DeepMount00
commited on
Commit
•
bf1295d
1
Parent(s):
d01f1e7
Upload 7 files
Browse files- config.json +263 -0
- id2tag.json +1 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +16 -0
- vocab.txt +0 -0
config.json
ADDED
@@ -0,0 +1,263 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "dbmdz/bert-base-italian-xxl-cased",
|
3 |
+
"architectures": [
|
4 |
+
"BertForTokenClassification"
|
5 |
+
],
|
6 |
+
"attention_probs_dropout_prob": 0.1,
|
7 |
+
"classifier_dropout": null,
|
8 |
+
"hidden_act": "gelu",
|
9 |
+
"hidden_dropout_prob": 0.1,
|
10 |
+
"hidden_size": 768,
|
11 |
+
"id2label": {
|
12 |
+
"0": "O",
|
13 |
+
"1": "B-INDIRIZZO",
|
14 |
+
"2": "I-INDIRIZZO",
|
15 |
+
"3": "B-VALUTA",
|
16 |
+
"4": "I-VALUTA",
|
17 |
+
"5": "B-CVV",
|
18 |
+
"6": "I-CVV",
|
19 |
+
"7": "B-NUMERO_CONTO",
|
20 |
+
"8": "I-NUMERO_CONTO",
|
21 |
+
"9": "B-BIC",
|
22 |
+
"10": "I-BIC",
|
23 |
+
"11": "B-IBAN",
|
24 |
+
"12": "I-IBAN",
|
25 |
+
"13": "B-STATO",
|
26 |
+
"14": "I-STATO",
|
27 |
+
"15": "B-NOME",
|
28 |
+
"16": "I-NOME",
|
29 |
+
"17": "B-COGNOME",
|
30 |
+
"18": "I-COGNOME",
|
31 |
+
"19": "B-CODICE_POSTALE",
|
32 |
+
"20": "I-CODICE_POSTALE",
|
33 |
+
"21": "B-IP",
|
34 |
+
"22": "I-IP",
|
35 |
+
"23": "B-ORARIO",
|
36 |
+
"24": "I-ORARIO",
|
37 |
+
"25": "B-URL",
|
38 |
+
"26": "I-URL",
|
39 |
+
"27": "B-LUOGO",
|
40 |
+
"28": "I-LUOGO",
|
41 |
+
"29": "B-IMPORTO",
|
42 |
+
"30": "I-IMPORTO",
|
43 |
+
"31": "B-EMAIL",
|
44 |
+
"32": "I-EMAIL",
|
45 |
+
"33": "B-PASSWORD",
|
46 |
+
"34": "I-PASSWORD",
|
47 |
+
"35": "B-NUMERO_CARTA",
|
48 |
+
"36": "I-NUMERO_CARTA",
|
49 |
+
"37": "B-TARGA_VEICOLO",
|
50 |
+
"38": "I-TARGA_VEICOLO",
|
51 |
+
"39": "B-DATA_NASCITA",
|
52 |
+
"40": "I-DATA_NASCITA",
|
53 |
+
"41": "B-DATA_MORTE",
|
54 |
+
"42": "I-DATA_MORTE",
|
55 |
+
"43": "B-RAGIONE_SOCIALE",
|
56 |
+
"44": "I-RAGIONE_SOCIALE",
|
57 |
+
"45": "B-ETA",
|
58 |
+
"46": "I-ETA",
|
59 |
+
"47": "B-DATA",
|
60 |
+
"48": "I-DATA",
|
61 |
+
"49": "B-PROFESSIONE",
|
62 |
+
"50": "I-PROFESSIONE",
|
63 |
+
"51": "B-PIN",
|
64 |
+
"52": "I-PIN",
|
65 |
+
"53": "B-NUMERO_TELEFONO",
|
66 |
+
"54": "I-NUMERO_TELEFONO",
|
67 |
+
"55": "B-FOGLIO",
|
68 |
+
"56": "I-FOGLIO",
|
69 |
+
"57": "B-PARTICELLA",
|
70 |
+
"58": "I-PARTICELLA",
|
71 |
+
"59": "B-CARTELLA_CLINICA",
|
72 |
+
"60": "I-CARTELLA_CLINICA",
|
73 |
+
"61": "B-MALATTIA",
|
74 |
+
"62": "I-MALATTIA",
|
75 |
+
"63": "B-MEDICINA",
|
76 |
+
"64": "I-MEDICINA",
|
77 |
+
"65": "B-CODICE_FISCALE",
|
78 |
+
"66": "I-CODICE_FISCALE",
|
79 |
+
"67": "B-NUMERO_DOCUMENTO",
|
80 |
+
"68": "I-NUMERO_DOCUMENTO",
|
81 |
+
"69": "B-STORIA_CLINICA",
|
82 |
+
"70": "I-STORIA_CLINICA",
|
83 |
+
"71": "B-AVV_NOTAIO",
|
84 |
+
"72": "I-AVV_NOTAIO",
|
85 |
+
"73": "B-P_IVA",
|
86 |
+
"74": "I-P_IVA",
|
87 |
+
"75": "B-LEGGE",
|
88 |
+
"76": "I-LEGGE",
|
89 |
+
"77": "B-TASSO_MUTUO",
|
90 |
+
"78": "I-TASSO_MUTUO",
|
91 |
+
"79": "B-N_SENTENZA",
|
92 |
+
"80": "I-N_SENTENZA",
|
93 |
+
"81": "B-MAPPALE",
|
94 |
+
"82": "I-MAPPALE",
|
95 |
+
"83": "B-SUBALTERNO",
|
96 |
+
"84": "I-SUBALTERNO",
|
97 |
+
"85": "B-REGIME_PATRIMONIALE",
|
98 |
+
"86": "I-REGIME_PATRIMONIALE",
|
99 |
+
"87": "B-STATO_CIVILE",
|
100 |
+
"88": "I-STATO_CIVILE",
|
101 |
+
"89": "B-BANCA",
|
102 |
+
"90": "I-BANCA",
|
103 |
+
"91": "B-BRAND",
|
104 |
+
"92": "I-BRAND",
|
105 |
+
"93": "B-NUM_ASSEGNO_BANCARIO",
|
106 |
+
"94": "I-NUM_ASSEGNO_BANCARIO",
|
107 |
+
"95": "B-IMEI",
|
108 |
+
"96": "I-IMEI",
|
109 |
+
"97": "B-N_LICENZA",
|
110 |
+
"98": "I-N_LICENZA",
|
111 |
+
"99": "B-IPV6_1",
|
112 |
+
"100": "I-IPV6_1",
|
113 |
+
"101": "B-MAC",
|
114 |
+
"102": "I-MAC",
|
115 |
+
"103": "B-USER_AGENT",
|
116 |
+
"104": "I-USER_AGENT",
|
117 |
+
"105": "B-TRIBUNALE",
|
118 |
+
"106": "I-TRIBUNALE",
|
119 |
+
"107": "B-STRENGTH",
|
120 |
+
"108": "I-STRENGTH",
|
121 |
+
"109": "B-FREQUENZA",
|
122 |
+
"110": "I-FREQUENZA",
|
123 |
+
"111": "B-DURATION",
|
124 |
+
"112": "I-DURATION",
|
125 |
+
"113": "B-DOSAGGIO",
|
126 |
+
"114": "I-DOSAGGIO",
|
127 |
+
"115": "B-FORM",
|
128 |
+
"116": "I-FORM"
|
129 |
+
},
|
130 |
+
"initializer_range": 0.02,
|
131 |
+
"intermediate_size": 3072,
|
132 |
+
"label2id": {
|
133 |
+
"B-AVV_NOTAIO": 71,
|
134 |
+
"B-BANCA": 89,
|
135 |
+
"B-BIC": 9,
|
136 |
+
"B-BRAND": 91,
|
137 |
+
"B-CARTELLA_CLINICA": 59,
|
138 |
+
"B-CODICE_FISCALE": 65,
|
139 |
+
"B-CODICE_POSTALE": 19,
|
140 |
+
"B-COGNOME": 17,
|
141 |
+
"B-CVV": 5,
|
142 |
+
"B-DATA": 47,
|
143 |
+
"B-DATA_MORTE": 41,
|
144 |
+
"B-DATA_NASCITA": 39,
|
145 |
+
"B-DOSAGGIO": 113,
|
146 |
+
"B-DURATION": 111,
|
147 |
+
"B-EMAIL": 31,
|
148 |
+
"B-ETA": 45,
|
149 |
+
"B-FOGLIO": 55,
|
150 |
+
"B-FORM": 115,
|
151 |
+
"B-FREQUENZA": 109,
|
152 |
+
"B-IBAN": 11,
|
153 |
+
"B-IMEI": 95,
|
154 |
+
"B-IMPORTO": 29,
|
155 |
+
"B-INDIRIZZO": 1,
|
156 |
+
"B-IP": 21,
|
157 |
+
"B-IPV6_1": 99,
|
158 |
+
"B-LEGGE": 75,
|
159 |
+
"B-LUOGO": 27,
|
160 |
+
"B-MAC": 101,
|
161 |
+
"B-MALATTIA": 61,
|
162 |
+
"B-MAPPALE": 81,
|
163 |
+
"B-MEDICINA": 63,
|
164 |
+
"B-NOME": 15,
|
165 |
+
"B-NUMERO_CARTA": 35,
|
166 |
+
"B-NUMERO_CONTO": 7,
|
167 |
+
"B-NUMERO_DOCUMENTO": 67,
|
168 |
+
"B-NUMERO_TELEFONO": 53,
|
169 |
+
"B-NUM_ASSEGNO_BANCARIO": 93,
|
170 |
+
"B-N_LICENZA": 97,
|
171 |
+
"B-N_SENTENZA": 79,
|
172 |
+
"B-ORARIO": 23,
|
173 |
+
"B-PARTICELLA": 57,
|
174 |
+
"B-PASSWORD": 33,
|
175 |
+
"B-PIN": 51,
|
176 |
+
"B-PROFESSIONE": 49,
|
177 |
+
"B-P_IVA": 73,
|
178 |
+
"B-RAGIONE_SOCIALE": 43,
|
179 |
+
"B-REGIME_PATRIMONIALE": 85,
|
180 |
+
"B-STATO": 13,
|
181 |
+
"B-STATO_CIVILE": 87,
|
182 |
+
"B-STORIA_CLINICA": 69,
|
183 |
+
"B-STRENGTH": 107,
|
184 |
+
"B-SUBALTERNO": 83,
|
185 |
+
"B-TARGA_VEICOLO": 37,
|
186 |
+
"B-TASSO_MUTUO": 77,
|
187 |
+
"B-TRIBUNALE": 105,
|
188 |
+
"B-URL": 25,
|
189 |
+
"B-USER_AGENT": 103,
|
190 |
+
"B-VALUTA": 3,
|
191 |
+
"I-AVV_NOTAIO": 72,
|
192 |
+
"I-BANCA": 90,
|
193 |
+
"I-BIC": 10,
|
194 |
+
"I-BRAND": 92,
|
195 |
+
"I-CARTELLA_CLINICA": 60,
|
196 |
+
"I-CODICE_FISCALE": 66,
|
197 |
+
"I-CODICE_POSTALE": 20,
|
198 |
+
"I-COGNOME": 18,
|
199 |
+
"I-CVV": 6,
|
200 |
+
"I-DATA": 48,
|
201 |
+
"I-DATA_MORTE": 42,
|
202 |
+
"I-DATA_NASCITA": 40,
|
203 |
+
"I-DOSAGGIO": 114,
|
204 |
+
"I-DURATION": 112,
|
205 |
+
"I-EMAIL": 32,
|
206 |
+
"I-ETA": 46,
|
207 |
+
"I-FOGLIO": 56,
|
208 |
+
"I-FORM": 116,
|
209 |
+
"I-FREQUENZA": 110,
|
210 |
+
"I-IBAN": 12,
|
211 |
+
"I-IMEI": 96,
|
212 |
+
"I-IMPORTO": 30,
|
213 |
+
"I-INDIRIZZO": 2,
|
214 |
+
"I-IP": 22,
|
215 |
+
"I-IPV6_1": 100,
|
216 |
+
"I-LEGGE": 76,
|
217 |
+
"I-LUOGO": 28,
|
218 |
+
"I-MAC": 102,
|
219 |
+
"I-MALATTIA": 62,
|
220 |
+
"I-MAPPALE": 82,
|
221 |
+
"I-MEDICINA": 64,
|
222 |
+
"I-NOME": 16,
|
223 |
+
"I-NUMERO_CARTA": 36,
|
224 |
+
"I-NUMERO_CONTO": 8,
|
225 |
+
"I-NUMERO_DOCUMENTO": 68,
|
226 |
+
"I-NUMERO_TELEFONO": 54,
|
227 |
+
"I-NUM_ASSEGNO_BANCARIO": 94,
|
228 |
+
"I-N_LICENZA": 98,
|
229 |
+
"I-N_SENTENZA": 80,
|
230 |
+
"I-ORARIO": 24,
|
231 |
+
"I-PARTICELLA": 58,
|
232 |
+
"I-PASSWORD": 34,
|
233 |
+
"I-PIN": 52,
|
234 |
+
"I-PROFESSIONE": 50,
|
235 |
+
"I-P_IVA": 74,
|
236 |
+
"I-RAGIONE_SOCIALE": 44,
|
237 |
+
"I-REGIME_PATRIMONIALE": 86,
|
238 |
+
"I-STATO": 14,
|
239 |
+
"I-STATO_CIVILE": 88,
|
240 |
+
"I-STORIA_CLINICA": 70,
|
241 |
+
"I-STRENGTH": 108,
|
242 |
+
"I-SUBALTERNO": 84,
|
243 |
+
"I-TARGA_VEICOLO": 38,
|
244 |
+
"I-TASSO_MUTUO": 78,
|
245 |
+
"I-TRIBUNALE": 106,
|
246 |
+
"I-URL": 26,
|
247 |
+
"I-USER_AGENT": 104,
|
248 |
+
"I-VALUTA": 4,
|
249 |
+
"O": 0
|
250 |
+
},
|
251 |
+
"layer_norm_eps": 1e-12,
|
252 |
+
"max_position_embeddings": 512,
|
253 |
+
"model_type": "bert",
|
254 |
+
"num_attention_heads": 12,
|
255 |
+
"num_hidden_layers": 12,
|
256 |
+
"pad_token_id": 0,
|
257 |
+
"position_embedding_type": "absolute",
|
258 |
+
"torch_dtype": "float32",
|
259 |
+
"transformers_version": "4.33.0",
|
260 |
+
"type_vocab_size": 2,
|
261 |
+
"use_cache": true,
|
262 |
+
"vocab_size": 32102
|
263 |
+
}
|
id2tag.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{0: 'O', 1: 'B-INDIRIZZO', 2: 'I-INDIRIZZO', 3: 'B-VALUTA', 4: 'I-VALUTA', 5: 'B-CVV', 6: 'I-CVV', 7: 'B-NUMERO_CONTO', 8: 'I-NUMERO_CONTO', 9: 'B-BIC', 10: 'I-BIC', 11: 'B-IBAN', 12: 'I-IBAN', 13: 'B-STATO', 14: 'I-STATO', 15: 'B-NOME', 16: 'I-NOME', 17: 'B-COGNOME', 18: 'I-COGNOME', 19: 'B-CODICE_POSTALE', 20: 'I-CODICE_POSTALE', 21: 'B-IP', 22: 'I-IP', 23: 'B-ORARIO', 24: 'I-ORARIO', 25: 'B-URL', 26: 'I-URL', 27: 'B-LUOGO', 28: 'I-LUOGO', 29: 'B-IMPORTO', 30: 'I-IMPORTO', 31: 'B-EMAIL', 32: 'I-EMAIL', 33: 'B-PASSWORD', 34: 'I-PASSWORD', 35: 'B-NUMERO_CARTA', 36: 'I-NUMERO_CARTA', 37: 'B-TARGA_VEICOLO', 38: 'I-TARGA_VEICOLO', 39: 'B-DATA_NASCITA', 40: 'I-DATA_NASCITA', 41: 'B-DATA_MORTE', 42: 'I-DATA_MORTE', 43: 'B-RAGIONE_SOCIALE', 44: 'I-RAGIONE_SOCIALE', 45: 'B-ETA', 46: 'I-ETA', 47: 'B-DATA', 48: 'I-DATA', 49: 'B-PROFESSIONE', 50: 'I-PROFESSIONE', 51: 'B-PIN', 52: 'I-PIN', 53: 'B-NUMERO_TELEFONO', 54: 'I-NUMERO_TELEFONO', 55: 'B-FOGLIO', 56: 'I-FOGLIO', 57: 'B-PARTICELLA', 58: 'I-PARTICELLA', 59: 'B-CARTELLA_CLINICA', 60: 'I-CARTELLA_CLINICA', 61: 'B-MALATTIA', 62: 'I-MALATTIA', 63: 'B-MEDICINA', 64: 'I-MEDICINA', 65: 'B-CODICE_FISCALE', 66: 'I-CODICE_FISCALE', 67: 'B-NUMERO_DOCUMENTO', 68: 'I-NUMERO_DOCUMENTO', 69: 'B-STORIA_CLINICA', 70: 'I-STORIA_CLINICA', 71: 'B-AVV_NOTAIO', 72: 'I-AVV_NOTAIO', 73: 'B-P_IVA', 74: 'I-P_IVA', 75: 'B-LEGGE', 76: 'I-LEGGE', 77: 'B-TASSO_MUTUO', 78: 'I-TASSO_MUTUO', 79: 'B-N_SENTENZA', 80: 'I-N_SENTENZA', 81: 'B-MAPPALE', 82: 'I-MAPPALE', 83: 'B-SUBALTERNO', 84: 'I-SUBALTERNO', 85: 'B-REGIME_PATRIMONIALE', 86: 'I-REGIME_PATRIMONIALE', 87: 'B-STATO_CIVILE', 88: 'I-STATO_CIVILE', 89: 'B-BANCA', 90: 'I-BANCA', 91: 'B-BRAND', 92: 'I-BRAND', 93: 'B-NUM_ASSEGNO_BANCARIO', 94: 'I-NUM_ASSEGNO_BANCARIO', 95: 'B-IMEI', 96: 'I-IMEI', 97: 'B-N_LICENZA', 98: 'I-N_LICENZA', 99: 'B-IPV6_1', 100: 'I-IPV6_1', 101: 'B-MAC', 102: 'I-MAC', 103: 'B-USER_AGENT', 104: 'I-USER_AGENT', 105: 'B-TRIBUNALE', 106: 'I-TRIBUNALE', 107: 'B-STRENGTH', 108: 'I-STRENGTH', 109: 'B-FREQUENZA', 110: 'I-FREQUENZA', 111: 'B-DURATION', 112: 'I-DURATION', 113: 'B-DOSAGGIO', 114: 'I-DOSAGGIO', 115: 'B-FORM', 116: 'I-FORM'}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7b1e37f54bffe06b021e97159a7fa67c35f2acc129ff3ac427b5e2f87e695a45
|
3 |
+
size 440847977
|
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 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"clean_up_tokenization_spaces": true,
|
3 |
+
"cls_token": "[CLS]",
|
4 |
+
"do_basic_tokenize": true,
|
5 |
+
"do_lower_case": false,
|
6 |
+
"mask_token": "[MASK]",
|
7 |
+
"max_len": 512,
|
8 |
+
"model_max_length": 512,
|
9 |
+
"never_split": null,
|
10 |
+
"pad_token": "[PAD]",
|
11 |
+
"sep_token": "[SEP]",
|
12 |
+
"strip_accents": null,
|
13 |
+
"tokenize_chinese_chars": true,
|
14 |
+
"tokenizer_class": "BertTokenizer",
|
15 |
+
"unk_token": "[UNK]"
|
16 |
+
}
|
vocab.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|