Hugging Face's logo

MinhTriet
/
GenPII-framework-finetuned

Safetensors
Model card Files
xet
 Community
MinhTriet commited on
Commit
8d594ed
·
verified ·
1 Parent(s): 891bfd1

Upload folder using huggingface_hub

Browse files
Files changed (6) hide show
  1. gretel/GLiNER-NVIDIA/classification_report.json +266 -0
  2. gretel/GLiNER-NVIDIA/gliner_config.json +107 -0
  3. gretel/GLiNER-NVIDIA/predictions.json +0 -0
  4. gretel/GLiNER-NVIDIA/pytorch_model.bin +3 -0
  5. gretel/GLiNER-NVIDIA/tokenizer.json +0 -0
  6. gretel/GLiNER-NVIDIA/tokenizer_config.json +25 -0
gretel/GLiNER-NVIDIA/classification_report.json ADDED
@@ -0,0 +1,266 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "ACCOUNT_NUMBER": {
3
+ "precision": 0.9048,
4
+ "recall": 0.6129,
5
+ "f1-score": 0.7308,
6
+ "support": 31
7
+ },
8
+ "ADDRESS": {
9
+ "precision": 0.0,
10
+ "recall": 0.0,
11
+ "f1-score": 0.0,
12
+ "support": 110
13
+ },
14
+ "API_KEY": {
15
+ "precision": 0.8889,
16
+ "recall": 0.8,
17
+ "f1-score": 0.8421,
18
+ "support": 10
19
+ },
20
+ "BANK_ROUTING_NUMBER": {
21
+ "precision": 1.0,
22
+ "recall": 0.7115,
23
+ "f1-score": 0.8315,
24
+ "support": 52
25
+ },
26
+ "BIOMETRIC_IDENTIFIER": {
27
+ "precision": 0.871,
28
+ "recall": 0.9,
29
+ "f1-score": 0.8852,
30
+ "support": 30
31
+ },
32
+ "CERTIFICATE_LICENSE_NUMBER": {
33
+ "precision": 1.0,
34
+ "recall": 0.5417,
35
+ "f1-score": 0.7027,
36
+ "support": 24
37
+ },
38
+ "CITY": {
39
+ "precision": 0.099,
40
+ "recall": 0.8,
41
+ "f1-score": 0.1762,
42
+ "support": 25
43
+ },
44
+ "COMPANY_NAME": {
45
+ "precision": 0.8,
46
+ "recall": 0.6809,
47
+ "f1-score": 0.7356,
48
+ "support": 47
49
+ },
50
+ "COORDINATE": {
51
+ "precision": 0.875,
52
+ "recall": 0.6364,
53
+ "f1-score": 0.7368,
54
+ "support": 22
55
+ },
56
+ "COUNTRY": {
57
+ "precision": 0.16,
58
+ "recall": 0.6154,
59
+ "f1-score": 0.254,
60
+ "support": 13
61
+ },
62
+ "CREDIT_CARD_NUMBER": {
63
+ "precision": 0.9451,
64
+ "recall": 0.7288,
65
+ "f1-score": 0.823,
66
+ "support": 118
67
+ },
68
+ "CUSTOMER_ID": {
69
+ "precision": 0.9733,
70
+ "recall": 0.6854,
71
+ "f1-score": 0.8044,
72
+ "support": 213
73
+ },
74
+ "CVV": {
75
+ "precision": 0.0,
76
+ "recall": 0.0,
77
+ "f1-score": 0.0,
78
+ "support": 1
79
+ },
80
+ "DATE": {
81
+ "precision": 0.8297,
82
+ "recall": 0.6833,
83
+ "f1-score": 0.7494,
84
+ "support": 221
85
+ },
86
+ "DATE_OF_BIRTH": {
87
+ "precision": 0.9309,
88
+ "recall": 0.7592,
89
+ "f1-score": 0.8363,
90
+ "support": 461
91
+ },
92
+ "DATE_TIME": {
93
+ "precision": 0.8667,
94
+ "recall": 0.7027,
95
+ "f1-score": 0.7761,
96
+ "support": 37
97
+ },
98
+ "DEVICE_IDENTIFIER": {
99
+ "precision": 0.9429,
100
+ "recall": 0.7021,
101
+ "f1-score": 0.8049,
102
+ "support": 47
103
+ },
104
+ "EMAIL": {
105
+ "precision": 0.9867,
106
+ "recall": 0.7327,
107
+ "f1-score": 0.8409,
108
+ "support": 202
109
+ },
110
+ "EMPLOYEE_ID": {
111
+ "precision": 0.8938,
112
+ "recall": 0.7114,
113
+ "f1-score": 0.7922,
114
+ "support": 201
115
+ },
116
+ "FIRST_NAME": {
117
+ "precision": 0.4938,
118
+ "recall": 0.7692,
119
+ "f1-score": 0.6015,
120
+ "support": 260
121
+ },
122
+ "HEALTH_PLAN_BENEFICIARY_NUMBER": {
123
+ "precision": 1.0,
124
+ "recall": 0.75,
125
+ "f1-score": 0.8571,
126
+ "support": 16
127
+ },
128
+ "IPADDRESS": {
129
+ "precision": 0.9844,
130
+ "recall": 0.6667,
131
+ "f1-score": 0.795,
132
+ "support": 189
133
+ },
134
+ "LAST_NAME": {
135
+ "precision": 0.497,
136
+ "recall": 0.7401,
137
+ "f1-score": 0.5947,
138
+ "support": 227
139
+ },
140
+ "LICENSE_PLATE": {
141
+ "precision": 0.9394,
142
+ "recall": 0.6596,
143
+ "f1-score": 0.775,
144
+ "support": 141
145
+ },
146
+ "MEDICAL_RECORD_NUMBER": {
147
+ "precision": 0.9358,
148
+ "recall": 0.7317,
149
+ "f1-score": 0.8212,
150
+ "support": 518
151
+ },
152
+ "NAME": {
153
+ "precision": 0.973,
154
+ "recall": 0.1856,
155
+ "f1-score": 0.3117,
156
+ "support": 194
157
+ },
158
+ "NATIONAL_ID": {
159
+ "precision": 0.875,
160
+ "recall": 0.4667,
161
+ "f1-score": 0.6087,
162
+ "support": 15
163
+ },
164
+ "PASSWORD": {
165
+ "precision": 0.6923,
166
+ "recall": 0.6923,
167
+ "f1-score": 0.6923,
168
+ "support": 13
169
+ },
170
+ "PHONE_NUMBER": {
171
+ "precision": 0.9917,
172
+ "recall": 0.7018,
173
+ "f1-score": 0.8219,
174
+ "support": 171
175
+ },
176
+ "PIN": {
177
+ "precision": 0.0,
178
+ "recall": 0.0,
179
+ "f1-score": 0.0,
180
+ "support": 0
181
+ },
182
+ "POSTCODE": {
183
+ "precision": 0.3333,
184
+ "recall": 0.8235,
185
+ "f1-score": 0.4746,
186
+ "support": 17
187
+ },
188
+ "SSN": {
189
+ "precision": 0.9844,
190
+ "recall": 0.7463,
191
+ "f1-score": 0.849,
192
+ "support": 339
193
+ },
194
+ "STATE": {
195
+ "precision": 0.05,
196
+ "recall": 1.0,
197
+ "f1-score": 0.0952,
198
+ "support": 3
199
+ },
200
+ "STREET_ADDRESS": {
201
+ "precision": 0.5461,
202
+ "recall": 0.4368,
203
+ "f1-score": 0.4854,
204
+ "support": 190
205
+ },
206
+ "SWIFT_BIC": {
207
+ "precision": 0.5,
208
+ "recall": 1.0,
209
+ "f1-score": 0.6667,
210
+ "support": 6
211
+ },
212
+ "TAX_ID": {
213
+ "precision": 0.5,
214
+ "recall": 0.75,
215
+ "f1-score": 0.6,
216
+ "support": 4
217
+ },
218
+ "TIME": {
219
+ "precision": 0.4412,
220
+ "recall": 0.6522,
221
+ "f1-score": 0.5263,
222
+ "support": 23
223
+ },
224
+ "UNIQUE_IDENTIFIER": {
225
+ "precision": 0.9375,
226
+ "recall": 0.5556,
227
+ "f1-score": 0.6977,
228
+ "support": 27
229
+ },
230
+ "URL": {
231
+ "precision": 0.6,
232
+ "recall": 0.6,
233
+ "f1-score": 0.6,
234
+ "support": 5
235
+ },
236
+ "USER_NAME": {
237
+ "precision": 0.9211,
238
+ "recall": 0.7,
239
+ "f1-score": 0.7955,
240
+ "support": 50
241
+ },
242
+ "VEHICLE_IDENTIFIER": {
243
+ "precision": 1.0,
244
+ "recall": 0.5714,
245
+ "f1-score": 0.7273,
246
+ "support": 21
247
+ },
248
+ "OVERALL_MICRO": {
249
+ "precision": 0.75,
250
+ "recall": 0.6644,
251
+ "f1-score": 0.7046,
252
+ "support": 4294
253
+ },
254
+ "OVERALL_MACRO": {
255
+ "precision": 0.7113,
256
+ "recall": 0.6391,
257
+ "f1-score": 0.6273,
258
+ "support": 4294
259
+ },
260
+ "OVERALL_WEIGHTED": {
261
+ "precision": 0.8342,
262
+ "recall": 0.6644,
263
+ "f1-score": 0.7183,
264
+ "support": 4294
265
+ }
266
+ }
gretel/GLiNER-NVIDIA/gliner_config.json ADDED
@@ -0,0 +1,107 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "class_token_index": 128002,
3
+ "dropout": 0.4,
4
+ "embed_ent_token": true,
5
+ "encoder_config": {
6
+ "_name_or_path": "microsoft/deberta-v3-large",
7
+ "add_cross_attention": false,
8
+ "architectures": null,
9
+ "attention_probs_dropout_prob": 0.1,
10
+ "bos_token_id": null,
11
+ "chunk_size_feed_forward": 0,
12
+ "cross_attention_hidden_size": null,
13
+ "decoder_start_token_id": null,
14
+ "dtype": null,
15
+ "eos_token_id": null,
16
+ "finetuning_task": null,
17
+ "hidden_act": "gelu",
18
+ "hidden_dropout_prob": 0.1,
19
+ "hidden_size": 1024,
20
+ "id2label": {
21
+ "0": "LABEL_0",
22
+ "1": "LABEL_1"
23
+ },
24
+ "initializer_range": 0.02,
25
+ "intermediate_size": 4096,
26
+ "is_decoder": false,
27
+ "is_encoder_decoder": false,
28
+ "label2id": {
29
+ "LABEL_0": 0,
30
+ "LABEL_1": 1
31
+ },
32
+ "layer_norm_eps": 1e-07,
33
+ "legacy": true,
34
+ "max_position_embeddings": 512,
35
+ "max_relative_positions": -1,
36
+ "model_type": "deberta-v2",
37
+ "norm_rel_ebd": "layer_norm",
38
+ "num_attention_heads": 16,
39
+ "num_hidden_layers": 24,
40
+ "output_attentions": false,
41
+ "output_hidden_states": false,
42
+ "pad_token_id": 0,
43
+ "pooler_dropout": 0,
44
+ "pooler_hidden_act": "gelu",
45
+ "pooler_hidden_size": 1024,
46
+ "pos_att_type": [
47
+ "p2c",
48
+ "c2p"
49
+ ],
50
+ "position_biased_input": false,
51
+ "position_buckets": 256,
52
+ "prefix": null,
53
+ "problem_type": null,
54
+ "pruned_heads": {},
55
+ "relative_attention": true,
56
+ "return_dict": true,
57
+ "sep_token_id": null,
58
+ "share_att_key": true,
59
+ "task_specific_params": null,
60
+ "tf_legacy_loss": false,
61
+ "tie_encoder_decoder": false,
62
+ "tie_word_embeddings": true,
63
+ "tokenizer_class": null,
64
+ "torchscript": false,
65
+ "type_vocab_size": 0,
66
+ "use_bfloat16": false,
67
+ "vocab_size": 128004
68
+ },
69
+ "ent_token": "<<ENT>>",
70
+ "eval_every": 5000,
71
+ "fine_tune": true,
72
+ "fuse_layers": false,
73
+ "has_rnn": true,
74
+ "hidden_size": 512,
75
+ "id_to_classes": null,
76
+ "labels_encoder": null,
77
+ "labels_encoder_config": null,
78
+ "lr_encoder": "1e-5",
79
+ "lr_others": "5e-5",
80
+ "max_len": 384,
81
+ "max_neg_type_ratio": 1,
82
+ "max_types": 25,
83
+ "max_width": 12,
84
+ "model_name": "microsoft/deberta-v3-large",
85
+ "model_type": null,
86
+ "name": "correct",
87
+ "neg_spans_ratio": 1.0,
88
+ "num_post_fusion_layers": 1,
89
+ "num_rnn_layers": 1,
90
+ "num_steps": 30000,
91
+ "post_fusion_schema": "",
92
+ "precomputed_prompts_mode": null,
93
+ "random_drop": true,
94
+ "represent_spans": false,
95
+ "sep_token": "<<SEP>>",
96
+ "shuffle_types": true,
97
+ "size_sup": -1,
98
+ "span_loss_coef": 1.0,
99
+ "span_mode": "markerV0",
100
+ "subtoken_pooling": "first",
101
+ "token_loss_coef": 1.0,
102
+ "train_batch_size": 8,
103
+ "transformers_version": "5.6.2",
104
+ "vocab_size": 128004,
105
+ "warmup_ratio": 3000,
106
+ "words_splitter_type": "whitespace"
107
+ }
gretel/GLiNER-NVIDIA/predictions.json ADDED
The diff for this file is too large to render. See raw diff
 
gretel/GLiNER-NVIDIA/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1281b8f6e68dc68b19f6563122c7b1325f869535c7649c702f7c07e1c84f6b5b
3
+ size 1782005383
gretel/GLiNER-NVIDIA/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
gretel/GLiNER-NVIDIA/tokenizer_config.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": true,
3
+ "backend": "tokenizers",
4
+ "bos_token": "[CLS]",
5
+ "clean_up_tokenization_spaces": false,
6
+ "cls_token": "[CLS]",
7
+ "do_lower_case": false,
8
+ "eos_token": "[SEP]",
9
+ "is_local": true,
10
+ "local_files_only": false,
11
+ "mask_token": "[MASK]",
12
+ "max_length": null,
13
+ "model_max_length": 1000000000000000019884624838656,
14
+ "pad_to_multiple_of": null,
15
+ "pad_token": "[PAD]",
16
+ "pad_token_type_id": 0,
17
+ "padding_side": "right",
18
+ "sep_token": "[SEP]",
19
+ "sp_model_kwargs": {},
20
+ "split_by_punct": false,
21
+ "tokenizer_class": "DebertaV2Tokenizer",
22
+ "unk_id": 3,
23
+ "unk_token": "[UNK]",
24
+ "vocab_type": "spm"
25
+ }