rouabelgacem
commited on
Commit
·
a58f074
1
Parent(s):
1b03e83
commit1
Browse files- .gitignore +1 -0
- config.json +264 -0
- pytorch_model.bin +3 -0
- runs/Mar17_09-09-37_5f63d64e18cb/1679044192.7661684/events.out.tfevents.1679044192.5f63d64e18cb.1524.3 +3 -0
- runs/Mar17_09-09-37_5f63d64e18cb/events.out.tfevents.1679044192.5f63d64e18cb.1524.2 +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +13 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
checkpoint-*/
|
config.json
ADDED
@@ -0,0 +1,264 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "bert-base-cased",
|
3 |
+
"architectures": [
|
4 |
+
"BertForTokenClassification"
|
5 |
+
],
|
6 |
+
"attention_probs_dropout_prob": 0.1,
|
7 |
+
"classifier_dropout": null,
|
8 |
+
"gradient_checkpointing": false,
|
9 |
+
"hidden_act": "gelu",
|
10 |
+
"hidden_dropout_prob": 0.1,
|
11 |
+
"hidden_size": 768,
|
12 |
+
"id2label": {
|
13 |
+
"0": "O",
|
14 |
+
"1": "B-O",
|
15 |
+
"2": "I-O",
|
16 |
+
"3": "L-O",
|
17 |
+
"4": "U-O",
|
18 |
+
"5": "B-PER",
|
19 |
+
"6": "I-PER",
|
20 |
+
"7": "L-PER",
|
21 |
+
"8": "U-PER",
|
22 |
+
"9": "B-LOC",
|
23 |
+
"10": "I-LOC",
|
24 |
+
"11": "L-LOC",
|
25 |
+
"12": "U-LOC",
|
26 |
+
"13": "B-ORG",
|
27 |
+
"14": "I-ORG",
|
28 |
+
"15": "L-ORG",
|
29 |
+
"16": "U-ORG",
|
30 |
+
"17": "B-NRP",
|
31 |
+
"18": "I-NRP",
|
32 |
+
"19": "L-NRP",
|
33 |
+
"20": "U-NRP",
|
34 |
+
"21": "B-DATE_TIME",
|
35 |
+
"22": "I-DATE_TIME",
|
36 |
+
"23": "L-DATE_TIME",
|
37 |
+
"24": "U-DATE_TIME",
|
38 |
+
"25": "B-CREDIT_CARD",
|
39 |
+
"26": "I-CREDIT_CARD",
|
40 |
+
"27": "L-CREDIT_CARD",
|
41 |
+
"28": "U-CREDIT_CARD",
|
42 |
+
"29": "B-URL",
|
43 |
+
"30": "I-URL",
|
44 |
+
"31": "L-URL",
|
45 |
+
"32": "U-URL",
|
46 |
+
"33": "B-IBAN_CODE",
|
47 |
+
"34": "I-IBAN_CODE",
|
48 |
+
"35": "L-IBAN_CODE",
|
49 |
+
"36": "U-IBAN_CODE",
|
50 |
+
"37": "B-US_BANK_NUMBER",
|
51 |
+
"38": "I-US_BANK_NUMBER",
|
52 |
+
"39": "L-US_BANK_NUMBER",
|
53 |
+
"40": "U-US_BANK_NUMBER",
|
54 |
+
"41": "B-PHONE_NUMBER",
|
55 |
+
"42": "I-PHONE_NUMBER",
|
56 |
+
"43": "L-PHONE_NUMBER",
|
57 |
+
"44": "U-PHONE_NUMBER",
|
58 |
+
"45": "B-US_SSN",
|
59 |
+
"46": "I-US_SSN",
|
60 |
+
"47": "L-US_SSN",
|
61 |
+
"48": "U-US_SSN",
|
62 |
+
"49": "B-US_PASSPORT",
|
63 |
+
"50": "I-US_PASSPORT",
|
64 |
+
"51": "L-US_PASSPORT",
|
65 |
+
"52": "U-US_PASSPORT",
|
66 |
+
"53": "B-US_DRIVER_LICENSE",
|
67 |
+
"54": "I-US_DRIVER_LICENSE",
|
68 |
+
"55": "L-US_DRIVER_LICENSE",
|
69 |
+
"56": "U-US_DRIVER_LICENSE",
|
70 |
+
"57": "B-US_LICENSE_PLATE",
|
71 |
+
"58": "I-US_LICENSE_PLATE",
|
72 |
+
"59": "L-US_LICENSE_PLATE",
|
73 |
+
"60": "U-US_LICENSE_PLATE",
|
74 |
+
"61": "B-IP_ADDRESS",
|
75 |
+
"62": "I-IP_ADDRESS",
|
76 |
+
"63": "L-IP_ADDRESS",
|
77 |
+
"64": "U-IP_ADDRESS",
|
78 |
+
"65": "B-US_ITIN",
|
79 |
+
"66": "I-US_ITIN",
|
80 |
+
"67": "L-US_ITIN",
|
81 |
+
"68": "U-US_ITIN",
|
82 |
+
"69": "B-EMAIL_ADDRESS",
|
83 |
+
"70": "I-EMAIL_ADDRESS",
|
84 |
+
"71": "L-EMAIL_ADDRESS",
|
85 |
+
"72": "U-EMAIL_ADDRESS",
|
86 |
+
"73": "B-TITLE",
|
87 |
+
"74": "I-TITLE",
|
88 |
+
"75": "L-TITLE",
|
89 |
+
"76": "U-TITLE",
|
90 |
+
"77": "B-COORDINATE",
|
91 |
+
"78": "I-COORDINATE",
|
92 |
+
"79": "L-COORDINATE",
|
93 |
+
"80": "U-COORDINATE",
|
94 |
+
"81": "B-IMEI",
|
95 |
+
"82": "I-IMEI",
|
96 |
+
"83": "L-IMEI",
|
97 |
+
"84": "U-IMEI",
|
98 |
+
"85": "B-PASSWORD",
|
99 |
+
"86": "I-PASSWORD",
|
100 |
+
"87": "L-PASSWORD",
|
101 |
+
"88": "U-PASSWORD",
|
102 |
+
"89": "B-LICENSE_PLATE",
|
103 |
+
"90": "I-LICENSE_PLATE",
|
104 |
+
"91": "L-LICENSE_PLATE",
|
105 |
+
"92": "U-LICENSE_PLATE",
|
106 |
+
"93": "B-CURRENCY",
|
107 |
+
"94": "I-CURRENCY",
|
108 |
+
"95": "L-CURRENCY",
|
109 |
+
"96": "U-CURRENCY",
|
110 |
+
"97": "B-FINANCIAL",
|
111 |
+
"98": "I-FINANCIAL",
|
112 |
+
"99": "L-FINANCIAL",
|
113 |
+
"100": "U-FINANCIAL",
|
114 |
+
"101": "B-ROUTING_NUMBER",
|
115 |
+
"102": "I-ROUTING_NUMBER",
|
116 |
+
"103": "L-ROUTING_NUMBER",
|
117 |
+
"104": "U-ROUTING_NUMBER",
|
118 |
+
"105": "B-SWIFT_CODE",
|
119 |
+
"106": "I-SWIFT_CODE",
|
120 |
+
"107": "L-SWIFT_CODE",
|
121 |
+
"108": "U-SWIFT_CODE",
|
122 |
+
"109": "B-MAC_ADDRESS",
|
123 |
+
"110": "I-MAC_ADDRESS",
|
124 |
+
"111": "L-MAC_ADDRESS",
|
125 |
+
"112": "U-MAC_ADDRESS",
|
126 |
+
"113": "B-AGE",
|
127 |
+
"114": "I-AGE",
|
128 |
+
"115": "L-AGE",
|
129 |
+
"116": "U-AGE"
|
130 |
+
},
|
131 |
+
"initializer_range": 0.02,
|
132 |
+
"intermediate_size": 3072,
|
133 |
+
"label2id": {
|
134 |
+
"B-AGE": 113,
|
135 |
+
"B-COORDINATE": 77,
|
136 |
+
"B-CREDIT_CARD": 25,
|
137 |
+
"B-CURRENCY": 93,
|
138 |
+
"B-DATE_TIME": 21,
|
139 |
+
"B-EMAIL_ADDRESS": 69,
|
140 |
+
"B-FINANCIAL": 97,
|
141 |
+
"B-IBAN_CODE": 33,
|
142 |
+
"B-IMEI": 81,
|
143 |
+
"B-IP_ADDRESS": 61,
|
144 |
+
"B-LICENSE_PLATE": 89,
|
145 |
+
"B-LOC": 9,
|
146 |
+
"B-MAC_ADDRESS": 109,
|
147 |
+
"B-NRP": 17,
|
148 |
+
"B-O": 1,
|
149 |
+
"B-ORG": 13,
|
150 |
+
"B-PASSWORD": 85,
|
151 |
+
"B-PER": 5,
|
152 |
+
"B-PHONE_NUMBER": 41,
|
153 |
+
"B-ROUTING_NUMBER": 101,
|
154 |
+
"B-SWIFT_CODE": 105,
|
155 |
+
"B-TITLE": 73,
|
156 |
+
"B-URL": 29,
|
157 |
+
"B-US_BANK_NUMBER": 37,
|
158 |
+
"B-US_DRIVER_LICENSE": 53,
|
159 |
+
"B-US_ITIN": 65,
|
160 |
+
"B-US_LICENSE_PLATE": 57,
|
161 |
+
"B-US_PASSPORT": 49,
|
162 |
+
"B-US_SSN": 45,
|
163 |
+
"I-AGE": 114,
|
164 |
+
"I-COORDINATE": 78,
|
165 |
+
"I-CREDIT_CARD": 26,
|
166 |
+
"I-CURRENCY": 94,
|
167 |
+
"I-DATE_TIME": 22,
|
168 |
+
"I-EMAIL_ADDRESS": 70,
|
169 |
+
"I-FINANCIAL": 98,
|
170 |
+
"I-IBAN_CODE": 34,
|
171 |
+
"I-IMEI": 82,
|
172 |
+
"I-IP_ADDRESS": 62,
|
173 |
+
"I-LICENSE_PLATE": 90,
|
174 |
+
"I-LOC": 10,
|
175 |
+
"I-MAC_ADDRESS": 110,
|
176 |
+
"I-NRP": 18,
|
177 |
+
"I-O": 2,
|
178 |
+
"I-ORG": 14,
|
179 |
+
"I-PASSWORD": 86,
|
180 |
+
"I-PER": 6,
|
181 |
+
"I-PHONE_NUMBER": 42,
|
182 |
+
"I-ROUTING_NUMBER": 102,
|
183 |
+
"I-SWIFT_CODE": 106,
|
184 |
+
"I-TITLE": 74,
|
185 |
+
"I-URL": 30,
|
186 |
+
"I-US_BANK_NUMBER": 38,
|
187 |
+
"I-US_DRIVER_LICENSE": 54,
|
188 |
+
"I-US_ITIN": 66,
|
189 |
+
"I-US_LICENSE_PLATE": 58,
|
190 |
+
"I-US_PASSPORT": 50,
|
191 |
+
"I-US_SSN": 46,
|
192 |
+
"L-AGE": 115,
|
193 |
+
"L-COORDINATE": 79,
|
194 |
+
"L-CREDIT_CARD": 27,
|
195 |
+
"L-CURRENCY": 95,
|
196 |
+
"L-DATE_TIME": 23,
|
197 |
+
"L-EMAIL_ADDRESS": 71,
|
198 |
+
"L-FINANCIAL": 99,
|
199 |
+
"L-IBAN_CODE": 35,
|
200 |
+
"L-IMEI": 83,
|
201 |
+
"L-IP_ADDRESS": 63,
|
202 |
+
"L-LICENSE_PLATE": 91,
|
203 |
+
"L-LOC": 11,
|
204 |
+
"L-MAC_ADDRESS": 111,
|
205 |
+
"L-NRP": 19,
|
206 |
+
"L-O": 3,
|
207 |
+
"L-ORG": 15,
|
208 |
+
"L-PASSWORD": 87,
|
209 |
+
"L-PER": 7,
|
210 |
+
"L-PHONE_NUMBER": 43,
|
211 |
+
"L-ROUTING_NUMBER": 103,
|
212 |
+
"L-SWIFT_CODE": 107,
|
213 |
+
"L-TITLE": 75,
|
214 |
+
"L-URL": 31,
|
215 |
+
"L-US_BANK_NUMBER": 39,
|
216 |
+
"L-US_DRIVER_LICENSE": 55,
|
217 |
+
"L-US_ITIN": 67,
|
218 |
+
"L-US_LICENSE_PLATE": 59,
|
219 |
+
"L-US_PASSPORT": 51,
|
220 |
+
"L-US_SSN": 47,
|
221 |
+
"O": 0,
|
222 |
+
"U-AGE": 116,
|
223 |
+
"U-COORDINATE": 80,
|
224 |
+
"U-CREDIT_CARD": 28,
|
225 |
+
"U-CURRENCY": 96,
|
226 |
+
"U-DATE_TIME": 24,
|
227 |
+
"U-EMAIL_ADDRESS": 72,
|
228 |
+
"U-FINANCIAL": 100,
|
229 |
+
"U-IBAN_CODE": 36,
|
230 |
+
"U-IMEI": 84,
|
231 |
+
"U-IP_ADDRESS": 64,
|
232 |
+
"U-LICENSE_PLATE": 92,
|
233 |
+
"U-LOC": 12,
|
234 |
+
"U-MAC_ADDRESS": 112,
|
235 |
+
"U-NRP": 20,
|
236 |
+
"U-O": 4,
|
237 |
+
"U-ORG": 16,
|
238 |
+
"U-PASSWORD": 88,
|
239 |
+
"U-PER": 8,
|
240 |
+
"U-PHONE_NUMBER": 44,
|
241 |
+
"U-ROUTING_NUMBER": 104,
|
242 |
+
"U-SWIFT_CODE": 108,
|
243 |
+
"U-TITLE": 76,
|
244 |
+
"U-URL": 32,
|
245 |
+
"U-US_BANK_NUMBER": 40,
|
246 |
+
"U-US_DRIVER_LICENSE": 56,
|
247 |
+
"U-US_ITIN": 68,
|
248 |
+
"U-US_LICENSE_PLATE": 60,
|
249 |
+
"U-US_PASSPORT": 52,
|
250 |
+
"U-US_SSN": 48
|
251 |
+
},
|
252 |
+
"layer_norm_eps": 1e-12,
|
253 |
+
"max_position_embeddings": 512,
|
254 |
+
"model_type": "bert",
|
255 |
+
"num_attention_heads": 12,
|
256 |
+
"num_hidden_layers": 12,
|
257 |
+
"pad_token_id": 0,
|
258 |
+
"position_embedding_type": "absolute",
|
259 |
+
"torch_dtype": "float32",
|
260 |
+
"transformers_version": "4.27.1",
|
261 |
+
"type_vocab_size": 2,
|
262 |
+
"use_cache": true,
|
263 |
+
"vocab_size": 28996
|
264 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9e7b5280b18a20d66e5da7f968771e2f5f09b202acc43920e3fee0589b9257e9
|
3 |
+
size 431310765
|
runs/Mar17_09-09-37_5f63d64e18cb/1679044192.7661684/events.out.tfevents.1679044192.5f63d64e18cb.1524.3
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4933728ad1a1e71d48734fdf8e5d99058e7991bcc79ec39cbe11f9d7151c0ff1
|
3 |
+
size 5784
|
runs/Mar17_09-09-37_5f63d64e18cb/events.out.tfevents.1679044192.5f63d64e18cb.1524.2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:31eaf71a373acc0e47c2cfc59c7a241deed2f87aae4dd084cc3b75c04bf2f27e
|
3 |
+
size 18307
|
special_tokens_map.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"cls_token": "[CLS]",
|
3 |
+
"mask_token": "[MASK]",
|
4 |
+
"pad_token": "[PAD]",
|
5 |
+
"sep_token": "[SEP]",
|
6 |
+
"unk_token": "[UNK]"
|
7 |
+
}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"cls_token": "[CLS]",
|
3 |
+
"do_lower_case": false,
|
4 |
+
"mask_token": "[MASK]",
|
5 |
+
"model_max_length": 512,
|
6 |
+
"pad_token": "[PAD]",
|
7 |
+
"sep_token": "[SEP]",
|
8 |
+
"special_tokens_map_file": null,
|
9 |
+
"strip_accents": null,
|
10 |
+
"tokenize_chinese_chars": true,
|
11 |
+
"tokenizer_class": "BertTokenizer",
|
12 |
+
"unk_token": "[UNK]"
|
13 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:93fcaf02ef0aa880f1033d1888348669968df95dc8cb1e4b5f9b95307e8e6706
|
3 |
+
size 3579
|
vocab.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|