Training done
Browse files- added_tokens.json +25 -35
- tokenizer.json +24 -114
added_tokens.json
CHANGED
@@ -1,44 +1,34 @@
|
|
1 |
{
|
2 |
-
"</
|
3 |
-
"</
|
4 |
-
"</
|
5 |
-
"</
|
6 |
-
"</s_Race>": 57542,
|
7 |
"</s_State file #>": 57528,
|
8 |
-
"</
|
9 |
-
"</
|
10 |
-
"</
|
11 |
-
"</
|
12 |
-
"</
|
13 |
-
"</
|
14 |
-
"</s_death>": 57536,
|
15 |
-
"</s_father_L>": 57562,
|
16 |
-
"</s_father_R>": 57560,
|
17 |
-
"</s_name>": 57534,
|
18 |
"</s_person>": 57526,
|
19 |
-
"</s_person_data>":
|
20 |
-
"</s_relation>":
|
21 |
-
"<
|
22 |
-
"<
|
23 |
-
"<
|
24 |
-
"<
|
25 |
-
"<s_Race>": 57541,
|
26 |
"<s_State file #>": 57527,
|
27 |
-
"<
|
28 |
-
"<
|
29 |
-
"<
|
30 |
-
"<
|
31 |
-
"<
|
32 |
-
"<
|
33 |
-
"<s_county>": 57529,
|
34 |
-
"<s_death>": 57535,
|
35 |
-
"<s_father_L>": 57561,
|
36 |
-
"<s_father_R>": 57559,
|
37 |
"<s_iitcdip>": 57523,
|
38 |
-
"<s_name>":
|
39 |
"<s_person>": 57525,
|
40 |
-
"<s_person_data>":
|
41 |
-
"<s_relation>":
|
42 |
"<s_synthdog>": 57524,
|
43 |
"<sep/>": 57522
|
44 |
}
|
|
|
1 |
{
|
2 |
+
"</s_Age>": 57546,
|
3 |
+
"</s_Gender>": 57538,
|
4 |
+
"</s_Mother>": 57550,
|
5 |
+
"</s_Race>": 57540,
|
|
|
6 |
"</s_State file #>": 57528,
|
7 |
+
"</s_birth_day>": 57542,
|
8 |
+
"</s_birthplace>": 57544,
|
9 |
+
"</s_city>": 57530,
|
10 |
+
"</s_death>": 57534,
|
11 |
+
"</s_father>": 57552,
|
12 |
+
"</s_name>": 57532,
|
|
|
|
|
|
|
|
|
13 |
"</s_person>": 57526,
|
14 |
+
"</s_person_data>": 57536,
|
15 |
+
"</s_relation>": 57548,
|
16 |
+
"<s_Age>": 57545,
|
17 |
+
"<s_Gender>": 57537,
|
18 |
+
"<s_Mother>": 57549,
|
19 |
+
"<s_Race>": 57539,
|
|
|
20 |
"<s_State file #>": 57527,
|
21 |
+
"<s_birth_day>": 57541,
|
22 |
+
"<s_birthplace>": 57543,
|
23 |
+
"<s_city>": 57529,
|
24 |
+
"<s_cord-v2>": 57553,
|
25 |
+
"<s_death>": 57533,
|
26 |
+
"<s_father>": 57551,
|
|
|
|
|
|
|
|
|
27 |
"<s_iitcdip>": 57523,
|
28 |
+
"<s_name>": 57531,
|
29 |
"<s_person>": 57525,
|
30 |
+
"<s_person_data>": 57535,
|
31 |
+
"<s_relation>": 57547,
|
32 |
"<s_synthdog>": 57524,
|
33 |
"<sep/>": 57522
|
34 |
}
|
tokenizer.json
CHANGED
@@ -113,7 +113,7 @@
|
|
113 |
},
|
114 |
{
|
115 |
"id": 57529,
|
116 |
-
"content": "<
|
117 |
"single_word": false,
|
118 |
"lstrip": false,
|
119 |
"rstrip": false,
|
@@ -122,7 +122,7 @@
|
|
122 |
},
|
123 |
{
|
124 |
"id": 57530,
|
125 |
-
"content": "</
|
126 |
"single_word": false,
|
127 |
"lstrip": false,
|
128 |
"rstrip": false,
|
@@ -131,7 +131,7 @@
|
|
131 |
},
|
132 |
{
|
133 |
"id": 57531,
|
134 |
-
"content": "<
|
135 |
"single_word": false,
|
136 |
"lstrip": false,
|
137 |
"rstrip": false,
|
@@ -140,7 +140,7 @@
|
|
140 |
},
|
141 |
{
|
142 |
"id": 57532,
|
143 |
-
"content": "</
|
144 |
"single_word": false,
|
145 |
"lstrip": false,
|
146 |
"rstrip": false,
|
@@ -149,7 +149,7 @@
|
|
149 |
},
|
150 |
{
|
151 |
"id": 57533,
|
152 |
-
"content": "<
|
153 |
"single_word": false,
|
154 |
"lstrip": false,
|
155 |
"rstrip": false,
|
@@ -158,7 +158,7 @@
|
|
158 |
},
|
159 |
{
|
160 |
"id": 57534,
|
161 |
-
"content": "</
|
162 |
"single_word": false,
|
163 |
"lstrip": false,
|
164 |
"rstrip": false,
|
@@ -167,7 +167,7 @@
|
|
167 |
},
|
168 |
{
|
169 |
"id": 57535,
|
170 |
-
"content": "<
|
171 |
"single_word": false,
|
172 |
"lstrip": false,
|
173 |
"rstrip": false,
|
@@ -176,7 +176,7 @@
|
|
176 |
},
|
177 |
{
|
178 |
"id": 57536,
|
179 |
-
"content": "</
|
180 |
"single_word": false,
|
181 |
"lstrip": false,
|
182 |
"rstrip": false,
|
@@ -185,7 +185,7 @@
|
|
185 |
},
|
186 |
{
|
187 |
"id": 57537,
|
188 |
-
"content": "<
|
189 |
"single_word": false,
|
190 |
"lstrip": false,
|
191 |
"rstrip": false,
|
@@ -194,7 +194,7 @@
|
|
194 |
},
|
195 |
{
|
196 |
"id": 57538,
|
197 |
-
"content": "</
|
198 |
"single_word": false,
|
199 |
"lstrip": false,
|
200 |
"rstrip": false,
|
@@ -203,7 +203,7 @@
|
|
203 |
},
|
204 |
{
|
205 |
"id": 57539,
|
206 |
-
"content": "<
|
207 |
"single_word": false,
|
208 |
"lstrip": false,
|
209 |
"rstrip": false,
|
@@ -212,7 +212,7 @@
|
|
212 |
},
|
213 |
{
|
214 |
"id": 57540,
|
215 |
-
"content": "</
|
216 |
"single_word": false,
|
217 |
"lstrip": false,
|
218 |
"rstrip": false,
|
@@ -221,7 +221,7 @@
|
|
221 |
},
|
222 |
{
|
223 |
"id": 57541,
|
224 |
-
"content": "<
|
225 |
"single_word": false,
|
226 |
"lstrip": false,
|
227 |
"rstrip": false,
|
@@ -230,7 +230,7 @@
|
|
230 |
},
|
231 |
{
|
232 |
"id": 57542,
|
233 |
-
"content": "</
|
234 |
"single_word": false,
|
235 |
"lstrip": false,
|
236 |
"rstrip": false,
|
@@ -239,7 +239,7 @@
|
|
239 |
},
|
240 |
{
|
241 |
"id": 57543,
|
242 |
-
"content": "<
|
243 |
"single_word": false,
|
244 |
"lstrip": false,
|
245 |
"rstrip": false,
|
@@ -248,7 +248,7 @@
|
|
248 |
},
|
249 |
{
|
250 |
"id": 57544,
|
251 |
-
"content": "</
|
252 |
"single_word": false,
|
253 |
"lstrip": false,
|
254 |
"rstrip": false,
|
@@ -257,7 +257,7 @@
|
|
257 |
},
|
258 |
{
|
259 |
"id": 57545,
|
260 |
-
"content": "<
|
261 |
"single_word": false,
|
262 |
"lstrip": false,
|
263 |
"rstrip": false,
|
@@ -266,7 +266,7 @@
|
|
266 |
},
|
267 |
{
|
268 |
"id": 57546,
|
269 |
-
"content": "</
|
270 |
"single_word": false,
|
271 |
"lstrip": false,
|
272 |
"rstrip": false,
|
@@ -275,7 +275,7 @@
|
|
275 |
},
|
276 |
{
|
277 |
"id": 57547,
|
278 |
-
"content": "<
|
279 |
"single_word": false,
|
280 |
"lstrip": false,
|
281 |
"rstrip": false,
|
@@ -284,7 +284,7 @@
|
|
284 |
},
|
285 |
{
|
286 |
"id": 57548,
|
287 |
-
"content": "</
|
288 |
"single_word": false,
|
289 |
"lstrip": false,
|
290 |
"rstrip": false,
|
@@ -293,7 +293,7 @@
|
|
293 |
},
|
294 |
{
|
295 |
"id": 57549,
|
296 |
-
"content": "<
|
297 |
"single_word": false,
|
298 |
"lstrip": false,
|
299 |
"rstrip": false,
|
@@ -302,7 +302,7 @@
|
|
302 |
},
|
303 |
{
|
304 |
"id": 57550,
|
305 |
-
"content": "</
|
306 |
"single_word": false,
|
307 |
"lstrip": false,
|
308 |
"rstrip": false,
|
@@ -311,7 +311,7 @@
|
|
311 |
},
|
312 |
{
|
313 |
"id": 57551,
|
314 |
-
"content": "<
|
315 |
"single_word": false,
|
316 |
"lstrip": false,
|
317 |
"rstrip": false,
|
@@ -320,7 +320,7 @@
|
|
320 |
},
|
321 |
{
|
322 |
"id": 57552,
|
323 |
-
"content": "</
|
324 |
"single_word": false,
|
325 |
"lstrip": false,
|
326 |
"rstrip": false,
|
@@ -329,96 +329,6 @@
|
|
329 |
},
|
330 |
{
|
331 |
"id": 57553,
|
332 |
-
"content": "<s_Age_R>",
|
333 |
-
"single_word": false,
|
334 |
-
"lstrip": false,
|
335 |
-
"rstrip": false,
|
336 |
-
"normalized": true,
|
337 |
-
"special": false
|
338 |
-
},
|
339 |
-
{
|
340 |
-
"id": 57554,
|
341 |
-
"content": "</s_Age_R>",
|
342 |
-
"single_word": false,
|
343 |
-
"lstrip": false,
|
344 |
-
"rstrip": false,
|
345 |
-
"normalized": true,
|
346 |
-
"special": false
|
347 |
-
},
|
348 |
-
{
|
349 |
-
"id": 57555,
|
350 |
-
"content": "<s_relation>",
|
351 |
-
"single_word": false,
|
352 |
-
"lstrip": false,
|
353 |
-
"rstrip": false,
|
354 |
-
"normalized": true,
|
355 |
-
"special": false
|
356 |
-
},
|
357 |
-
{
|
358 |
-
"id": 57556,
|
359 |
-
"content": "</s_relation>",
|
360 |
-
"single_word": false,
|
361 |
-
"lstrip": false,
|
362 |
-
"rstrip": false,
|
363 |
-
"normalized": true,
|
364 |
-
"special": false
|
365 |
-
},
|
366 |
-
{
|
367 |
-
"id": 57557,
|
368 |
-
"content": "<s_Mother>",
|
369 |
-
"single_word": false,
|
370 |
-
"lstrip": false,
|
371 |
-
"rstrip": false,
|
372 |
-
"normalized": true,
|
373 |
-
"special": false
|
374 |
-
},
|
375 |
-
{
|
376 |
-
"id": 57558,
|
377 |
-
"content": "</s_Mother>",
|
378 |
-
"single_word": false,
|
379 |
-
"lstrip": false,
|
380 |
-
"rstrip": false,
|
381 |
-
"normalized": true,
|
382 |
-
"special": false
|
383 |
-
},
|
384 |
-
{
|
385 |
-
"id": 57559,
|
386 |
-
"content": "<s_father_R>",
|
387 |
-
"single_word": false,
|
388 |
-
"lstrip": false,
|
389 |
-
"rstrip": false,
|
390 |
-
"normalized": true,
|
391 |
-
"special": false
|
392 |
-
},
|
393 |
-
{
|
394 |
-
"id": 57560,
|
395 |
-
"content": "</s_father_R>",
|
396 |
-
"single_word": false,
|
397 |
-
"lstrip": false,
|
398 |
-
"rstrip": false,
|
399 |
-
"normalized": true,
|
400 |
-
"special": false
|
401 |
-
},
|
402 |
-
{
|
403 |
-
"id": 57561,
|
404 |
-
"content": "<s_father_L>",
|
405 |
-
"single_word": false,
|
406 |
-
"lstrip": false,
|
407 |
-
"rstrip": false,
|
408 |
-
"normalized": true,
|
409 |
-
"special": false
|
410 |
-
},
|
411 |
-
{
|
412 |
-
"id": 57562,
|
413 |
-
"content": "</s_father_L>",
|
414 |
-
"single_word": false,
|
415 |
-
"lstrip": false,
|
416 |
-
"rstrip": false,
|
417 |
-
"normalized": true,
|
418 |
-
"special": false
|
419 |
-
},
|
420 |
-
{
|
421 |
-
"id": 57563,
|
422 |
"content": "<s_cord-v2>",
|
423 |
"single_word": false,
|
424 |
"lstrip": false,
|
|
|
113 |
},
|
114 |
{
|
115 |
"id": 57529,
|
116 |
+
"content": "<s_city>",
|
117 |
"single_word": false,
|
118 |
"lstrip": false,
|
119 |
"rstrip": false,
|
|
|
122 |
},
|
123 |
{
|
124 |
"id": 57530,
|
125 |
+
"content": "</s_city>",
|
126 |
"single_word": false,
|
127 |
"lstrip": false,
|
128 |
"rstrip": false,
|
|
|
131 |
},
|
132 |
{
|
133 |
"id": 57531,
|
134 |
+
"content": "<s_name>",
|
135 |
"single_word": false,
|
136 |
"lstrip": false,
|
137 |
"rstrip": false,
|
|
|
140 |
},
|
141 |
{
|
142 |
"id": 57532,
|
143 |
+
"content": "</s_name>",
|
144 |
"single_word": false,
|
145 |
"lstrip": false,
|
146 |
"rstrip": false,
|
|
|
149 |
},
|
150 |
{
|
151 |
"id": 57533,
|
152 |
+
"content": "<s_death>",
|
153 |
"single_word": false,
|
154 |
"lstrip": false,
|
155 |
"rstrip": false,
|
|
|
158 |
},
|
159 |
{
|
160 |
"id": 57534,
|
161 |
+
"content": "</s_death>",
|
162 |
"single_word": false,
|
163 |
"lstrip": false,
|
164 |
"rstrip": false,
|
|
|
167 |
},
|
168 |
{
|
169 |
"id": 57535,
|
170 |
+
"content": "<s_person_data>",
|
171 |
"single_word": false,
|
172 |
"lstrip": false,
|
173 |
"rstrip": false,
|
|
|
176 |
},
|
177 |
{
|
178 |
"id": 57536,
|
179 |
+
"content": "</s_person_data>",
|
180 |
"single_word": false,
|
181 |
"lstrip": false,
|
182 |
"rstrip": false,
|
|
|
185 |
},
|
186 |
{
|
187 |
"id": 57537,
|
188 |
+
"content": "<s_Gender>",
|
189 |
"single_word": false,
|
190 |
"lstrip": false,
|
191 |
"rstrip": false,
|
|
|
194 |
},
|
195 |
{
|
196 |
"id": 57538,
|
197 |
+
"content": "</s_Gender>",
|
198 |
"single_word": false,
|
199 |
"lstrip": false,
|
200 |
"rstrip": false,
|
|
|
203 |
},
|
204 |
{
|
205 |
"id": 57539,
|
206 |
+
"content": "<s_Race>",
|
207 |
"single_word": false,
|
208 |
"lstrip": false,
|
209 |
"rstrip": false,
|
|
|
212 |
},
|
213 |
{
|
214 |
"id": 57540,
|
215 |
+
"content": "</s_Race>",
|
216 |
"single_word": false,
|
217 |
"lstrip": false,
|
218 |
"rstrip": false,
|
|
|
221 |
},
|
222 |
{
|
223 |
"id": 57541,
|
224 |
+
"content": "<s_birth_day>",
|
225 |
"single_word": false,
|
226 |
"lstrip": false,
|
227 |
"rstrip": false,
|
|
|
230 |
},
|
231 |
{
|
232 |
"id": 57542,
|
233 |
+
"content": "</s_birth_day>",
|
234 |
"single_word": false,
|
235 |
"lstrip": false,
|
236 |
"rstrip": false,
|
|
|
239 |
},
|
240 |
{
|
241 |
"id": 57543,
|
242 |
+
"content": "<s_birthplace>",
|
243 |
"single_word": false,
|
244 |
"lstrip": false,
|
245 |
"rstrip": false,
|
|
|
248 |
},
|
249 |
{
|
250 |
"id": 57544,
|
251 |
+
"content": "</s_birthplace>",
|
252 |
"single_word": false,
|
253 |
"lstrip": false,
|
254 |
"rstrip": false,
|
|
|
257 |
},
|
258 |
{
|
259 |
"id": 57545,
|
260 |
+
"content": "<s_Age>",
|
261 |
"single_word": false,
|
262 |
"lstrip": false,
|
263 |
"rstrip": false,
|
|
|
266 |
},
|
267 |
{
|
268 |
"id": 57546,
|
269 |
+
"content": "</s_Age>",
|
270 |
"single_word": false,
|
271 |
"lstrip": false,
|
272 |
"rstrip": false,
|
|
|
275 |
},
|
276 |
{
|
277 |
"id": 57547,
|
278 |
+
"content": "<s_relation>",
|
279 |
"single_word": false,
|
280 |
"lstrip": false,
|
281 |
"rstrip": false,
|
|
|
284 |
},
|
285 |
{
|
286 |
"id": 57548,
|
287 |
+
"content": "</s_relation>",
|
288 |
"single_word": false,
|
289 |
"lstrip": false,
|
290 |
"rstrip": false,
|
|
|
293 |
},
|
294 |
{
|
295 |
"id": 57549,
|
296 |
+
"content": "<s_Mother>",
|
297 |
"single_word": false,
|
298 |
"lstrip": false,
|
299 |
"rstrip": false,
|
|
|
302 |
},
|
303 |
{
|
304 |
"id": 57550,
|
305 |
+
"content": "</s_Mother>",
|
306 |
"single_word": false,
|
307 |
"lstrip": false,
|
308 |
"rstrip": false,
|
|
|
311 |
},
|
312 |
{
|
313 |
"id": 57551,
|
314 |
+
"content": "<s_father>",
|
315 |
"single_word": false,
|
316 |
"lstrip": false,
|
317 |
"rstrip": false,
|
|
|
320 |
},
|
321 |
{
|
322 |
"id": 57552,
|
323 |
+
"content": "</s_father>",
|
324 |
"single_word": false,
|
325 |
"lstrip": false,
|
326 |
"rstrip": false,
|
|
|
329 |
},
|
330 |
{
|
331 |
"id": 57553,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
332 |
"content": "<s_cord-v2>",
|
333 |
"single_word": false,
|
334 |
"lstrip": false,
|