PuxAI commited on
Commit
be2ece8
·
verified ·
1 Parent(s): 9fc1f1d

Upload folder using huggingface_hub

Browse files
Files changed (32) hide show
  1. gretel-pii-ready/TokenBased-BERT/checkpoint-1563/config.json +208 -0
  2. gretel-pii-ready/TokenBased-BERT/checkpoint-1563/model.safetensors +3 -0
  3. gretel-pii-ready/TokenBased-BERT/checkpoint-1563/optimizer.pt +3 -0
  4. gretel-pii-ready/TokenBased-BERT/checkpoint-1563/rng_state.pth +3 -0
  5. gretel-pii-ready/TokenBased-BERT/checkpoint-1563/scheduler.pt +3 -0
  6. gretel-pii-ready/TokenBased-BERT/checkpoint-1563/tokenizer.json +0 -0
  7. gretel-pii-ready/TokenBased-BERT/checkpoint-1563/tokenizer_config.json +15 -0
  8. gretel-pii-ready/TokenBased-BERT/checkpoint-1563/trainer_state.json +63 -0
  9. gretel-pii-ready/TokenBased-BERT/checkpoint-1563/training_args.bin +3 -0
  10. gretel-pii-ready/TokenBased-BERT/checkpoint-3126/config.json +208 -0
  11. gretel-pii-ready/TokenBased-BERT/checkpoint-3126/model.safetensors +3 -0
  12. gretel-pii-ready/TokenBased-BERT/checkpoint-3126/optimizer.pt +3 -0
  13. gretel-pii-ready/TokenBased-BERT/checkpoint-3126/rng_state.pth +3 -0
  14. gretel-pii-ready/TokenBased-BERT/checkpoint-3126/scheduler.pt +3 -0
  15. gretel-pii-ready/TokenBased-BERT/checkpoint-3126/tokenizer.json +0 -0
  16. gretel-pii-ready/TokenBased-BERT/checkpoint-3126/tokenizer_config.json +15 -0
  17. gretel-pii-ready/TokenBased-BERT/checkpoint-3126/trainer_state.json +92 -0
  18. gretel-pii-ready/TokenBased-BERT/checkpoint-3126/training_args.bin +3 -0
  19. gretel-pii-ready/TokenBased-BERT/checkpoint-4689/config.json +208 -0
  20. gretel-pii-ready/TokenBased-BERT/checkpoint-4689/model.safetensors +3 -0
  21. gretel-pii-ready/TokenBased-BERT/checkpoint-4689/optimizer.pt +3 -0
  22. gretel-pii-ready/TokenBased-BERT/checkpoint-4689/rng_state.pth +3 -0
  23. gretel-pii-ready/TokenBased-BERT/checkpoint-4689/scheduler.pt +3 -0
  24. gretel-pii-ready/TokenBased-BERT/checkpoint-4689/tokenizer.json +0 -0
  25. gretel-pii-ready/TokenBased-BERT/checkpoint-4689/tokenizer_config.json +15 -0
  26. gretel-pii-ready/TokenBased-BERT/checkpoint-4689/trainer_state.json +121 -0
  27. gretel-pii-ready/TokenBased-BERT/checkpoint-4689/training_args.bin +3 -0
  28. gretel-pii-ready/TokenBased-BERT/config.json +208 -0
  29. gretel-pii-ready/TokenBased-BERT/model.safetensors +3 -0
  30. gretel-pii-ready/TokenBased-BERT/tokenizer.json +0 -0
  31. gretel-pii-ready/TokenBased-BERT/tokenizer_config.json +15 -0
  32. gretel-pii-ready/TokenBased-BERT/training_args.bin +3 -0
gretel-pii-ready/TokenBased-BERT/checkpoint-1563/config.json ADDED
@@ -0,0 +1,208 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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-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-PIN",
78
+ "62": "I-PIN",
79
+ "63": "B-POSTCODE",
80
+ "64": "I-POSTCODE",
81
+ "65": "B-SSN",
82
+ "66": "I-SSN",
83
+ "67": "B-STATE",
84
+ "68": "I-STATE",
85
+ "69": "B-STREET_ADDRESS",
86
+ "70": "I-STREET_ADDRESS",
87
+ "71": "B-SWIFT_BIC",
88
+ "72": "I-SWIFT_BIC",
89
+ "73": "B-TAX_ID",
90
+ "74": "I-TAX_ID",
91
+ "75": "B-TIME",
92
+ "76": "I-TIME",
93
+ "77": "B-UNIQUE_IDENTIFIER",
94
+ "78": "I-UNIQUE_IDENTIFIER",
95
+ "79": "B-URL",
96
+ "80": "I-URL",
97
+ "81": "B-USER_NAME",
98
+ "82": "I-USER_NAME",
99
+ "83": "B-VEHICLE_IDENTIFIER",
100
+ "84": "I-VEHICLE_IDENTIFIER"
101
+ },
102
+ "initializer_range": 0.02,
103
+ "intermediate_size": 3072,
104
+ "is_decoder": false,
105
+ "label2id": {
106
+ "B-ACCOUNT_NUMBER": 1,
107
+ "B-ADDRESS": 3,
108
+ "B-API_KEY": 5,
109
+ "B-BANK_ROUTING_NUMBER": 7,
110
+ "B-BIOMETRIC_IDENTIFIER": 9,
111
+ "B-CERTIFICATE_LICENSE_NUMBER": 11,
112
+ "B-CITY": 13,
113
+ "B-COMPANY_NAME": 15,
114
+ "B-COORDINATE": 17,
115
+ "B-COUNTRY": 19,
116
+ "B-CREDIT_CARD_NUMBER": 21,
117
+ "B-CUSTOMER_ID": 23,
118
+ "B-CVV": 25,
119
+ "B-DATE": 27,
120
+ "B-DATE_OF_BIRTH": 29,
121
+ "B-DATE_TIME": 31,
122
+ "B-DEVICE_IDENTIFIER": 33,
123
+ "B-EMAIL": 35,
124
+ "B-EMPLOYEE_ID": 37,
125
+ "B-FIRST_NAME": 39,
126
+ "B-HEALTH_PLAN_BENEFICIARY_NUMBER": 41,
127
+ "B-IPV4": 43,
128
+ "B-IPV6": 45,
129
+ "B-LAST_NAME": 47,
130
+ "B-LICENSE_PLATE": 49,
131
+ "B-MEDICAL_RECORD_NUMBER": 51,
132
+ "B-NAME": 53,
133
+ "B-NATIONAL_ID": 55,
134
+ "B-PASSWORD": 57,
135
+ "B-PHONE_NUMBER": 59,
136
+ "B-PIN": 61,
137
+ "B-POSTCODE": 63,
138
+ "B-SSN": 65,
139
+ "B-STATE": 67,
140
+ "B-STREET_ADDRESS": 69,
141
+ "B-SWIFT_BIC": 71,
142
+ "B-TAX_ID": 73,
143
+ "B-TIME": 75,
144
+ "B-UNIQUE_IDENTIFIER": 77,
145
+ "B-URL": 79,
146
+ "B-USER_NAME": 81,
147
+ "B-VEHICLE_IDENTIFIER": 83,
148
+ "I-ACCOUNT_NUMBER": 2,
149
+ "I-ADDRESS": 4,
150
+ "I-API_KEY": 6,
151
+ "I-BANK_ROUTING_NUMBER": 8,
152
+ "I-BIOMETRIC_IDENTIFIER": 10,
153
+ "I-CERTIFICATE_LICENSE_NUMBER": 12,
154
+ "I-CITY": 14,
155
+ "I-COMPANY_NAME": 16,
156
+ "I-COORDINATE": 18,
157
+ "I-COUNTRY": 20,
158
+ "I-CREDIT_CARD_NUMBER": 22,
159
+ "I-CUSTOMER_ID": 24,
160
+ "I-CVV": 26,
161
+ "I-DATE": 28,
162
+ "I-DATE_OF_BIRTH": 30,
163
+ "I-DATE_TIME": 32,
164
+ "I-DEVICE_IDENTIFIER": 34,
165
+ "I-EMAIL": 36,
166
+ "I-EMPLOYEE_ID": 38,
167
+ "I-FIRST_NAME": 40,
168
+ "I-HEALTH_PLAN_BENEFICIARY_NUMBER": 42,
169
+ "I-IPV4": 44,
170
+ "I-IPV6": 46,
171
+ "I-LAST_NAME": 48,
172
+ "I-LICENSE_PLATE": 50,
173
+ "I-MEDICAL_RECORD_NUMBER": 52,
174
+ "I-NAME": 54,
175
+ "I-NATIONAL_ID": 56,
176
+ "I-PASSWORD": 58,
177
+ "I-PHONE_NUMBER": 60,
178
+ "I-PIN": 62,
179
+ "I-POSTCODE": 64,
180
+ "I-SSN": 66,
181
+ "I-STATE": 68,
182
+ "I-STREET_ADDRESS": 70,
183
+ "I-SWIFT_BIC": 72,
184
+ "I-TAX_ID": 74,
185
+ "I-TIME": 76,
186
+ "I-UNIQUE_IDENTIFIER": 78,
187
+ "I-URL": 80,
188
+ "I-USER_NAME": 82,
189
+ "I-VEHICLE_IDENTIFIER": 84,
190
+ "O": 0
191
+ },
192
+ "layer_norm_eps": 1e-12,
193
+ "max_position_embeddings": 512,
194
+ "model_type": "bert",
195
+ "num_attention_heads": 12,
196
+ "num_hidden_layers": 12,
197
+ "pad_token_id": 0,
198
+ "pooler_fc_size": 768,
199
+ "pooler_num_attention_heads": 12,
200
+ "pooler_num_fc_layers": 3,
201
+ "pooler_size_per_head": 128,
202
+ "pooler_type": "first_token_transform",
203
+ "tie_word_embeddings": true,
204
+ "transformers_version": "5.3.0",
205
+ "type_vocab_size": 2,
206
+ "use_cache": false,
207
+ "vocab_size": 119547
208
+ }
gretel-pii-ready/TokenBased-BERT/checkpoint-1563/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5eb6cc67644321b90b9b3b15b5f9567db003237b4bba880133c316fe5ef97a8
3
+ size 709336188
gretel-pii-ready/TokenBased-BERT/checkpoint-1563/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e5239a729425598f4b63e82d2393265e0240ca5b6fbbb514f2afedf022cf6d6f
3
+ size 1418795403
gretel-pii-ready/TokenBased-BERT/checkpoint-1563/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37dde603be145693ef20b356f678f5c2b3cad8efbcde6e919506fa230211e4ea
3
+ size 14645
gretel-pii-ready/TokenBased-BERT/checkpoint-1563/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:396e3d88a15c6e695b0f09eb00f8183d6f7541b046e9d4bc11dee4e9a850183c
3
+ size 1465
gretel-pii-ready/TokenBased-BERT/checkpoint-1563/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
gretel-pii-ready/TokenBased-BERT/checkpoint-1563/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-BERT/checkpoint-1563/trainer_state.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 1563,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.3198976327575176,
14
+ "grad_norm": 1.666126012802124,
15
+ "learning_rate": 1.7871614416719985e-05,
16
+ "loss": 0.36542596435546876,
17
+ "step": 500
18
+ },
19
+ {
20
+ "epoch": 0.6397952655150352,
21
+ "grad_norm": 1.2852263450622559,
22
+ "learning_rate": 1.5738963531669867e-05,
23
+ "loss": 0.07889308929443359,
24
+ "step": 1000
25
+ },
26
+ {
27
+ "epoch": 0.9596928982725528,
28
+ "grad_norm": 1.8407200574874878,
29
+ "learning_rate": 1.3606312646619751e-05,
30
+ "loss": 0.05602884674072266,
31
+ "step": 1500
32
+ },
33
+ {
34
+ "epoch": 1.0,
35
+ "eval_loss": 0.04717080667614937,
36
+ "eval_runtime": 16.2297,
37
+ "eval_samples_per_second": 308.078,
38
+ "eval_steps_per_second": 38.51,
39
+ "step": 1563
40
+ }
41
+ ],
42
+ "logging_steps": 500,
43
+ "max_steps": 4689,
44
+ "num_input_tokens_seen": 0,
45
+ "num_train_epochs": 3,
46
+ "save_steps": 500,
47
+ "stateful_callbacks": {
48
+ "TrainerControl": {
49
+ "args": {
50
+ "should_epoch_stop": false,
51
+ "should_evaluate": false,
52
+ "should_log": false,
53
+ "should_save": true,
54
+ "should_training_stop": false
55
+ },
56
+ "attributes": {}
57
+ }
58
+ },
59
+ "total_flos": 1.3074641664e+16,
60
+ "train_batch_size": 32,
61
+ "trial_name": null,
62
+ "trial_params": null
63
+ }
gretel-pii-ready/TokenBased-BERT/checkpoint-1563/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:94312732eff9efbc5cfb93e3ca98a015e5429d10739537c5b4cb08702f1fb1d1
3
+ size 5201
gretel-pii-ready/TokenBased-BERT/checkpoint-3126/config.json ADDED
@@ -0,0 +1,208 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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-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-PIN",
78
+ "62": "I-PIN",
79
+ "63": "B-POSTCODE",
80
+ "64": "I-POSTCODE",
81
+ "65": "B-SSN",
82
+ "66": "I-SSN",
83
+ "67": "B-STATE",
84
+ "68": "I-STATE",
85
+ "69": "B-STREET_ADDRESS",
86
+ "70": "I-STREET_ADDRESS",
87
+ "71": "B-SWIFT_BIC",
88
+ "72": "I-SWIFT_BIC",
89
+ "73": "B-TAX_ID",
90
+ "74": "I-TAX_ID",
91
+ "75": "B-TIME",
92
+ "76": "I-TIME",
93
+ "77": "B-UNIQUE_IDENTIFIER",
94
+ "78": "I-UNIQUE_IDENTIFIER",
95
+ "79": "B-URL",
96
+ "80": "I-URL",
97
+ "81": "B-USER_NAME",
98
+ "82": "I-USER_NAME",
99
+ "83": "B-VEHICLE_IDENTIFIER",
100
+ "84": "I-VEHICLE_IDENTIFIER"
101
+ },
102
+ "initializer_range": 0.02,
103
+ "intermediate_size": 3072,
104
+ "is_decoder": false,
105
+ "label2id": {
106
+ "B-ACCOUNT_NUMBER": 1,
107
+ "B-ADDRESS": 3,
108
+ "B-API_KEY": 5,
109
+ "B-BANK_ROUTING_NUMBER": 7,
110
+ "B-BIOMETRIC_IDENTIFIER": 9,
111
+ "B-CERTIFICATE_LICENSE_NUMBER": 11,
112
+ "B-CITY": 13,
113
+ "B-COMPANY_NAME": 15,
114
+ "B-COORDINATE": 17,
115
+ "B-COUNTRY": 19,
116
+ "B-CREDIT_CARD_NUMBER": 21,
117
+ "B-CUSTOMER_ID": 23,
118
+ "B-CVV": 25,
119
+ "B-DATE": 27,
120
+ "B-DATE_OF_BIRTH": 29,
121
+ "B-DATE_TIME": 31,
122
+ "B-DEVICE_IDENTIFIER": 33,
123
+ "B-EMAIL": 35,
124
+ "B-EMPLOYEE_ID": 37,
125
+ "B-FIRST_NAME": 39,
126
+ "B-HEALTH_PLAN_BENEFICIARY_NUMBER": 41,
127
+ "B-IPV4": 43,
128
+ "B-IPV6": 45,
129
+ "B-LAST_NAME": 47,
130
+ "B-LICENSE_PLATE": 49,
131
+ "B-MEDICAL_RECORD_NUMBER": 51,
132
+ "B-NAME": 53,
133
+ "B-NATIONAL_ID": 55,
134
+ "B-PASSWORD": 57,
135
+ "B-PHONE_NUMBER": 59,
136
+ "B-PIN": 61,
137
+ "B-POSTCODE": 63,
138
+ "B-SSN": 65,
139
+ "B-STATE": 67,
140
+ "B-STREET_ADDRESS": 69,
141
+ "B-SWIFT_BIC": 71,
142
+ "B-TAX_ID": 73,
143
+ "B-TIME": 75,
144
+ "B-UNIQUE_IDENTIFIER": 77,
145
+ "B-URL": 79,
146
+ "B-USER_NAME": 81,
147
+ "B-VEHICLE_IDENTIFIER": 83,
148
+ "I-ACCOUNT_NUMBER": 2,
149
+ "I-ADDRESS": 4,
150
+ "I-API_KEY": 6,
151
+ "I-BANK_ROUTING_NUMBER": 8,
152
+ "I-BIOMETRIC_IDENTIFIER": 10,
153
+ "I-CERTIFICATE_LICENSE_NUMBER": 12,
154
+ "I-CITY": 14,
155
+ "I-COMPANY_NAME": 16,
156
+ "I-COORDINATE": 18,
157
+ "I-COUNTRY": 20,
158
+ "I-CREDIT_CARD_NUMBER": 22,
159
+ "I-CUSTOMER_ID": 24,
160
+ "I-CVV": 26,
161
+ "I-DATE": 28,
162
+ "I-DATE_OF_BIRTH": 30,
163
+ "I-DATE_TIME": 32,
164
+ "I-DEVICE_IDENTIFIER": 34,
165
+ "I-EMAIL": 36,
166
+ "I-EMPLOYEE_ID": 38,
167
+ "I-FIRST_NAME": 40,
168
+ "I-HEALTH_PLAN_BENEFICIARY_NUMBER": 42,
169
+ "I-IPV4": 44,
170
+ "I-IPV6": 46,
171
+ "I-LAST_NAME": 48,
172
+ "I-LICENSE_PLATE": 50,
173
+ "I-MEDICAL_RECORD_NUMBER": 52,
174
+ "I-NAME": 54,
175
+ "I-NATIONAL_ID": 56,
176
+ "I-PASSWORD": 58,
177
+ "I-PHONE_NUMBER": 60,
178
+ "I-PIN": 62,
179
+ "I-POSTCODE": 64,
180
+ "I-SSN": 66,
181
+ "I-STATE": 68,
182
+ "I-STREET_ADDRESS": 70,
183
+ "I-SWIFT_BIC": 72,
184
+ "I-TAX_ID": 74,
185
+ "I-TIME": 76,
186
+ "I-UNIQUE_IDENTIFIER": 78,
187
+ "I-URL": 80,
188
+ "I-USER_NAME": 82,
189
+ "I-VEHICLE_IDENTIFIER": 84,
190
+ "O": 0
191
+ },
192
+ "layer_norm_eps": 1e-12,
193
+ "max_position_embeddings": 512,
194
+ "model_type": "bert",
195
+ "num_attention_heads": 12,
196
+ "num_hidden_layers": 12,
197
+ "pad_token_id": 0,
198
+ "pooler_fc_size": 768,
199
+ "pooler_num_attention_heads": 12,
200
+ "pooler_num_fc_layers": 3,
201
+ "pooler_size_per_head": 128,
202
+ "pooler_type": "first_token_transform",
203
+ "tie_word_embeddings": true,
204
+ "transformers_version": "5.3.0",
205
+ "type_vocab_size": 2,
206
+ "use_cache": false,
207
+ "vocab_size": 119547
208
+ }
gretel-pii-ready/TokenBased-BERT/checkpoint-3126/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8d4964c570ed64b561938c2b33ba2dcc2f27381169398e86f6cf7bfbd85657ca
3
+ size 709336188
gretel-pii-ready/TokenBased-BERT/checkpoint-3126/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a73c22538b745612b213bd3eefa4c92242310d08f4aeec4a3d8b527754fdd3e
3
+ size 1418795403
gretel-pii-ready/TokenBased-BERT/checkpoint-3126/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e1a5eb3e39f334c88b83ede23c7f803996072776f3bac8fcb49eebacdd10731
3
+ size 14645
gretel-pii-ready/TokenBased-BERT/checkpoint-3126/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad052e737f9e85d34b148a4255a9f41b3f2abf70b1c4994f6c860a9741e229de
3
+ size 1465
gretel-pii-ready/TokenBased-BERT/checkpoint-3126/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
gretel-pii-ready/TokenBased-BERT/checkpoint-3126/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-BERT/checkpoint-3126/trainer_state.json ADDED
@@ -0,0 +1,92 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 3126,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.3198976327575176,
14
+ "grad_norm": 1.666126012802124,
15
+ "learning_rate": 1.7871614416719985e-05,
16
+ "loss": 0.36542596435546876,
17
+ "step": 500
18
+ },
19
+ {
20
+ "epoch": 0.6397952655150352,
21
+ "grad_norm": 1.2852263450622559,
22
+ "learning_rate": 1.5738963531669867e-05,
23
+ "loss": 0.07889308929443359,
24
+ "step": 1000
25
+ },
26
+ {
27
+ "epoch": 0.9596928982725528,
28
+ "grad_norm": 1.8407200574874878,
29
+ "learning_rate": 1.3606312646619751e-05,
30
+ "loss": 0.05602884674072266,
31
+ "step": 1500
32
+ },
33
+ {
34
+ "epoch": 1.0,
35
+ "eval_loss": 0.04717080667614937,
36
+ "eval_runtime": 16.2297,
37
+ "eval_samples_per_second": 308.078,
38
+ "eval_steps_per_second": 38.51,
39
+ "step": 1563
40
+ },
41
+ {
42
+ "epoch": 1.2795905310300704,
43
+ "grad_norm": 0.979367733001709,
44
+ "learning_rate": 1.1473661761569632e-05,
45
+ "loss": 0.03934717559814453,
46
+ "step": 2000
47
+ },
48
+ {
49
+ "epoch": 1.599488163787588,
50
+ "grad_norm": 0.3883236050605774,
51
+ "learning_rate": 9.341010876519514e-06,
52
+ "loss": 0.036172042846679686,
53
+ "step": 2500
54
+ },
55
+ {
56
+ "epoch": 1.9193857965451055,
57
+ "grad_norm": 1.526019811630249,
58
+ "learning_rate": 7.2083599914693966e-06,
59
+ "loss": 0.03161228942871094,
60
+ "step": 3000
61
+ },
62
+ {
63
+ "epoch": 2.0,
64
+ "eval_loss": 0.03497769311070442,
65
+ "eval_runtime": 16.1263,
66
+ "eval_samples_per_second": 310.052,
67
+ "eval_steps_per_second": 38.757,
68
+ "step": 3126
69
+ }
70
+ ],
71
+ "logging_steps": 500,
72
+ "max_steps": 4689,
73
+ "num_input_tokens_seen": 0,
74
+ "num_train_epochs": 3,
75
+ "save_steps": 500,
76
+ "stateful_callbacks": {
77
+ "TrainerControl": {
78
+ "args": {
79
+ "should_epoch_stop": false,
80
+ "should_evaluate": false,
81
+ "should_log": false,
82
+ "should_save": true,
83
+ "should_training_stop": false
84
+ },
85
+ "attributes": {}
86
+ }
87
+ },
88
+ "total_flos": 2.6149283328e+16,
89
+ "train_batch_size": 32,
90
+ "trial_name": null,
91
+ "trial_params": null
92
+ }
gretel-pii-ready/TokenBased-BERT/checkpoint-3126/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:94312732eff9efbc5cfb93e3ca98a015e5429d10739537c5b4cb08702f1fb1d1
3
+ size 5201
gretel-pii-ready/TokenBased-BERT/checkpoint-4689/config.json ADDED
@@ -0,0 +1,208 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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-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-PIN",
78
+ "62": "I-PIN",
79
+ "63": "B-POSTCODE",
80
+ "64": "I-POSTCODE",
81
+ "65": "B-SSN",
82
+ "66": "I-SSN",
83
+ "67": "B-STATE",
84
+ "68": "I-STATE",
85
+ "69": "B-STREET_ADDRESS",
86
+ "70": "I-STREET_ADDRESS",
87
+ "71": "B-SWIFT_BIC",
88
+ "72": "I-SWIFT_BIC",
89
+ "73": "B-TAX_ID",
90
+ "74": "I-TAX_ID",
91
+ "75": "B-TIME",
92
+ "76": "I-TIME",
93
+ "77": "B-UNIQUE_IDENTIFIER",
94
+ "78": "I-UNIQUE_IDENTIFIER",
95
+ "79": "B-URL",
96
+ "80": "I-URL",
97
+ "81": "B-USER_NAME",
98
+ "82": "I-USER_NAME",
99
+ "83": "B-VEHICLE_IDENTIFIER",
100
+ "84": "I-VEHICLE_IDENTIFIER"
101
+ },
102
+ "initializer_range": 0.02,
103
+ "intermediate_size": 3072,
104
+ "is_decoder": false,
105
+ "label2id": {
106
+ "B-ACCOUNT_NUMBER": 1,
107
+ "B-ADDRESS": 3,
108
+ "B-API_KEY": 5,
109
+ "B-BANK_ROUTING_NUMBER": 7,
110
+ "B-BIOMETRIC_IDENTIFIER": 9,
111
+ "B-CERTIFICATE_LICENSE_NUMBER": 11,
112
+ "B-CITY": 13,
113
+ "B-COMPANY_NAME": 15,
114
+ "B-COORDINATE": 17,
115
+ "B-COUNTRY": 19,
116
+ "B-CREDIT_CARD_NUMBER": 21,
117
+ "B-CUSTOMER_ID": 23,
118
+ "B-CVV": 25,
119
+ "B-DATE": 27,
120
+ "B-DATE_OF_BIRTH": 29,
121
+ "B-DATE_TIME": 31,
122
+ "B-DEVICE_IDENTIFIER": 33,
123
+ "B-EMAIL": 35,
124
+ "B-EMPLOYEE_ID": 37,
125
+ "B-FIRST_NAME": 39,
126
+ "B-HEALTH_PLAN_BENEFICIARY_NUMBER": 41,
127
+ "B-IPV4": 43,
128
+ "B-IPV6": 45,
129
+ "B-LAST_NAME": 47,
130
+ "B-LICENSE_PLATE": 49,
131
+ "B-MEDICAL_RECORD_NUMBER": 51,
132
+ "B-NAME": 53,
133
+ "B-NATIONAL_ID": 55,
134
+ "B-PASSWORD": 57,
135
+ "B-PHONE_NUMBER": 59,
136
+ "B-PIN": 61,
137
+ "B-POSTCODE": 63,
138
+ "B-SSN": 65,
139
+ "B-STATE": 67,
140
+ "B-STREET_ADDRESS": 69,
141
+ "B-SWIFT_BIC": 71,
142
+ "B-TAX_ID": 73,
143
+ "B-TIME": 75,
144
+ "B-UNIQUE_IDENTIFIER": 77,
145
+ "B-URL": 79,
146
+ "B-USER_NAME": 81,
147
+ "B-VEHICLE_IDENTIFIER": 83,
148
+ "I-ACCOUNT_NUMBER": 2,
149
+ "I-ADDRESS": 4,
150
+ "I-API_KEY": 6,
151
+ "I-BANK_ROUTING_NUMBER": 8,
152
+ "I-BIOMETRIC_IDENTIFIER": 10,
153
+ "I-CERTIFICATE_LICENSE_NUMBER": 12,
154
+ "I-CITY": 14,
155
+ "I-COMPANY_NAME": 16,
156
+ "I-COORDINATE": 18,
157
+ "I-COUNTRY": 20,
158
+ "I-CREDIT_CARD_NUMBER": 22,
159
+ "I-CUSTOMER_ID": 24,
160
+ "I-CVV": 26,
161
+ "I-DATE": 28,
162
+ "I-DATE_OF_BIRTH": 30,
163
+ "I-DATE_TIME": 32,
164
+ "I-DEVICE_IDENTIFIER": 34,
165
+ "I-EMAIL": 36,
166
+ "I-EMPLOYEE_ID": 38,
167
+ "I-FIRST_NAME": 40,
168
+ "I-HEALTH_PLAN_BENEFICIARY_NUMBER": 42,
169
+ "I-IPV4": 44,
170
+ "I-IPV6": 46,
171
+ "I-LAST_NAME": 48,
172
+ "I-LICENSE_PLATE": 50,
173
+ "I-MEDICAL_RECORD_NUMBER": 52,
174
+ "I-NAME": 54,
175
+ "I-NATIONAL_ID": 56,
176
+ "I-PASSWORD": 58,
177
+ "I-PHONE_NUMBER": 60,
178
+ "I-PIN": 62,
179
+ "I-POSTCODE": 64,
180
+ "I-SSN": 66,
181
+ "I-STATE": 68,
182
+ "I-STREET_ADDRESS": 70,
183
+ "I-SWIFT_BIC": 72,
184
+ "I-TAX_ID": 74,
185
+ "I-TIME": 76,
186
+ "I-UNIQUE_IDENTIFIER": 78,
187
+ "I-URL": 80,
188
+ "I-USER_NAME": 82,
189
+ "I-VEHICLE_IDENTIFIER": 84,
190
+ "O": 0
191
+ },
192
+ "layer_norm_eps": 1e-12,
193
+ "max_position_embeddings": 512,
194
+ "model_type": "bert",
195
+ "num_attention_heads": 12,
196
+ "num_hidden_layers": 12,
197
+ "pad_token_id": 0,
198
+ "pooler_fc_size": 768,
199
+ "pooler_num_attention_heads": 12,
200
+ "pooler_num_fc_layers": 3,
201
+ "pooler_size_per_head": 128,
202
+ "pooler_type": "first_token_transform",
203
+ "tie_word_embeddings": true,
204
+ "transformers_version": "5.3.0",
205
+ "type_vocab_size": 2,
206
+ "use_cache": false,
207
+ "vocab_size": 119547
208
+ }
gretel-pii-ready/TokenBased-BERT/checkpoint-4689/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c8ef31b04e7e318825a513a0949fd8f0e66385605578071f555b5829541e28c1
3
+ size 709336188
gretel-pii-ready/TokenBased-BERT/checkpoint-4689/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62da5ebb81ef8c721772ca4a060d33750fb9829952fe38989e3135eb9df97b2c
3
+ size 1418795403
gretel-pii-ready/TokenBased-BERT/checkpoint-4689/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8991c053d100dc794aa6b41c4778ae17e6c395dd3206dd4a646df5e355505150
3
+ size 14645
gretel-pii-ready/TokenBased-BERT/checkpoint-4689/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:819022c2704a2c9667807f7d5537d7d4a91c40d504ba8627512c8c96aaf50c01
3
+ size 1465
gretel-pii-ready/TokenBased-BERT/checkpoint-4689/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
gretel-pii-ready/TokenBased-BERT/checkpoint-4689/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-BERT/checkpoint-4689/trainer_state.json ADDED
@@ -0,0 +1,121 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 4689,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.3198976327575176,
14
+ "grad_norm": 1.666126012802124,
15
+ "learning_rate": 1.7871614416719985e-05,
16
+ "loss": 0.36542596435546876,
17
+ "step": 500
18
+ },
19
+ {
20
+ "epoch": 0.6397952655150352,
21
+ "grad_norm": 1.2852263450622559,
22
+ "learning_rate": 1.5738963531669867e-05,
23
+ "loss": 0.07889308929443359,
24
+ "step": 1000
25
+ },
26
+ {
27
+ "epoch": 0.9596928982725528,
28
+ "grad_norm": 1.8407200574874878,
29
+ "learning_rate": 1.3606312646619751e-05,
30
+ "loss": 0.05602884674072266,
31
+ "step": 1500
32
+ },
33
+ {
34
+ "epoch": 1.0,
35
+ "eval_loss": 0.04717080667614937,
36
+ "eval_runtime": 16.2297,
37
+ "eval_samples_per_second": 308.078,
38
+ "eval_steps_per_second": 38.51,
39
+ "step": 1563
40
+ },
41
+ {
42
+ "epoch": 1.2795905310300704,
43
+ "grad_norm": 0.979367733001709,
44
+ "learning_rate": 1.1473661761569632e-05,
45
+ "loss": 0.03934717559814453,
46
+ "step": 2000
47
+ },
48
+ {
49
+ "epoch": 1.599488163787588,
50
+ "grad_norm": 0.3883236050605774,
51
+ "learning_rate": 9.341010876519514e-06,
52
+ "loss": 0.036172042846679686,
53
+ "step": 2500
54
+ },
55
+ {
56
+ "epoch": 1.9193857965451055,
57
+ "grad_norm": 1.526019811630249,
58
+ "learning_rate": 7.2083599914693966e-06,
59
+ "loss": 0.03161228942871094,
60
+ "step": 3000
61
+ },
62
+ {
63
+ "epoch": 2.0,
64
+ "eval_loss": 0.03497769311070442,
65
+ "eval_runtime": 16.1263,
66
+ "eval_samples_per_second": 310.052,
67
+ "eval_steps_per_second": 38.757,
68
+ "step": 3126
69
+ },
70
+ {
71
+ "epoch": 2.239283429302623,
72
+ "grad_norm": 1.1337404251098633,
73
+ "learning_rate": 5.07570910641928e-06,
74
+ "loss": 0.02470443534851074,
75
+ "step": 3500
76
+ },
77
+ {
78
+ "epoch": 2.5591810620601407,
79
+ "grad_norm": 1.1263959407806396,
80
+ "learning_rate": 2.943058221369162e-06,
81
+ "loss": 0.02315133285522461,
82
+ "step": 4000
83
+ },
84
+ {
85
+ "epoch": 2.8790786948176583,
86
+ "grad_norm": 1.0614573955535889,
87
+ "learning_rate": 8.104073363190447e-07,
88
+ "loss": 0.021389280319213867,
89
+ "step": 4500
90
+ },
91
+ {
92
+ "epoch": 3.0,
93
+ "eval_loss": 0.03137294203042984,
94
+ "eval_runtime": 16.6967,
95
+ "eval_samples_per_second": 299.461,
96
+ "eval_steps_per_second": 37.433,
97
+ "step": 4689
98
+ }
99
+ ],
100
+ "logging_steps": 500,
101
+ "max_steps": 4689,
102
+ "num_input_tokens_seen": 0,
103
+ "num_train_epochs": 3,
104
+ "save_steps": 500,
105
+ "stateful_callbacks": {
106
+ "TrainerControl": {
107
+ "args": {
108
+ "should_epoch_stop": false,
109
+ "should_evaluate": false,
110
+ "should_log": false,
111
+ "should_save": true,
112
+ "should_training_stop": true
113
+ },
114
+ "attributes": {}
115
+ }
116
+ },
117
+ "total_flos": 3.9223924992e+16,
118
+ "train_batch_size": 32,
119
+ "trial_name": null,
120
+ "trial_params": null
121
+ }
gretel-pii-ready/TokenBased-BERT/checkpoint-4689/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:94312732eff9efbc5cfb93e3ca98a015e5429d10739537c5b4cb08702f1fb1d1
3
+ size 5201
gretel-pii-ready/TokenBased-BERT/config.json ADDED
@@ -0,0 +1,208 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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-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-PIN",
78
+ "62": "I-PIN",
79
+ "63": "B-POSTCODE",
80
+ "64": "I-POSTCODE",
81
+ "65": "B-SSN",
82
+ "66": "I-SSN",
83
+ "67": "B-STATE",
84
+ "68": "I-STATE",
85
+ "69": "B-STREET_ADDRESS",
86
+ "70": "I-STREET_ADDRESS",
87
+ "71": "B-SWIFT_BIC",
88
+ "72": "I-SWIFT_BIC",
89
+ "73": "B-TAX_ID",
90
+ "74": "I-TAX_ID",
91
+ "75": "B-TIME",
92
+ "76": "I-TIME",
93
+ "77": "B-UNIQUE_IDENTIFIER",
94
+ "78": "I-UNIQUE_IDENTIFIER",
95
+ "79": "B-URL",
96
+ "80": "I-URL",
97
+ "81": "B-USER_NAME",
98
+ "82": "I-USER_NAME",
99
+ "83": "B-VEHICLE_IDENTIFIER",
100
+ "84": "I-VEHICLE_IDENTIFIER"
101
+ },
102
+ "initializer_range": 0.02,
103
+ "intermediate_size": 3072,
104
+ "is_decoder": false,
105
+ "label2id": {
106
+ "B-ACCOUNT_NUMBER": 1,
107
+ "B-ADDRESS": 3,
108
+ "B-API_KEY": 5,
109
+ "B-BANK_ROUTING_NUMBER": 7,
110
+ "B-BIOMETRIC_IDENTIFIER": 9,
111
+ "B-CERTIFICATE_LICENSE_NUMBER": 11,
112
+ "B-CITY": 13,
113
+ "B-COMPANY_NAME": 15,
114
+ "B-COORDINATE": 17,
115
+ "B-COUNTRY": 19,
116
+ "B-CREDIT_CARD_NUMBER": 21,
117
+ "B-CUSTOMER_ID": 23,
118
+ "B-CVV": 25,
119
+ "B-DATE": 27,
120
+ "B-DATE_OF_BIRTH": 29,
121
+ "B-DATE_TIME": 31,
122
+ "B-DEVICE_IDENTIFIER": 33,
123
+ "B-EMAIL": 35,
124
+ "B-EMPLOYEE_ID": 37,
125
+ "B-FIRST_NAME": 39,
126
+ "B-HEALTH_PLAN_BENEFICIARY_NUMBER": 41,
127
+ "B-IPV4": 43,
128
+ "B-IPV6": 45,
129
+ "B-LAST_NAME": 47,
130
+ "B-LICENSE_PLATE": 49,
131
+ "B-MEDICAL_RECORD_NUMBER": 51,
132
+ "B-NAME": 53,
133
+ "B-NATIONAL_ID": 55,
134
+ "B-PASSWORD": 57,
135
+ "B-PHONE_NUMBER": 59,
136
+ "B-PIN": 61,
137
+ "B-POSTCODE": 63,
138
+ "B-SSN": 65,
139
+ "B-STATE": 67,
140
+ "B-STREET_ADDRESS": 69,
141
+ "B-SWIFT_BIC": 71,
142
+ "B-TAX_ID": 73,
143
+ "B-TIME": 75,
144
+ "B-UNIQUE_IDENTIFIER": 77,
145
+ "B-URL": 79,
146
+ "B-USER_NAME": 81,
147
+ "B-VEHICLE_IDENTIFIER": 83,
148
+ "I-ACCOUNT_NUMBER": 2,
149
+ "I-ADDRESS": 4,
150
+ "I-API_KEY": 6,
151
+ "I-BANK_ROUTING_NUMBER": 8,
152
+ "I-BIOMETRIC_IDENTIFIER": 10,
153
+ "I-CERTIFICATE_LICENSE_NUMBER": 12,
154
+ "I-CITY": 14,
155
+ "I-COMPANY_NAME": 16,
156
+ "I-COORDINATE": 18,
157
+ "I-COUNTRY": 20,
158
+ "I-CREDIT_CARD_NUMBER": 22,
159
+ "I-CUSTOMER_ID": 24,
160
+ "I-CVV": 26,
161
+ "I-DATE": 28,
162
+ "I-DATE_OF_BIRTH": 30,
163
+ "I-DATE_TIME": 32,
164
+ "I-DEVICE_IDENTIFIER": 34,
165
+ "I-EMAIL": 36,
166
+ "I-EMPLOYEE_ID": 38,
167
+ "I-FIRST_NAME": 40,
168
+ "I-HEALTH_PLAN_BENEFICIARY_NUMBER": 42,
169
+ "I-IPV4": 44,
170
+ "I-IPV6": 46,
171
+ "I-LAST_NAME": 48,
172
+ "I-LICENSE_PLATE": 50,
173
+ "I-MEDICAL_RECORD_NUMBER": 52,
174
+ "I-NAME": 54,
175
+ "I-NATIONAL_ID": 56,
176
+ "I-PASSWORD": 58,
177
+ "I-PHONE_NUMBER": 60,
178
+ "I-PIN": 62,
179
+ "I-POSTCODE": 64,
180
+ "I-SSN": 66,
181
+ "I-STATE": 68,
182
+ "I-STREET_ADDRESS": 70,
183
+ "I-SWIFT_BIC": 72,
184
+ "I-TAX_ID": 74,
185
+ "I-TIME": 76,
186
+ "I-UNIQUE_IDENTIFIER": 78,
187
+ "I-URL": 80,
188
+ "I-USER_NAME": 82,
189
+ "I-VEHICLE_IDENTIFIER": 84,
190
+ "O": 0
191
+ },
192
+ "layer_norm_eps": 1e-12,
193
+ "max_position_embeddings": 512,
194
+ "model_type": "bert",
195
+ "num_attention_heads": 12,
196
+ "num_hidden_layers": 12,
197
+ "pad_token_id": 0,
198
+ "pooler_fc_size": 768,
199
+ "pooler_num_attention_heads": 12,
200
+ "pooler_num_fc_layers": 3,
201
+ "pooler_size_per_head": 128,
202
+ "pooler_type": "first_token_transform",
203
+ "tie_word_embeddings": true,
204
+ "transformers_version": "5.3.0",
205
+ "type_vocab_size": 2,
206
+ "use_cache": false,
207
+ "vocab_size": 119547
208
+ }
gretel-pii-ready/TokenBased-BERT/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c8ef31b04e7e318825a513a0949fd8f0e66385605578071f555b5829541e28c1
3
+ size 709336188
gretel-pii-ready/TokenBased-BERT/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
gretel-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
+ }
gretel-pii-ready/TokenBased-BERT/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:94312732eff9efbc5cfb93e3ca98a015e5429d10739537c5b4cb08702f1fb1d1
3
+ size 5201