ "B-VEHICLE_IDENTIFIER": 109,
187
+ "I-ACCOUNT_NUMBER": 2,
188
+ "I-AGE": 4,
189
+ "I-API_KEY": 6,
190
+ "I-BANK_ROUTING_NUMBER": 8,
191
+ "I-BIOMETRIC_IDENTIFIER": 10,
192
+ "I-BLOOD_TYPE": 12,
193
+ "I-CERTIFICATE_LICENSE_NUMBER": 14,
194
+ "I-CITY": 16,
195
+ "I-COMPANY_NAME": 18,
196
+ "I-COORDINATE": 20,
197
+ "I-COUNTRY": 22,
198
+ "I-COUNTY": 24,
199
+ "I-CREDIT_DEBIT_CARD": 26,
200
+ "I-CUSTOMER_ID": 28,
201
+ "I-CVV": 30,
202
+ "I-DATE": 32,
203
+ "I-DATE_OF_BIRTH": 34,
204
+ "I-DATE_TIME": 36,
205
+ "I-DEVICE_IDENTIFIER": 38,
206
+ "I-EDUCATION_LEVEL": 40,
207
+ "I-EMAIL": 42,
208
+ "I-EMPLOYEE_ID": 44,
209
+ "I-EMPLOYMENT_STATUS": 46,
210
+ "I-FAX_NUMBER": 48,
211
+ "I-FIRST_NAME": 50,
212
+ "I-GENDER": 52,
213
+ "I-HEALTH_PLAN_BENEFICIARY_NUMBER": 54,
214
+ "I-HTTP_COOKIE": 56,
215
+ "I-IPV4": 58,
216
+ "I-IPV6": 60,
217
+ "I-LANGUAGE": 62,
218
+ "I-LAST_NAME": 64,
219
+ "I-LICENSE_PLATE": 66,
220
+ "I-MAC_ADDRESS": 68,
221
+ "I-MEDICAL_RECORD_NUMBER": 70,
222
+ "I-NATIONAL_ID": 72,
223
+ "I-OCCUPATION": 74,
224
+ "I-PASSWORD": 76,
225
+ "I-PHONE_NUMBER": 78,
226
+ "I-PIN": 80,
227
+ "I-POLITICAL_VIEW": 82,
228
+ "I-POSTCODE": 84,
229
+ "I-RACE_ETHNICITY": 86,
230
+ "I-RELIGIOUS_BELIEF": 88,
231
+ "I-SEXUALITY": 90,
232
+ "I-SSN": 92,
233
+ "I-STATE": 94,
234
+ "I-STREET_ADDRESS": 96,
235
+ "I-SWIFT_BIC": 98,
236
+ "I-TAX_ID": 100,
237
+ "I-TIME": 102,
238
+ "I-UNIQUE_ID": 104,
239
+ "I-URL": 106,
240
+ "I-USER_NAME": 108,
241
+ "I-VEHICLE_IDENTIFIER": 110,
242
+ "O": 0
243
+ },
244
+ "layer_norm_eps": 1e-12,
245
+ "max_position_embeddings": 512,
246
+ "model_type": "bert",
247
+ "num_attention_heads": 12,
248
+ "num_hidden_layers": 12,
249
+ "pad_token_id": 0,
250
+ "pooler_fc_size": 768,
251
+ "pooler_num_attention_heads": 12,
252
+ "pooler_num_fc_layers": 3,
253
+ "pooler_size_per_head": 128,
254
+ "pooler_type": "first_token_transform",
255
+ "tie_word_embeddings": true,
256
+ "transformers_version": "5.3.0",
257
+ "type_vocab_size": 2,
258
+ "use_cache": false,
259
+ "vocab_size": 119547
260
+ }
nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ecadc1d3a0982c73b8c1c0dac81700b93df1ece6f1acfdcdd4f99abaa56216b
3
+ size 709416164
nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:793ef19316c629ffbb05a08b1ea447cce67c230ae88c335bdf9bfe76f6a65914
3
+ size 1418955275
nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c0af9750a81ff32f7f9dde78071ca77d155e8004f11102e0baa890814b3e9d69
3
+ size 14645
nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:32c80219e876dd7af92bb807a5228e367cbf4ca1f261222952102ad14cd47606
3
+ size 1465
nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/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-BERT/checkpoint-6250/trainer_state.json ADDED
@@ -0,0 +1,134 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 2.0,
6
+ "eval_steps": 500,
7
+ "global_step": 6250,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.16,
14
+ "grad_norm": 0.8835734128952026,
15
+ "learning_rate": 1.893546666666667e-05,
16
+ "loss": 0.27433673095703126,
17
+ "step": 500
18
+ },
19
+ {
20
+ "epoch": 0.32,
21
+ "grad_norm": 0.8788529634475708,
22
+ "learning_rate": 1.78688e-05,
23
+ "loss": 0.051882022857666014,
24
+ "step": 1000
25
+ },
26
+ {
27
+ "epoch": 0.48,
28
+ "grad_norm": 0.35897135734558105,
29
+ "learning_rate": 1.6802133333333336e-05,
30
+ "loss": 0.04031004333496094,
31
+ "step": 1500
32
+ },
33
+ {
34
+ "epoch": 0.64,
35
+ "grad_norm": 0.2843490540981293,
36
+ "learning_rate": 1.573546666666667e-05,
37
+ "loss": 0.034954586029052735,
38
+ "step": 2000
39
+ },
40
+ {
41
+ "epoch": 0.8,
42
+ "grad_norm": 0.2961885333061218,
43
+ "learning_rate": 1.4668800000000001e-05,
44
+ "loss": 0.03310459899902344,
45
+ "step": 2500
46
+ },
47
+ {
48
+ "epoch": 0.96,
49
+ "grad_norm": 0.6217581033706665,
50
+ "learning_rate": 1.3602133333333333e-05,
51
+ "loss": 0.02999618911743164,
52
+ "step": 3000
53
+ },
54
+ {
55
+ "epoch": 1.0,
56
+ "eval_loss": 0.032662831246852875,
57
+ "eval_runtime": 68.1445,
58
+ "eval_samples_per_second": 293.494,
59
+ "eval_steps_per_second": 36.687,
60
+ "step": 3125
61
+ },
62
+ {
63
+ "epoch": 1.12,
64
+ "grad_norm": 0.39666157960891724,
65
+ "learning_rate": 1.2535466666666667e-05,
66
+ "loss": 0.02726400375366211,
67
+ "step": 3500
68
+ },
69
+ {
70
+ "epoch": 1.28,
71
+ "grad_norm": 0.45859163999557495,
72
+ "learning_rate": 1.14688e-05,
73
+ "loss": 0.02493283271789551,
74
+ "step": 4000
75
+ },
76
+ {
77
+ "epoch": 1.44,
78
+ "grad_norm": 0.26409512758255005,
79
+ "learning_rate": 1.0402133333333335e-05,
80
+ "loss": 0.024670473098754885,
81
+ "step": 4500
82
+ },
83
+ {
84
+ "epoch": 1.6,
85
+ "grad_norm": 0.8010464906692505,
86
+ "learning_rate": 9.335466666666667e-06,
87
+ "loss": 0.02451570510864258,
88
+ "step": 5000
89
+ },
90
+ {
91
+ "epoch": 1.76,
92
+ "grad_norm": 0.550110399723053,
93
+ "learning_rate": 8.2688e-06,
94
+ "loss": 0.022706100463867188,
95
+ "step": 5500
96
+ },
97
+ {
98
+ "epoch": 1.92,
99
+ "grad_norm": 0.34001559019088745,
100
+ "learning_rate": 7.202133333333334e-06,
101
+ "loss": 0.0224871883392334,
102
+ "step": 6000
103
+ },
104
+ {
105
+ "epoch": 2.0,
106
+ "eval_loss": 0.02736319601535797,
107
+ "eval_runtime": 71.557,
108
+ "eval_samples_per_second": 279.498,
109
+ "eval_steps_per_second": 34.937,
110
+ "step": 6250
111
+ }
112
+ ],
113
+ "logging_steps": 500,
114
+ "max_steps": 9375,
115
+ "num_input_tokens_seen": 0,
116
+ "num_train_epochs": 3,
117
+ "save_steps": 500,
118
+ "stateful_callbacks": {
119
+ "TrainerControl": {
120
+ "args": {
121
+ "should_epoch_stop": false,
122
+ "should_evaluate": false,
123
+ "should_log": false,
124
+ "should_save": true,
125
+ "should_training_stop": false
126
+ },
127
+ "attributes": {}
128
+ }
129
+ },
130
+ "total_flos": 5.23108509696e+16,
131
+ "train_batch_size": 32,
132
+ "trial_name": null,
133
+ "trial_params": null
134
+ }
nemotron-pii-ready/TokenBased-BERT/checkpoint-6250/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54258d8cd6ea85d5af9ffa138c92bf8a8c9ec44c6cb90d9f1043a2f004d015d6
3
+ size 5201
nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/config.json ADDED
@@ -0,0 +1,260 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_cross_attention": false,
3
+ "architectures": [
4
+ "BertForTokenClassification"
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-NATIONAL_ID",
88
+ "72": "I-NATIONAL_ID",
89
+ "73": "B-OCCUPATION",
90
+ "74": "I-OCCUPATION",
91
+ "75": "B-PASSWORD",
92
+ "76": "I-PASSWORD",
93
+ "77": "B-PHONE_NUMBER",
94
+ "78": "I-PHONE_NUMBER",
95
+ "79": "B-PIN",
96
+ "80": "I-PIN",
97
+ "81": "B-POLITICAL_VIEW",
98
+ "82": "I-POLITICAL_VIEW",
99
+ "83": "B-POSTCODE",
100
+ "84": "I-POSTCODE",
101
+ "85": "B-RACE_ETHNICITY",
102
+ "86": "I-RACE_ETHNICITY",
103
+ "87": "B-RELIGIOUS_BELIEF",
104
+ "88": "I-RELIGIOUS_BELIEF",
105
+ "89": "B-SEXUALITY",
106
+ "90": "I-SEXUALITY",
107
+ "91": "B-SSN",
108
+ "92": "I-SSN",
109
+ "93": "B-STATE",
110
+ "94": "I-STATE",
111
+ "95": "B-STREET_ADDRESS",
112
+ "96": "I-STREET_ADDRESS",
113
+ "97": "B-SWIFT_BIC",
114
+ "98": "I-SWIFT_BIC",
115
+ "99": "B-TAX_ID",
116
+ "100": "I-TAX_ID",
117
+ "101": "B-TIME",
118
+ "102": "I-TIME",
119
+ "103": "B-UNIQUE_ID",
120
+ "104": "I-UNIQUE_ID",
121
+ "105": "B-URL",
122
+ "106": "I-URL",
123
+ "107": "B-USER_NAME",
124
+ "108": "I-USER_NAME",
125
+ "109": "B-VEHICLE_IDENTIFIER",
126
+ "110": "I-VEHICLE_IDENTIFIER"
127
+ },
128
+ "initializer_range": 0.02,
129
+ "intermediate_size": 3072,
130
+ "is_decoder": false,
131
+ "label2id": {
132
+ "B-ACCOUNT_NUMBER": 1,
133
+ "B-AGE": 3,
134
+ "B-API_KEY": 5,
135
+ "B-BANK_ROUTING_NUMBER": 7,
136
+ "B-BIOMETRIC_IDENTIFIER": 9,
137
+ "B-BLOOD_TYPE": 11,
138
+ "B-CERTIFICATE_LICENSE_NUMBER": 13,
139
+ "B-CITY": 15,
140
+ "B-COMPANY_NAME": 17,
141
+ "B-COORDINATE": 19,
142
+ "B-COUNTRY": 21,
143
+ "B-COUNTY": 23,
144
+ "B-CREDIT_DEBIT_CARD": 25,
145
+ "B-CUSTOMER_ID": 27,
146
+ "B-CVV": 29,
147
+ "B-DATE": 31,
148
+ "B-DATE_OF_BIRTH": 33,
149
+ "B-DATE_TIME": 35,
150
+ "B-DEVICE_IDENTIFIER": 37,
151
+ "B-EDUCATION_LEVEL": 39,
152
+ "B-EMAIL": 41,
153
+ "B-EMPLOYEE_ID": 43,
154
+ "B-EMPLOYMENT_STATUS": 45,
155
+ "B-FAX_NUMBER": 47,
156
+ "B-FIRST_NAME": 49,
157
+ "B-GENDER": 51,
158
+ "B-HEALTH_PLAN_BENEFICIARY_NUMBER": 53,
159
+ "B-HTTP_COOKIE": 55,
160
+ "B-IPV4": 57,
161
+ "B-IPV6": 59,
162
+ "B-LANGUAGE": 61,
163
+ "B-LAST_NAME": 63,
164
+ "B-LICENSE_PLATE": 65,
165
+ "B-MAC_ADDRESS": 67,
166
+ "B-MEDICAL_RECORD_NUMBER": 69,
167
+ "B-NATIONAL_ID": 71,
168
+ "B-OCCUPATION": 73,
169
+ "B-PASSWORD": 75,
170
+ "B-PHONE_NUMBER": 77,
171
+ "B-PIN": 79,
172
+ "B-POLITICAL_VIEW": 81,
173
+ "B-POSTCODE": 83,
174
+ "B-RACE_ETHNICITY": 85,
175
+ "B-RELIGIOUS_BELIEF": 87,
176
+ "B-SEXUALITY": 89,
177
+ "B-SSN": 91,
178
+ "B-STATE": 93,
179
+ "B-STREET_ADDRESS": 95,
180
+ "B-SWIFT_BIC": 97,
181
+ "B-TAX_ID": 99,
182
+ "B-TIME": 101,
183
+ "B-UNIQUE_ID": 103,
184
+ "B-URL": 105,
185
+ "B-USER_NAME": 107,
186
+ "B-VEHICLE_IDENTIFIER": 109,
187
+ "I-ACCOUNT_NUMBER": 2,
188
+ "I-AGE": 4,
189
+ "I-API_KEY": 6,
190
+ "I-BANK_ROUTING_NUMBER": 8,
191
+ "I-BIOMETRIC_IDENTIFIER": 10,
192
+ "I-BLOOD_TYPE": 12,
193
+ "I-CERTIFICATE_LICENSE_NUMBER": 14,
194
+ "I-CITY": 16,
195
+ "I-COMPANY_NAME": 18,
196
+ "I-COORDINATE": 20,
197
+ "I-COUNTRY": 22,
198
+ "I-COUNTY": 24,
199
+ "I-CREDIT_DEBIT_CARD": 26,
200
+ "I-CUSTOMER_ID": 28,
201
+ "I-CVV": 30,
202
+ "I-DATE": 32,
203
+ "I-DATE_OF_BIRTH": 34,
204
+ "I-DATE_TIME": 36,
205
+ "I-DEVICE_IDENTIFIER": 38,
206
+ "I-EDUCATION_LEVEL": 40,
207
+ "I-EMAIL": 42,
208
+ "I-EMPLOYEE_ID": 44,
209
+ "I-EMPLOYMENT_STATUS": 46,
210
+ "I-FAX_NUMBER": 48,
211
+ "I-FIRST_NAME": 50,
212
+ "I-GENDER": 52,
213
+ "I-HEALTH_PLAN_BENEFICIARY_NUMBER": 54,
214
+ "I-HTTP_COOKIE": 56,
215
+ "I-IPV4": 58,
216
+ "I-IPV6": 60,
217
+ "I-LANGUAGE": 62,
218
+ "I-LAST_NAME": 64,
219
+ "I-LICENSE_PLATE": 66,
220
+ "I-MAC_ADDRESS": 68,
221
+ "I-MEDICAL_RECORD_NUMBER": 70,
222
+ "I-NATIONAL_ID": 72,
223
+ "I-OCCUPATION": 74,
224
+ "I-PASSWORD": 76,
225
+ "I-PHONE_NUMBER": 78,
226
+ "I-PIN": 80,
227
+ "I-POLITICAL_VIEW": 82,
228
+ "I-POSTCODE": 84,
229
+ "I-RACE_ETHNICITY": 86,
230
+ "I-RELIGIOUS_BELIEF": 88,
231
+ "I-SEXUALITY": 90,
232
+ "I-SSN": 92,
233
+ "I-STATE": 94,
234
+ "I-STREET_ADDRESS": 96,
235
+ "I-SWIFT_BIC": 98,
236
+ "I-TAX_ID": 100,
237
+ "I-TIME": 102,
238
+ "I-UNIQUE_ID": 104,
239
+ "I-URL": 106,
240
+ "I-USER_NAME": 108,
241
+ "I-VEHICLE_IDENTIFIER": 110,
242
+ "O": 0
243
+ },
244
+ "layer_norm_eps": 1e-12,
245
+ "max_position_embeddings": 512,
246
+ "model_type": "bert",
247
+ "num_attention_heads": 12,
248
+ "num_hidden_layers": 12,
249
+ "pad_token_id": 0,
250
+ "pooler_fc_size": 768,
251
+ "pooler_num_attention_heads": 12,
252
+ "pooler_num_fc_layers": 3,
253
+ "pooler_size_per_head": 128,
254
+ "pooler_type": "first_token_transform",
255
+ "tie_word_embeddings": true,
256
+ "transformers_version": "5.3.0",
257
+ "type_vocab_size": 2,
258
+ "use_cache": false,
259
+ "vocab_size": 119547
260
+ }
nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:260ff5ff37bb85d1af3c0b1d780ea4a36d7ac2b2009a7087e6451c18ba7e0741
3
+ size 709416164
nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b97632b15a3bea9754c085b3c7591d162f8e366c9c2f8385e448c13827db8bb3
3
+ size 1418955275
nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d13f7f343ad6761ef3362358fd7ca51b0abba297d1b390935a3fc25cc58933ff
3
+ size 14645
nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9a05ede63ffbd919598e306777d0c9bad809a60634784dd51e6ae2428c382b97
3
+ size 1465
nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/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-BERT/checkpoint-9375/trainer_state.json ADDED
@@ -0,0 +1,184 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 3.0,
6
+ "eval_steps": 500,
7
+ "global_step": 9375,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.16,
14
+ "grad_norm": 0.8835734128952026,
15
+ "learning_rate": 1.893546666666667e-05,
16
+ "loss": 0.27433673095703126,
17
+ "step": 500
18
+ },
19
+ {
20
+ "epoch": 0.32,
21
+ "grad_norm": 0.8788529634475708,
22
+ "learning_rate": 1.78688e-05,
23
+ "loss": 0.051882022857666014,
24
+ "step": 1000
25
+ },
26
+ {
27
+ "epoch": 0.48,
28
+ "grad_norm": 0.35897135734558105,
29
+ "learning_rate": 1.6802133333333336e-05,
30
+ "loss": 0.04031004333496094,
31
+ "step": 1500
32
+ },
33
+ {
34
+ "epoch": 0.64,
35
+ "grad_norm": 0.2843490540981293,
36
+ "learning_rate": 1.573546666666667e-05,
37
+ "loss": 0.034954586029052735,
38
+ "step": 2000
39
+ },
40
+ {
41
+ "epoch": 0.8,
42
+ "grad_norm": 0.2961885333061218,
43
+ "learning_rate": 1.4668800000000001e-05,
44
+ "loss": 0.03310459899902344,
45
+ "step": 2500
46
+ },
47
+ {
48
+ "epoch": 0.96,
49
+ "grad_norm": 0.6217581033706665,
50
+ "learning_rate": 1.3602133333333333e-05,
51
+ "loss": 0.02999618911743164,
52
+ "step": 3000
53
+ },
54
+ {
55
+ "epoch": 1.0,
56
+ "eval_loss": 0.032662831246852875,
57
+ "eval_runtime": 68.1445,
58
+ "eval_samples_per_second": 293.494,
59
+ "eval_steps_per_second": 36.687,
60
+ "step": 3125
61
+ },
62
+ {
63
+ "epoch": 1.12,
64
+ "grad_norm": 0.39666157960891724,
65
+ "learning_rate": 1.2535466666666667e-05,
66
+ "loss": 0.02726400375366211,
67
+ "step": 3500
68
+ },
69
+ {
70
+ "epoch": 1.28,
71
+ "grad_norm": 0.45859163999557495,
72
+ "learning_rate": 1.14688e-05,
73
+ "loss": 0.02493283271789551,
74
+ "step": 4000
75
+ },
76
+ {
77
+ "epoch": 1.44,
78
+ "grad_norm": 0.26409512758255005,
79
+ "learning_rate": 1.0402133333333335e-05,
80
+ "loss": 0.024670473098754885,
81
+ "step": 4500
82
+ },
83
+ {
84
+ "epoch": 1.6,
85
+ "grad_norm": 0.8010464906692505,
86
+ "learning_rate": 9.335466666666667e-06,
87
+ "loss": 0.02451570510864258,
88
+ "step": 5000
89
+ },
90
+ {
91
+ "epoch": 1.76,
92
+ "grad_norm": 0.550110399723053,
93
+ "learning_rate": 8.2688e-06,
94
+ "loss": 0.022706100463867188,
95
+ "step": 5500
96
+ },
97
+ {
98
+ "epoch": 1.92,
99
+ "grad_norm": 0.34001559019088745,
100
+ "learning_rate": 7.202133333333334e-06,
101
+ "loss": 0.0224871883392334,
102
+ "step": 6000
103
+ },
104
+ {
105
+ "epoch": 2.0,
106
+ "eval_loss": 0.02736319601535797,
107
+ "eval_runtime": 71.557,
108
+ "eval_samples_per_second": 279.498,
109
+ "eval_steps_per_second": 34.937,
110
+ "step": 6250
111
+ },
112
+ {
113
+ "epoch": 2.08,
114
+ "grad_norm": 0.6057286262512207,
115
+ "learning_rate": 6.135466666666667e-06,
116
+ "loss": 0.02155326271057129,
117
+ "step": 6500
118
+ },
119
+ {
120
+ "epoch": 2.24,
121
+ "grad_norm": 0.2852473855018616,
122
+ "learning_rate": 5.0688000000000005e-06,
123
+ "loss": 0.01856932258605957,
124
+ "step": 7000
125
+ },
126
+ {
127
+ "epoch": 2.4,
128
+ "grad_norm": 0.3777627646923065,
129
+ "learning_rate": 4.002133333333334e-06,
130
+ "loss": 0.019388118743896485,
131
+ "step": 7500
132
+ },
133
+ {
134
+ "epoch": 2.56,
135
+ "grad_norm": 0.2790425717830658,
136
+ "learning_rate": 2.935466666666667e-06,
137
+ "loss": 0.018188573837280272,
138
+ "step": 8000
139
+ },
140
+ {
141
+ "epoch": 2.7199999999999998,
142
+ "grad_norm": 0.2212079018354416,
143
+ "learning_rate": 1.8688e-06,
144
+ "loss": 0.018064674377441407,
145
+ "step": 8500
146
+ },
147
+ {
148
+ "epoch": 2.88,
149
+ "grad_norm": 0.2295033037662506,
150
+ "learning_rate": 8.021333333333334e-07,
151
+ "loss": 0.018154245376586914,
152
+ "step": 9000
153
+ },
154
+ {
155
+ "epoch": 3.0,
156
+ "eval_loss": 0.026027819141745567,
157
+ "eval_runtime": 66.1295,
158
+ "eval_samples_per_second": 302.437,
159
+ "eval_steps_per_second": 37.805,
160
+ "step": 9375
161
+ }
162
+ ],
163
+ "logging_steps": 500,
164
+ "max_steps": 9375,
165
+ "num_input_tokens_seen": 0,
166
+ "num_train_epochs": 3,
167
+ "save_steps": 500,
168
+ "stateful_callbacks": {
169
+ "TrainerControl": {
170
+ "args": {
171
+ "should_epoch_stop": false,
172
+ "should_evaluate": false,
173
+ "should_log": false,
174
+ "should_save": true,
175
+ "should_training_stop": true
176
+ },
177
+ "attributes": {}
178
+ }
179
+ },
180
+ "total_flos": 7.84662764544e+16,
181
+ "train_batch_size": 32,
182
+ "trial_name": null,
183
+ "trial_params": null
184
+ }
nemotron-pii-ready/TokenBased-BERT/checkpoint-9375/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54258d8cd6ea85d5af9ffa138c92bf8a8c9ec44c6cb90d9f1043a2f004d015d6
3
+ size 5201
nemotron-pii-ready/TokenBased-BERT/config.json ADDED
@@ -0,0 +1,260 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_cross_attention": false,
3
+ "architectures": [
4
+ "BertForTokenClassification"
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-NATIONAL_ID",
88
+ "72": "I-NATIONAL_ID",
89
+ "73": "B-OCCUPATION",
90
+ "74": "I-OCCUPATION",
91
+ "75": "B-PASSWORD",
92
+ "76": "I-PASSWORD",
93
+ "77": "B-PHONE_NUMBER",
94
+ "78": "I-PHONE_NUMBER",
95
+ "79": "B-PIN",
96
+ "80": "I-PIN",
97
+ "81": "B-POLITICAL_VIEW",
98
+ "82": "I-POLITICAL_VIEW",
99
+ "83": "B-POSTCODE",
100
+ "84": "I-POSTCODE",
101
+ "85": "B-RACE_ETHNICITY",
102
+ "86": "I-RACE_ETHNICITY",
103
+ "87": "B-RELIGIOUS_BELIEF",
104
+ "88": "I-RELIGIOUS_BELIEF",
105
+ "89": "B-SEXUALITY",
106
+ "90": "I-SEXUALITY",
107
+ "91": "B-SSN",
108
+ "92": "I-SSN",
109
+ "93": "B-STATE",
110
+ "94": "I-STATE",
111
+ "95": "B-STREET_ADDRESS",
112
+ "96": "I-STREET_ADDRESS",
113
+ "97": "B-SWIFT_BIC",
114
+ "98": "I-SWIFT_BIC",
115
+ "99": "B-TAX_ID",
116
+ "100": "I-TAX_ID",
117
+ "101": "B-TIME",
118
+ "102": "I-TIME",
119
+ "103": "B-UNIQUE_ID",
120
+ "104": "I-UNIQUE_ID",
121
+ "105": "B-URL",
122
+ "106": "I-URL",
123
+ "107": "B-USER_NAME",
124
+ "108": "I-USER_NAME",
125
+ "109": "B-VEHICLE_IDENTIFIER",
126
+ "110": "I-VEHICLE_IDENTIFIER"
127
+ },
128
+ "initializer_range": 0.02,
129
+ "intermediate_size": 3072,
130
+ "is_decoder": false,
131
+ "label2id": {
132
+ "B-ACCOUNT_NUMBER": 1,
133
+ "B-AGE": 3,
134
+ "B-API_KEY": 5,
135
+ "B-BANK_ROUTING_NUMBER": 7,
136
+ "B-BIOMETRIC_IDENTIFIER": 9,
137
+ "B-BLOOD_TYPE": 11,
138
+ "B-CERTIFICATE_LICENSE_NUMBER": 13,
139
+ "B-CITY": 15,
140
+ "B-COMPANY_NAME": 17,
141
+ "B-COORDINATE": 19,
142
+ "B-COUNTRY": 21,
143
+ "B-COUNTY": 23,
144
+ "B-CREDIT_DEBIT_CARD": 25,
145
+ "B-CUSTOMER_ID": 27,
146
+ "B-CVV": 29,
147
+ "B-DATE": 31,
148
+ "B-DATE_OF_BIRTH": 33,
149
+ "B-DATE_TIME": 35,
150
+ "B-DEVICE_IDENTIFIER": 37,
151
+ "B-EDUCATION_LEVEL": 39,
152
+ "B-EMAIL": 41,
153
+ "B-EMPLOYEE_ID": 43,
154
+ "B-EMPLOYMENT_STATUS": 45,
155
+ "B-FAX_NUMBER": 47,
156
+ "B-FIRST_NAME": 49,
157
+ "B-GENDER": 51,
158
+ "B-HEALTH_PLAN_BENEFICIARY_NUMBER": 53,
159
+ "B-HTTP_COOKIE": 55,
160
+ "B-IPV4": 57,
161
+ "B-IPV6": 59,
162
+ "B-LANGUAGE": 61,
163
+ "B-LAST_NAME": 63,
164
+ "B-LICENSE_PLATE": 65,
165
+ "B-MAC_ADDRESS": 67,
166
+ "B-MEDICAL_RECORD_NUMBER": 69,
167
+ "B-NATIONAL_ID": 71,
168
+ "B-OCCUPATION": 73,
169
+ "B-PASSWORD": 75,
170
+ "B-PHONE_NUMBER": 77,
171
+ "B-PIN": 79,
172
+ "B-POLITICAL_VIEW": 81,
173
+ "B-POSTCODE": 83,
174
+ "B-RACE_ETHNICITY": 85,
175
+ "B-RELIGIOUS_BELIEF": 87,
176
+ "B-SEXUALITY": 89,
177
+ "B-SSN": 91,
178
+ "B-STATE": 93,
179
+ "B-STREET_ADDRESS": 95,
180
+ "B-SWIFT_BIC": 97,
181
+ "B-TAX_ID": 99,
182
+ "B-TIME": 101,
183
+ "B-UNIQUE_ID": 103,
184
+ "B-URL": 105,
185
+ "B-USER_NAME": 107,
186
+ "B-VEHICLE_IDENTIFIER": 109,
187
+ "I-ACCOUNT_NUMBER": 2,
188
+ "I-AGE": 4,
189
+ "I-API_KEY": 6,
190
+ "I-BANK_ROUTING_NUMBER": 8,
191
+ "I-BIOMETRIC_IDENTIFIER": 10,
192
+ "I-BLOOD_TYPE": 12,
193
+ "I-CERTIFICATE_LICENSE_NUMBER": 14,
194
+ "I-CITY": 16,
195
+ "I-COMPANY_NAME": 18,
196
+ "I-COORDINATE": 20,
197
+ "I-COUNTRY": 22,
198
+ "I-COUNTY": 24,
199
+ "I-CREDIT_DEBIT_CARD": 26,
200
+ "I-CUSTOMER_ID": 28,
201
+ "I-CVV": 30,
202
+ "I-DATE": 32,
203
+ "I-DATE_OF_BIRTH": 34,
204
+ "I-DATE_TIME": 36,
205
+ "I-DEVICE_IDENTIFIER": 38,
206
+ "I-EDUCATION_LEVEL": 40,
207
+ "I-EMAIL": 42,
208
+ "I-EMPLOYEE_ID": 44,
209
+ "I-EMPLOYMENT_STATUS": 46,
210
+ "I-FAX_NUMBER": 48,
211
+ "I-FIRST_NAME": 50,
212
+ "I-GENDER": 52,
213
+ "I-HEALTH_PLAN_BENEFICIARY_NUMBER": 54,
214
+ "I-HTTP_COOKIE": 56,
215
+ "I-IPV4": 58,
216
+ "I-IPV6": 60,
217
+ "I-LANGUAGE": 62,
218
+ "I-LAST_NAME": 64,
219
+ "I-LICENSE_PLATE": 66,
220
+ "I-MAC_ADDRESS": 68,
221
+ "I-MEDICAL_RECORD_NUMBER": 70,
222
+ "I-NATIONAL_ID": 72,
223
+ "I-OCCUPATION": 74,
224
+ "I-PASSWORD": 76,
225
+ "I-PHONE_NUMBER": 78,
226
+ "I-PIN": 80,
227
+ "I-POLITICAL_VIEW": 82,
228
+ "I-POSTCODE": 84,
229
+ "I-RACE_ETHNICITY": 86,
230
+ "I-RELIGIOUS_BELIEF": 88,
231
+ "I-SEXUALITY": 90,
232
+ "I-SSN": 92,
233
+ "I-STATE": 94,
234
+ "I-STREET_ADDRESS": 96,
235
+ "I-SWIFT_BIC": 98,
236
+ "I-TAX_ID": 100,
237
+ "I-TIME": 102,
238
+ "I-UNIQUE_ID": 104,
239
+ "I-URL": 106,
240
+ "I-USER_NAME": 108,
241
+ "I-VEHICLE_IDENTIFIER": 110,
242
+ "O": 0
243
+ },
244
+ "layer_norm_eps": 1e-12,
245
+ "max_position_embeddings": 512,
246
+ "model_type": "bert",
247
+ "num_attention_heads": 12,
248
+ "num_hidden_layers": 12,
249
+ "pad_token_id": 0,
250
+ "pooler_fc_size": 768,
251
+ "pooler_num_attention_heads": 12,
252
+ "pooler_num_fc_layers": 3,
253
+ "pooler_size_per_head": 128,
254
+ "pooler_type": "first_token_transform",
255
+ "tie_word_embeddings": true,
256
+ "transformers_version": "5.3.0",
257
+ "type_vocab_size": 2,
258
+ "use_cache": false,
259
+ "vocab_size": 119547
260
+ }
nemotron-pii-ready/TokenBased-BERT/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:260ff5ff37bb85d1af3c0b1d780ea4a36d7ac2b2009a7087e6451c18ba7e0741
3
+ size 709416164
nemotron-pii-ready/TokenBased-BERT/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
nemotron-pii-ready/TokenBased-BERT/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-BERT/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54258d8cd6ea85d5af9ffa138c92bf8a8c9ec44c6cb90d9f1043a2f004d015d6
3
+ size 5201