KoichiYasuoka
commited on
Commit
•
e8d7850
1
Parent(s):
fcffeee
model improved
Browse files- config.json +471 -451
- pytorch_model.bin +2 -2
- special_tokens_map.json +42 -6
- tokenizer.json +2 -1
- tokenizer_config.json +48 -8
config.json
CHANGED
@@ -16,461 +16,481 @@
|
|
16 |
"hidden_size": 768,
|
17 |
"id2label": {
|
18 |
"0": "-|_|dep",
|
19 |
-
"1": "
|
20 |
-
"2": "ADP|\u526f\u52a9\u8a5e|
|
21 |
-
"3": "ADP|\u526f\u52a9\u8a5e|
|
22 |
-
"4": "ADP|\u526f\u52a9\u8a5e|
|
23 |
-
"5": "ADP|\
|
24 |
-
"6": "ADP|\u683c\u52a9\u8a5e|case",
|
25 |
-
"7": "ADP|\u683c\u52a9\u8a5e|
|
26 |
-
"8": "ADP|\u683c\u52a9\u8a5e|
|
27 |
-
"9": "ADP|\u683c\u52a9\u8a5e|
|
28 |
-
"10": "
|
29 |
-
"11": "ADV|\u526f\u8a5e|
|
30 |
-
"12": "ADV|\u526f\u8a5e|
|
31 |
-
"13": "ADV|\u526f\u8a5e|
|
32 |
-
"14": "ADV|\u526f\u8a5e|
|
33 |
-
"15": "ADV|\u526f\u8a5e|
|
34 |
-
"16": "ADV|\u526f\u8a5e|
|
35 |
-
"17": "ADV|\u526f\u8a5e|
|
36 |
-
"18": "
|
37 |
-
"19": "AUX|\u30c7\u30a2\u30eb\u52d5\u8a5e|
|
38 |
-
"20": "AUX|\
|
39 |
-
"21": "AUX|\
|
40 |
-
"22": "AUX|\u52d5\u8a5e
|
41 |
-
"23": "
|
42 |
-
"24": "
|
43 |
-
"25": "CCONJ|\u63a5\u7d9a\u8a5e|
|
44 |
-
"26": "
|
45 |
-
"27": "
|
46 |
-
"28": "DET
|
47 |
-
"29": "DET|\u9023\u4f53\u8a5e|
|
48 |
-
"30": "DET|\u9023\u4f53\u8a5e|
|
49 |
-
"31": "DET|\u9023\u4f53\u8a5e|
|
50 |
-
"32": "DET|\u9023\u4f53\u8a5e|
|
51 |
-
"33": "DET|\u9023\u4f53\u8a5e|
|
52 |
-
"34": "DET|\u9023\u4f53\u8a5e|
|
53 |
-
"35": "DET|\u9023\u4f53\u8a5e|
|
54 |
-
"36": "DET|\u9023\u4f53\u8a5e|
|
55 |
-
"37": "
|
56 |
-
"38": "
|
57 |
-
"39": "
|
58 |
-
"40": "INTJ|\u9593\u6295\u8a5e|
|
59 |
-
"41": "INTJ|\u9593\u6295\u8a5e|
|
60 |
-
"42": "
|
61 |
-
"43": "
|
62 |
-
"44": "
|
63 |
-
"45": "NOUN
|
64 |
-
"46": "NOUN
|
65 |
-
"47": "NOUN
|
66 |
-
"48": "NOUN|\u4f4d\u7f6e\u540d\u8a5e|
|
67 |
-
"49": "NOUN|\u4f4d\u7f6e\u540d\u8a5e|
|
68 |
-
"50": "NOUN|\u4f4d\u7f6e\u540d\u8a5e|
|
69 |
-
"51": "NOUN|\u4f4d\u7f6e\u540d\u8a5e|
|
70 |
-
"52": "NOUN|\
|
71 |
-
"53": "NOUN|\
|
72 |
-
"54": "NOUN|\
|
73 |
-
"55": "NOUN|\
|
74 |
-
"56": "NOUN|\
|
75 |
-
"57": "NOUN|\
|
76 |
-
"58": "NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\
|
77 |
-
"59": "NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\
|
78 |
-
"60": "NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\
|
79 |
-
"61": "NOUN|\u540d\u8a5e|
|
80 |
-
"62": "NOUN|\u540d\u8a5e|
|
81 |
-
"63": "NOUN|\u540d\u8a5e|
|
82 |
-
"64": "NOUN|\u540d\u8a5e|
|
83 |
-
"65": "NOUN|\u540d\u8a5e|
|
84 |
-
"66": "NOUN|\u540d\u8a5e|
|
85 |
-
"67": "NOUN|\u540d\u8a5e|clf",
|
86 |
-
"68": "NOUN|\u540d\u8a5e|
|
87 |
-
"69": "NOUN|\u540d\u8a5e|
|
88 |
-
"70": "NOUN|\u540d\u8a5e|
|
89 |
-
"71": "NOUN|\u540d\u8a5e|
|
90 |
-
"72": "NOUN|\u540d\u8a5e|
|
91 |
-
"73": "NOUN|\u540d\u8a5e|
|
92 |
-
"74": "NOUN|\u540d\u8a5e|
|
93 |
-
"75": "NOUN|\u540d\u8a5e|
|
94 |
-
"76": "NOUN|\u540d\u8a5e|
|
95 |
-
"77": "NOUN|\u540d\u8a5e|
|
96 |
-
"78": "NOUN|\u540d\u8a5e|
|
97 |
-
"79": "NOUN|\u540d\u8a5e|
|
98 |
-
"80": "NOUN|\
|
99 |
-
"81": "NOUN|\
|
100 |
-
"82": "NOUN|\
|
101 |
-
"83": "NOUN|\
|
102 |
-
"84": "NOUN|\
|
103 |
-
"85": "NOUN|\
|
104 |
-
"86": "NOUN|\
|
105 |
-
"87": "NOUN|\u5f62\u5f0f\u540d\u8a5e|
|
106 |
-
"88": "NOUN|\u5f62\u5f0f\u540d\u8a5e|
|
107 |
-
"89": "NOUN|\u5f62\u5f0f\u540d\u8a5e|
|
108 |
-
"90": "NOUN|\
|
109 |
-
"91": "NOUN|\
|
110 |
-
"92": "NOUN|\
|
111 |
-
"93": "
|
112 |
-
"94": "
|
113 |
-
"95": "
|
114 |
-
"96": "
|
115 |
-
"97": "
|
116 |
-
"98": "
|
117 |
-
"99": "
|
118 |
-
"100": "
|
119 |
-
"101": "NUM|\u6570\u8a5e|
|
120 |
-
"102": "NUM|\u6570\u8a5e|
|
121 |
-
"103": "
|
122 |
-
"104": "
|
123 |
-
"105": "
|
124 |
-
"106": "
|
125 |
-
"107": "
|
126 |
-
"108": "
|
127 |
-
"109": "
|
128 |
-
"110": "
|
129 |
-
"111": "PART|\u4eba\u79f0\u63a5\u8f9e|
|
130 |
-
"112": "PART|\u4eba\u79f0\u63a5\u8f9e|
|
131 |
-
"113": "PART|\u4eba\u79f0\u63a5\u8f9e|
|
132 |
-
"114": "PART|\u4eba\u79f0\u63a5\u8f9e|
|
133 |
-
"115": "PART|\u63a5\
|
134 |
-
"116": "PART|\u63a5\
|
135 |
-
"117": "PART|\u63a5\
|
136 |
-
"118": "PART|\u63a5\
|
137 |
-
"119": "PART|\u63a5\
|
138 |
-
"120": "PART|\u63a5\
|
139 |
-
"121": "PART|\u63a5\
|
140 |
-
"122": "PART|\u63a5\
|
141 |
-
"123": "PART|\u63a5\
|
142 |
-
"124": "PART|\u63a5\u5c3e\u8f9e|
|
143 |
-
"125": "PART|\u63a5\u5c3e\u8f9e|
|
144 |
-
"126": "PART|\u63a5\
|
145 |
-
"127": "PART|\u63a5\
|
146 |
-
"128": "PART|\u63a5\
|
147 |
-
"129": "PART|\u63a5\
|
148 |
-
"130": "PART|\
|
149 |
-
"131": "PART|\
|
150 |
-
"132": "PART|\
|
151 |
-
"133": "PART|\
|
152 |
-
"134": "PART|\
|
153 |
-
"135": "
|
154 |
-
"136": "
|
155 |
-
"137": "
|
156 |
-
"138": "
|
157 |
-
"139": "
|
158 |
-
"140": "
|
159 |
-
"141": "
|
160 |
-
"142": "
|
161 |
-
"143": "
|
162 |
-
"144": "PRON|\u4ee3\u540d\u8a5e|
|
163 |
-
"145": "PRON|\u4ee3\u540d\u8a5e|
|
164 |
-
"146": "PRON|\u4ee3\u540d\u8a5e|
|
165 |
-
"147": "
|
166 |
-
"148": "
|
167 |
-
"149": "
|
168 |
-
"150": "
|
169 |
-
"151": "
|
170 |
-
"152": "
|
171 |
-
"153": "
|
172 |
-
"154": "
|
173 |
-
"155": "
|
174 |
-
"156": "PROPN
|
175 |
-
"157": "PROPN|\u56fa\u6709\u540d\u8a5e|
|
176 |
-
"158": "PROPN|\u56fa\u6709\u540d\u8a5e|
|
177 |
-
"159": "
|
178 |
-
"160": "
|
179 |
-
"161": "
|
180 |
-
"162": "
|
181 |
-
"163": "
|
182 |
-
"164": "
|
183 |
-
"165": "
|
184 |
-
"166": "
|
185 |
-
"167": "
|
186 |
-
"168": "
|
187 |
-
"169": "
|
188 |
-
"170": "
|
189 |
-
"171": "
|
190 |
-
"172": "
|
191 |
-
"173": "
|
192 |
-
"174": "
|
193 |
-
"175": "
|
194 |
-
"176": "
|
195 |
-
"177": "
|
196 |
-
"178": "
|
197 |
-
"179": "
|
198 |
-
"180": "
|
199 |
-
"181": "SCONJ|\
|
200 |
-
"182": "SCONJ|\
|
201 |
-
"183": "SCONJ|\
|
202 |
-
"184": "SCONJ|\
|
203 |
-
"185": "SCONJ|\
|
204 |
-
"186": "SCONJ|\
|
205 |
-
"187": "SCONJ|\
|
206 |
-
"188": "SCONJ|\
|
207 |
-
"189": "
|
208 |
-
"190": "
|
209 |
-
"191": "
|
210 |
-
"192": "
|
211 |
-
"193": "
|
212 |
-
"194": "
|
213 |
-
"195": "
|
214 |
-
"196": "
|
215 |
-
"197": "
|
216 |
-
"198": "
|
217 |
-
"199": "
|
218 |
-
"200": "
|
219 |
-
"201": "
|
220 |
-
"202": "VERB|\u4ed6\u52d5\u8a5e|
|
221 |
-
"203": "VERB|\u4ed6\u52d5\u8a5e|
|
222 |
-
"204": "VERB|\
|
223 |
-
"205": "VERB|\
|
224 |
-
"206": "VERB|\
|
225 |
-
"207": "VERB|\
|
226 |
-
"208": "VERB|\
|
227 |
-
"209": "VERB|\
|
228 |
-
"210": "VERB|\
|
229 |
-
"211": "VERB|\
|
230 |
-
"212": "VERB|\
|
231 |
-
"213": "VERB|\
|
232 |
-
"214": "VERB|\
|
233 |
-
"215": "VERB|\
|
234 |
-
"216": "VERB|\
|
235 |
-
"217": "VERB|\
|
236 |
-
"218": "VERB|\
|
237 |
-
"219": "VERB|\
|
238 |
-
"220": "VERB|\
|
239 |
-
"221": "VERB|\u81ea\u52d5\u8a5e|
|
240 |
-
"222": "
|
241 |
-
"223": "
|
242 |
-
"224": "
|
243 |
-
"225": "
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
244 |
},
|
245 |
"initializer_range": 0.02,
|
246 |
"intermediate_size": 3072,
|
247 |
"label2id": {
|
248 |
"-|_|dep": 0,
|
249 |
-
"
|
250 |
-
"ADP|\u526f\u52a9\u8a5e|
|
251 |
-
"ADP|\u526f\u52a9\u8a5e|
|
252 |
-
"ADP|\u526f\u52a9\u8a5e|
|
253 |
-
"ADP|\
|
254 |
-
"ADP|\u683c\u52a9\u8a5e|case": 6,
|
255 |
-
"ADP|\u683c\u52a9\u8a5e|
|
256 |
-
"ADP|\u683c\u52a9\u8a5e|
|
257 |
-
"ADP|\u683c\u52a9\u8a5e|
|
258 |
-
"
|
259 |
-
"ADV|\u526f\u8a5e|
|
260 |
-
"ADV|\u526f\u8a5e|
|
261 |
-
"ADV|\u526f\u8a5e|
|
262 |
-
"ADV|\u526f\u8a5e|
|
263 |
-
"ADV|\u526f\u8a5e|
|
264 |
-
"ADV|\u526f\u8a5e|
|
265 |
-
"ADV|\u526f\u8a5e|
|
266 |
-
"
|
267 |
-
"AUX|\u30c7\u30a2\u30eb\u52d5\u8a5e|
|
268 |
-
"AUX|\
|
269 |
-
"AUX|\
|
270 |
-
"AUX|\u52d5\u8a5e
|
271 |
-
"
|
272 |
-
"
|
273 |
-
"CCONJ|\u63a5\u7d9a\u8a5e|
|
274 |
-
"
|
275 |
-
"
|
276 |
-
"DET
|
277 |
-
"DET|\u9023\u4f53\u8a5e|
|
278 |
-
"DET|\u9023\u4f53\u8a5e|
|
279 |
-
"DET|\u9023\u4f53\u8a5e|
|
280 |
-
"DET|\u9023\u4f53\u8a5e|
|
281 |
-
"DET|\u9023\u4f53\u8a5e|
|
282 |
-
"DET|\u9023\u4f53\u8a5e|
|
283 |
-
"DET|\u9023\u4f53\u8a5e|
|
284 |
-
"DET|\u9023\u4f53\u8a5e|
|
285 |
-
"
|
286 |
-
"
|
287 |
-
"
|
288 |
-
"INTJ|\u9593\u6295\u8a5e|
|
289 |
-
"INTJ|\u9593\u6295\u8a5e|
|
290 |
-
"
|
291 |
-
"
|
292 |
-
"
|
293 |
-
"NOUN
|
294 |
-
"NOUN
|
295 |
-
"NOUN
|
296 |
-
"NOUN|\u4f4d\u7f6e\u540d\u8a5e|
|
297 |
-
"NOUN|\u4f4d\u7f6e\u540d\u8a5e|
|
298 |
-
"NOUN|\u4f4d\u7f6e\u540d\u8a5e|
|
299 |
-
"NOUN|\u4f4d\u7f6e\u540d\u8a5e|
|
300 |
-
"NOUN|\
|
301 |
-
"NOUN|\
|
302 |
-
"NOUN|\
|
303 |
-
"NOUN|\
|
304 |
-
"NOUN|\
|
305 |
-
"NOUN|\
|
306 |
-
"NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\
|
307 |
-
"NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\
|
308 |
-
"NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\
|
309 |
-
"NOUN|\u540d\u8a5e|
|
310 |
-
"NOUN|\u540d\u8a5e|
|
311 |
-
"NOUN|\u540d\u8a5e|
|
312 |
-
"NOUN|\u540d\u8a5e|
|
313 |
-
"NOUN|\u540d\u8a5e|
|
314 |
-
"NOUN|\u540d\u8a5e|
|
315 |
-
"NOUN|\u540d\u8a5e|clf": 67,
|
316 |
-
"NOUN|\u540d\u8a5e|
|
317 |
-
"NOUN|\u540d\u8a5e|
|
318 |
-
"NOUN|\u540d\u8a5e|
|
319 |
-
"NOUN|\u540d\u8a5e|
|
320 |
-
"NOUN|\u540d\u8a5e|
|
321 |
-
"NOUN|\u540d\u8a5e|
|
322 |
-
"NOUN|\u540d\u8a5e|
|
323 |
-
"NOUN|\u540d\u8a5e|
|
324 |
-
"NOUN|\u540d\u8a5e|
|
325 |
-
"NOUN|\u540d\u8a5e|
|
326 |
-
"NOUN|\u540d\u8a5e|
|
327 |
-
"NOUN|\u540d\u8a5e|
|
328 |
-
"NOUN|\
|
329 |
-
"NOUN|\
|
330 |
-
"NOUN|\
|
331 |
-
"NOUN|\
|
332 |
-
"NOUN|\
|
333 |
-
"NOUN|\
|
334 |
-
"NOUN|\
|
335 |
-
"NOUN|\u5f62\u5f0f\u540d\u8a5e|
|
336 |
-
"NOUN|\u5f62\u5f0f\u540d\u8a5e|
|
337 |
-
"NOUN|\u5f62\u5f0f\u540d\u8a5e|
|
338 |
-
"NOUN|\
|
339 |
-
"NOUN|\
|
340 |
-
"NOUN|\
|
341 |
-
"
|
342 |
-
"
|
343 |
-
"
|
344 |
-
"
|
345 |
-
"
|
346 |
-
"
|
347 |
-
"
|
348 |
-
"
|
349 |
-
"NUM|\u6570\u8a5e|
|
350 |
-
"NUM|\u6570\u8a5e|
|
351 |
-
"
|
352 |
-
"
|
353 |
-
"
|
354 |
-
"
|
355 |
-
"
|
356 |
-
"
|
357 |
-
"
|
358 |
-
"
|
359 |
-
"PART|\u4eba\u79f0\u63a5\u8f9e|
|
360 |
-
"PART|\u4eba\u79f0\u63a5\u8f9e|
|
361 |
-
"PART|\u4eba\u79f0\u63a5\u8f9e|
|
362 |
-
"PART|\u4eba\u79f0\u63a5\u8f9e|
|
363 |
-
"PART|\u63a5\
|
364 |
-
"PART|\u63a5\
|
365 |
-
"PART|\u63a5\
|
366 |
-
"PART|\u63a5\
|
367 |
-
"PART|\u63a5\
|
368 |
-
"PART|\u63a5\
|
369 |
-
"PART|\u63a5\
|
370 |
-
"PART|\u63a5\
|
371 |
-
"PART|\u63a5\
|
372 |
-
"PART|\u63a5\u5c3e\u8f9e|
|
373 |
-
"PART|\u63a5\u5c3e\u8f9e|
|
374 |
-
"PART|\u63a5\
|
375 |
-
"PART|\u63a5\
|
376 |
-
"PART|\u63a5\
|
377 |
-
"PART|\u63a5\
|
378 |
-
"PART|\
|
379 |
-
"PART|\
|
380 |
-
"PART|\
|
381 |
-
"PART|\
|
382 |
-
"PART|\
|
383 |
-
"
|
384 |
-
"
|
385 |
-
"
|
386 |
-
"
|
387 |
-
"
|
388 |
-
"
|
389 |
-
"
|
390 |
-
"
|
391 |
-
"
|
392 |
-
"PRON|\u4ee3\u540d\u8a5e|
|
393 |
-
"PRON|\u4ee3\u540d\u8a5e|
|
394 |
-
"PRON|\u4ee3\u540d\u8a5e|
|
395 |
-
"
|
396 |
-
"
|
397 |
-
"
|
398 |
-
"
|
399 |
-
"
|
400 |
-
"
|
401 |
-
"
|
402 |
-
"
|
403 |
-
"
|
404 |
-
"PROPN
|
405 |
-
"PROPN|\u56fa\u6709\u540d\u8a5e|
|
406 |
-
"PROPN|\u56fa\u6709\u540d\u8a5e|
|
407 |
-
"
|
408 |
-
"
|
409 |
-
"
|
410 |
-
"
|
411 |
-
"
|
412 |
-
"
|
413 |
-
"
|
414 |
-
"
|
415 |
-
"
|
416 |
-
"
|
417 |
-
"
|
418 |
-
"
|
419 |
-
"
|
420 |
-
"
|
421 |
-
"
|
422 |
-
"
|
423 |
-
"
|
424 |
-
"
|
425 |
-
"
|
426 |
-
"
|
427 |
-
"
|
428 |
-
"
|
429 |
-
"SCONJ|\
|
430 |
-
"SCONJ|\
|
431 |
-
"SCONJ|\
|
432 |
-
"SCONJ|\
|
433 |
-
"SCONJ|\
|
434 |
-
"SCONJ|\
|
435 |
-
"SCONJ|\
|
436 |
-
"SCONJ|\
|
437 |
-
"
|
438 |
-
"
|
439 |
-
"
|
440 |
-
"
|
441 |
-
"
|
442 |
-
"
|
443 |
-
"
|
444 |
-
"
|
445 |
-
"
|
446 |
-
"
|
447 |
-
"
|
448 |
-
"
|
449 |
-
"
|
450 |
-
"VERB|\u4ed6\u52d5\u8a5e|
|
451 |
-
"VERB|\u4ed6\u52d5\u8a5e|
|
452 |
-
"VERB|\
|
453 |
-
"VERB|\
|
454 |
-
"VERB|\
|
455 |
-
"VERB|\
|
456 |
-
"VERB|\
|
457 |
-
"VERB|\
|
458 |
-
"VERB|\
|
459 |
-
"VERB|\
|
460 |
-
"VERB|\
|
461 |
-
"VERB|\
|
462 |
-
"VERB|\
|
463 |
-
"VERB|\
|
464 |
-
"VERB|\
|
465 |
-
"VERB|\
|
466 |
-
"VERB|\
|
467 |
-
"VERB|\
|
468 |
-
"VERB|\
|
469 |
-
"VERB|\u81ea\u52d5\u8a5e|
|
470 |
-
"
|
471 |
-
"
|
472 |
-
"
|
473 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
474 |
},
|
475 |
"layer_norm_eps": 1e-12,
|
476 |
"max_position_embeddings": 512,
|
@@ -481,7 +501,7 @@
|
|
481 |
"position_embedding_type": "absolute",
|
482 |
"tokenizer_class": "RemBertTokenizerFast",
|
483 |
"torch_dtype": "float32",
|
484 |
-
"transformers_version": "4.
|
485 |
"type_vocab_size": 2,
|
486 |
"use_cache": true,
|
487 |
"vocab_size": 6877
|
|
|
16 |
"hidden_size": 768,
|
17 |
"id2label": {
|
18 |
"0": "-|_|dep",
|
19 |
+
"1": "ADJ|_|amod",
|
20 |
+
"2": "ADP|\u526f\u52a9\u8a5e|case",
|
21 |
+
"3": "ADP|\u526f\u52a9\u8a5e|mark",
|
22 |
+
"4": "ADP|\u526f\u52a9\u8a5e|nsubj",
|
23 |
+
"5": "ADP|\u526f\u52a9\u8a5e|root",
|
24 |
+
"6": "ADP|\u52a9\u8a5e-\u683c\u52a9\u8a5e|case",
|
25 |
+
"7": "ADP|\u683c\u52a9\u8a5e|case",
|
26 |
+
"8": "ADP|\u683c\u52a9\u8a5e|mark",
|
27 |
+
"9": "ADP|\u683c\u52a9\u8a5e|parataxis",
|
28 |
+
"10": "ADP|\u683c\u52a9\u8a5e|root",
|
29 |
+
"11": "ADV|\u526f\u8a5e|acl",
|
30 |
+
"12": "ADV|\u526f\u8a5e|advcl",
|
31 |
+
"13": "ADV|\u526f\u8a5e|advmod",
|
32 |
+
"14": "ADV|\u526f\u8a5e|amod",
|
33 |
+
"15": "ADV|\u526f\u8a5e|conj",
|
34 |
+
"16": "ADV|\u526f\u8a5e|discourse",
|
35 |
+
"17": "ADV|\u526f\u8a5e|parataxis",
|
36 |
+
"18": "ADV|\u526f\u8a5e|root",
|
37 |
+
"19": "AUX|\u30c7\u30a2\u30eb\u52d5\u8a5e|aux",
|
38 |
+
"20": "AUX|\u30c7\u30a2\u30eb\u52d5\u8a5e|case",
|
39 |
+
"21": "AUX|\u30c7\u30a2\u30eb\u52d5\u8a5e|cop",
|
40 |
+
"22": "AUX|\u52a9\u52d5\u8a5e|aux",
|
41 |
+
"23": "AUX|\u52a9\u52d5\u8a5e|orphan",
|
42 |
+
"24": "AUX|\u52d5\u8a5e-\u975e\u81ea\u7acb\u53ef\u80fd|aux",
|
43 |
+
"25": "CCONJ|\u63a5\u7d9a\u8a5e|cc",
|
44 |
+
"26": "CCONJ|\u63a5\u7d9a\u8a5e|mark",
|
45 |
+
"27": "CCONJ|\u63a5\u7d9a\u8a5e|root",
|
46 |
+
"28": "DET|_|det",
|
47 |
+
"29": "DET|\u9023\u4f53\u8a5e|advcl",
|
48 |
+
"30": "DET|\u9023\u4f53\u8a5e|advmod",
|
49 |
+
"31": "DET|\u9023\u4f53\u8a5e|amod",
|
50 |
+
"32": "DET|\u9023\u4f53\u8a5e|cop",
|
51 |
+
"33": "DET|\u9023\u4f53\u8a5e|det",
|
52 |
+
"34": "DET|\u9023\u4f53\u8a5e|nmod",
|
53 |
+
"35": "DET|\u9023\u4f53\u8a5e|nsubj",
|
54 |
+
"36": "DET|\u9023\u4f53\u8a5e|obj",
|
55 |
+
"37": "DET|\u9023\u4f53\u8a5e|obl",
|
56 |
+
"38": "DET|\u9023\u4f53\u8a5e|parataxis",
|
57 |
+
"39": "DET|\u9023\u4f53\u8a5e|root",
|
58 |
+
"40": "INTJ|\u9593\u6295\u8a5e|advcl",
|
59 |
+
"41": "INTJ|\u9593\u6295\u8a5e|conj",
|
60 |
+
"42": "INTJ|\u9593\u6295\u8a5e|discourse",
|
61 |
+
"43": "INTJ|\u9593\u6295\u8a5e|parataxis",
|
62 |
+
"44": "INTJ|\u9593\u6295\u8a5e|root",
|
63 |
+
"45": "NOUN|_|advcl",
|
64 |
+
"46": "NOUN|_|ccomp",
|
65 |
+
"47": "NOUN|_|nmod",
|
66 |
+
"48": "NOUN|\u4f4d\u7f6e\u540d\u8a5e|advcl",
|
67 |
+
"49": "NOUN|\u4f4d\u7f6e\u540d\u8a5e|conj",
|
68 |
+
"50": "NOUN|\u4f4d\u7f6e\u540d\u8a5e|discourse",
|
69 |
+
"51": "NOUN|\u4f4d\u7f6e\u540d\u8a5e|iobj",
|
70 |
+
"52": "NOUN|\u4f4d\u7f6e\u540d\u8a5e|nmod",
|
71 |
+
"53": "NOUN|\u4f4d\u7f6e\u540d\u8a5e|nsubj",
|
72 |
+
"54": "NOUN|\u4f4d\u7f6e\u540d\u8a5e|obj",
|
73 |
+
"55": "NOUN|\u4f4d\u7f6e\u540d\u8a5e|obl",
|
74 |
+
"56": "NOUN|\u4f4d\u7f6e\u540d\u8a5e|parataxis",
|
75 |
+
"57": "NOUN|\u4f4d\u7f6e\u540d\u8a5e|root",
|
76 |
+
"58": "NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u30b5\u5909\u53ef\u80fd|compound",
|
77 |
+
"59": "NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u30b5\u5909\u53ef\u80fd|obj",
|
78 |
+
"60": "NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u4e00\u822c|advcl",
|
79 |
+
"61": "NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u4e00\u822c|compound",
|
80 |
+
"62": "NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u4e00\u822c|conj",
|
81 |
+
"63": "NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u4e00\u822c|nmod",
|
82 |
+
"64": "NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u4e00\u822c|obl",
|
83 |
+
"65": "NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u4e00\u822c|parataxis",
|
84 |
+
"66": "NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u52a9\u6570\u8a5e\u53ef\u80fd|appos",
|
85 |
+
"67": "NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u52a9\u6570\u8a5e\u53ef\u80fd|clf",
|
86 |
+
"68": "NOUN|\u540d\u8a5e|acl",
|
87 |
+
"69": "NOUN|\u540d\u8a5e|advcl",
|
88 |
+
"70": "NOUN|\u540d\u8a5e|amod",
|
89 |
+
"71": "NOUN|\u540d\u8a5e|appos",
|
90 |
+
"72": "NOUN|\u540d\u8a5e|case",
|
91 |
+
"73": "NOUN|\u540d\u8a5e|ccomp",
|
92 |
+
"74": "NOUN|\u540d\u8a5e|clf",
|
93 |
+
"75": "NOUN|\u540d\u8a5e|compound",
|
94 |
+
"76": "NOUN|\u540d\u8a5e|conj",
|
95 |
+
"77": "NOUN|\u540d\u8a5e|dislocated",
|
96 |
+
"78": "NOUN|\u540d\u8a5e|expl",
|
97 |
+
"79": "NOUN|\u540d\u8a5e|iobj",
|
98 |
+
"80": "NOUN|\u540d\u8a5e|nmod",
|
99 |
+
"81": "NOUN|\u540d\u8a5e|nsubj",
|
100 |
+
"82": "NOUN|\u540d\u8a5e|obj",
|
101 |
+
"83": "NOUN|\u540d\u8a5e|obl",
|
102 |
+
"84": "NOUN|\u540d\u8a5e|parataxis",
|
103 |
+
"85": "NOUN|\u540d\u8a5e|root",
|
104 |
+
"86": "NOUN|\u540d\u8a5e|vocative",
|
105 |
+
"87": "NOUN|\u5f62\u5f0f\u540d\u8a5e|acl",
|
106 |
+
"88": "NOUN|\u5f62\u5f0f\u540d\u8a5e|advcl",
|
107 |
+
"89": "NOUN|\u5f62\u5f0f\u540d\u8a5e|conj",
|
108 |
+
"90": "NOUN|\u5f62\u5f0f\u540d\u8a5e|iobj",
|
109 |
+
"91": "NOUN|\u5f62\u5f0f\u540d\u8a5e|nmod",
|
110 |
+
"92": "NOUN|\u5f62\u5f0f\u540d\u8a5e|nsubj",
|
111 |
+
"93": "NOUN|\u5f62\u5f0f\u540d\u8a5e|obj",
|
112 |
+
"94": "NOUN|\u5f62\u5f0f\u540d\u8a5e|obl",
|
113 |
+
"95": "NOUN|\u5f62\u5f0f\u540d\u8a5e|parataxis",
|
114 |
+
"96": "NOUN|\u5f62\u5f0f\u540d\u8a5e|root",
|
115 |
+
"97": "NOUN|\u63a5\u5c3e\u8f9e-\u540d\u8a5e\u7684-\u4e00\u822c|compound",
|
116 |
+
"98": "NOUN|\u63a5\u5c3e\u8f9e-\u540d\u8a5e\u7684-\u4e00\u822c|nmod",
|
117 |
+
"99": "NOUN|\u63a5\u5c3e\u8f9e-\u540d\u8a5e\u7684-\u4e00\u822c|nsubj",
|
118 |
+
"100": "NOUN|\u63a5\u982d\u8f9e-\u4e00\u822c|compound",
|
119 |
+
"101": "NUM|\u540d\u8a5e-\u6570\u8a5e|nummod",
|
120 |
+
"102": "NUM|\u6570\u8a5e|acl",
|
121 |
+
"103": "NUM|\u6570\u8a5e|advcl",
|
122 |
+
"104": "NUM|\u6570\u8a5e|case",
|
123 |
+
"105": "NUM|\u6570\u8a5e|mark",
|
124 |
+
"106": "NUM|\u6570\u8a5e|nmod",
|
125 |
+
"107": "NUM|\u6570\u8a5e|nsubj",
|
126 |
+
"108": "NUM|\u6570\u8a5e|nummod",
|
127 |
+
"109": "NUM|\u6570\u8a5e|obj",
|
128 |
+
"110": "NUM|\u6570\u8a5e|root",
|
129 |
+
"111": "PART|\u4eba\u79f0\u63a5\u8f9e|advcl",
|
130 |
+
"112": "PART|\u4eba\u79f0\u63a5\u8f9e|appos",
|
131 |
+
"113": "PART|\u4eba\u79f0\u63a5\u8f9e|case",
|
132 |
+
"114": "PART|\u4eba\u79f0\u63a5\u8f9e|det",
|
133 |
+
"115": "PART|\u4eba\u79f0\u63a5\u8f9e|expl",
|
134 |
+
"116": "PART|\u4eba\u79f0\u63a5\u8f9e|iobj",
|
135 |
+
"117": "PART|\u4eba\u79f0\u63a5\u8f9e|mark",
|
136 |
+
"118": "PART|\u4eba\u79f0\u63a5\u8f9e|nmod",
|
137 |
+
"119": "PART|\u4eba\u79f0\u63a5\u8f9e|nsubj",
|
138 |
+
"120": "PART|\u4eba\u79f0\u63a5\u8f9e|obj",
|
139 |
+
"121": "PART|\u4eba\u79f0\u63a5\u8f9e|obl",
|
140 |
+
"122": "PART|\u4eba\u79f0\u63a5\u8f9e|parataxis",
|
141 |
+
"123": "PART|\u4eba\u79f0\u63a5\u8f9e|root",
|
142 |
+
"124": "PART|\u63a5\u5c3e\u8f9e|acl",
|
143 |
+
"125": "PART|\u63a5\u5c3e\u8f9e|advcl",
|
144 |
+
"126": "PART|\u63a5\u5c3e\u8f9e|conj",
|
145 |
+
"127": "PART|\u63a5\u5c3e\u8f9e|fixed",
|
146 |
+
"128": "PART|\u63a5\u5c3e\u8f9e|mark",
|
147 |
+
"129": "PART|\u63a5\u5c3e\u8f9e|nmod",
|
148 |
+
"130": "PART|\u63a5\u5c3e\u8f9e|nsubj",
|
149 |
+
"131": "PART|\u63a5\u5c3e\u8f9e|obj",
|
150 |
+
"132": "PART|\u63a5\u5c3e\u8f9e|obl",
|
151 |
+
"133": "PART|\u63a5\u5c3e\u8f9e|parataxis",
|
152 |
+
"134": "PART|\u63a5\u5c3e\u8f9e|root",
|
153 |
+
"135": "PART|\u63a5\u982d\u8f9e|compound",
|
154 |
+
"136": "PART|\u63a5\u982d\u8f9e|det",
|
155 |
+
"137": "PART|\u63a5\u982d\u8f9e|iobj",
|
156 |
+
"138": "PART|\u63a5\u982d\u8f9e|parataxis",
|
157 |
+
"139": "PART|\u7d42\u52a9\u8a5e|discourse",
|
158 |
+
"140": "PART|\u7d42\u52a9\u8a5e|expl",
|
159 |
+
"141": "PART|\u7d42\u52a9\u8a5e|mark",
|
160 |
+
"142": "PART|\u7d42\u52a9\u8a5e|nsubj",
|
161 |
+
"143": "PART|\u7d42\u52a9\u8a5e|root",
|
162 |
+
"144": "PRON|\u4ee3\u540d\u8a5e|advcl",
|
163 |
+
"145": "PRON|\u4ee3\u540d\u8a5e|case",
|
164 |
+
"146": "PRON|\u4ee3\u540d\u8a5e|compound",
|
165 |
+
"147": "PRON|\u4ee3\u540d\u8a5e|det",
|
166 |
+
"148": "PRON|\u4ee3\u540d\u8a5e|dislocated",
|
167 |
+
"149": "PRON|\u4ee3\u540d\u8a5e|nmod",
|
168 |
+
"150": "PRON|\u4ee3\u540d\u8a5e|nsubj",
|
169 |
+
"151": "PRON|\u4ee3\u540d\u8a5e|obj",
|
170 |
+
"152": "PRON|\u4ee3\u540d\u8a5e|obl",
|
171 |
+
"153": "PRON|\u4ee3\u540d\u8a5e|parataxis",
|
172 |
+
"154": "PRON|\u4ee3\u540d\u8a5e|root",
|
173 |
+
"155": "PRON|\u4ee3\u540d\u8a5e|vocative",
|
174 |
+
"156": "PROPN|_|nmod",
|
175 |
+
"157": "PROPN|\u540d\u8a5e-\u56fa\u6709\u540d\u8a5e-\u4e00\u822c|compound",
|
176 |
+
"158": "PROPN|\u540d\u8a5e-\u56fa\u6709\u540d\u8a5e-\u4eba\u540d-\u540d|conj",
|
177 |
+
"159": "PROPN|\u540d\u8a5e-\u56fa\u6709\u540d\u8a5e-\u4eba\u540d-\u540d|flat",
|
178 |
+
"160": "PROPN|\u540d\u8a5e-\u56fa\u6709\u540d\u8a5e-\u4eba\u540d-\u59d3|nsubj",
|
179 |
+
"161": "PROPN|\u540d\u8a5e-\u56fa\u6709\u540d\u8a5e-\u5730\u540d-\u56fd|compound",
|
180 |
+
"162": "PROPN|\u56fa\u6709\u540d\u8a5e|advcl",
|
181 |
+
"163": "PROPN|\u56fa\u6709\u540d\u8a5e|appos",
|
182 |
+
"164": "PROPN|\u56fa\u6709\u540d\u8a5e|conj",
|
183 |
+
"165": "PROPN|\u56fa\u6709\u540d\u8a5e|flat",
|
184 |
+
"166": "PROPN|\u56fa\u6709\u540d\u8a5e|nmod",
|
185 |
+
"167": "PROPN|\u56fa\u6709\u540d\u8a5e|nsubj",
|
186 |
+
"168": "PROPN|\u56fa\u6709\u540d\u8a5e|obj",
|
187 |
+
"169": "PROPN|\u56fa\u6709\u540d\u8a5e|obl",
|
188 |
+
"170": "PROPN|\u56fa\u6709\u540d\u8a5e|root",
|
189 |
+
"171": "PUNCT|\u88dc\u52a9\u8a18\u53f7-\u62ec\u5f27\u9589|punct",
|
190 |
+
"172": "PUNCT|\u88dc\u52a9\u8a18\u53f7-\u62ec\u5f27\u958b|punct",
|
191 |
+
"173": "PUNCT|\u8a18\u53f7|case",
|
192 |
+
"174": "PUNCT|\u8a18\u53f7|conj",
|
193 |
+
"175": "PUNCT|\u8a18\u53f7|nmod",
|
194 |
+
"176": "PUNCT|\u8a18\u53f7|nsubj",
|
195 |
+
"177": "PUNCT|\u8a18\u53f7|obj",
|
196 |
+
"178": "PUNCT|\u8a18\u53f7|obl",
|
197 |
+
"179": "PUNCT|\u8a18\u53f7|punct",
|
198 |
+
"180": "PUNCT|\u8a18\u53f7|root",
|
199 |
+
"181": "SCONJ|\u5f8c\u7f6e\u526f\u8a5e|acl",
|
200 |
+
"182": "SCONJ|\u5f8c\u7f6e\u526f\u8a5e|advcl",
|
201 |
+
"183": "SCONJ|\u5f8c\u7f6e\u526f\u8a5e|advmod",
|
202 |
+
"184": "SCONJ|\u5f8c\u7f6e\u526f\u8a5e|amod",
|
203 |
+
"185": "SCONJ|\u5f8c\u7f6e\u526f\u8a5e|case",
|
204 |
+
"186": "SCONJ|\u5f8c\u7f6e\u526f\u8a5e|mark",
|
205 |
+
"187": "SCONJ|\u5f8c\u7f6e\u526f\u8a5e|parataxis",
|
206 |
+
"188": "SCONJ|\u5f8c\u7f6e\u526f\u8a5e|root",
|
207 |
+
"189": "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|advmod",
|
208 |
+
"190": "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|case",
|
209 |
+
"191": "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|mark",
|
210 |
+
"192": "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|nsubj",
|
211 |
+
"193": "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|obj",
|
212 |
+
"194": "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|parataxis",
|
213 |
+
"195": "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|punct",
|
214 |
+
"196": "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|root",
|
215 |
+
"197": "SCONJ|\u63a5\u7d9a\u8a5e|advmod",
|
216 |
+
"198": "SCONJ|\u63a5\u7d9a\u8a5e|case",
|
217 |
+
"199": "SCONJ|\u63a5\u7d9a\u8a5e|mark",
|
218 |
+
"200": "SYM|_|conj",
|
219 |
+
"201": "SYM|_|obj",
|
220 |
+
"202": "VERB|\u4ed6\u52d5\u8a5e|acl",
|
221 |
+
"203": "VERB|\u4ed6\u52d5\u8a5e|advcl",
|
222 |
+
"204": "VERB|\u4ed6\u52d5\u8a5e|amod",
|
223 |
+
"205": "VERB|\u4ed6\u52d5\u8a5e|appos",
|
224 |
+
"206": "VERB|\u4ed6\u52d5\u8a5e|case",
|
225 |
+
"207": "VERB|\u4ed6\u52d5\u8a5e|ccomp",
|
226 |
+
"208": "VERB|\u4ed6\u52d5\u8a5e|conj",
|
227 |
+
"209": "VERB|\u4ed6\u52d5\u8a5e|csubj",
|
228 |
+
"210": "VERB|\u4ed6\u52d5\u8a5e|mark",
|
229 |
+
"211": "VERB|\u4ed6\u52d5\u8a5e|obj",
|
230 |
+
"212": "VERB|\u4ed6\u52d5\u8a5e|obl",
|
231 |
+
"213": "VERB|\u4ed6\u52d5\u8a5e|parataxis",
|
232 |
+
"214": "VERB|\u4ed6\u52d5\u8a5e|root",
|
233 |
+
"215": "VERB|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u30b5\u5909\u53ef\u80fd|acl",
|
234 |
+
"216": "VERB|\u5b8c\u5168\u52d5\u8a5e|acl",
|
235 |
+
"217": "VERB|\u5b8c\u5168\u52d5\u8a5e|advcl",
|
236 |
+
"218": "VERB|\u5b8c\u5168\u52d5\u8a5e|amod",
|
237 |
+
"219": "VERB|\u5b8c\u5168\u52d5\u8a5e|parataxis",
|
238 |
+
"220": "VERB|\u5b8c\u5168\u52d5\u8a5e|root",
|
239 |
+
"221": "VERB|\u81ea\u52d5\u8a5e|acl",
|
240 |
+
"222": "VERB|\u81ea\u52d5\u8a5e|advcl",
|
241 |
+
"223": "VERB|\u81ea\u52d5\u8a5e|amod",
|
242 |
+
"224": "VERB|\u81ea\u52d5\u8a5e|ccomp",
|
243 |
+
"225": "VERB|\u81ea\u52d5\u8a5e|conj",
|
244 |
+
"226": "VERB|\u81ea\u52d5\u8a5e|csubj",
|
245 |
+
"227": "VERB|\u81ea\u52d5\u8a5e|det",
|
246 |
+
"228": "VERB|\u81ea\u52d5\u8a5e|mark",
|
247 |
+
"229": "VERB|\u81ea\u52d5\u8a5e|nmod",
|
248 |
+
"230": "VERB|\u81ea\u52d5\u8a5e|nsubj",
|
249 |
+
"231": "VERB|\u81ea\u52d5\u8a5e|parataxis",
|
250 |
+
"232": "VERB|\u81ea\u52d5\u8a5e|root",
|
251 |
+
"233": "X|_|conj",
|
252 |
+
"234": "X|_|goeswith",
|
253 |
+
"235": "X|_|nsubj"
|
254 |
},
|
255 |
"initializer_range": 0.02,
|
256 |
"intermediate_size": 3072,
|
257 |
"label2id": {
|
258 |
"-|_|dep": 0,
|
259 |
+
"ADJ|_|amod": 1,
|
260 |
+
"ADP|\u526f\u52a9\u8a5e|case": 2,
|
261 |
+
"ADP|\u526f\u52a9\u8a5e|mark": 3,
|
262 |
+
"ADP|\u526f\u52a9\u8a5e|nsubj": 4,
|
263 |
+
"ADP|\u526f\u52a9\u8a5e|root": 5,
|
264 |
+
"ADP|\u52a9\u8a5e-\u683c\u52a9\u8a5e|case": 6,
|
265 |
+
"ADP|\u683c\u52a9\u8a5e|case": 7,
|
266 |
+
"ADP|\u683c\u52a9\u8a5e|mark": 8,
|
267 |
+
"ADP|\u683c\u52a9\u8a5e|parataxis": 9,
|
268 |
+
"ADP|\u683c\u52a9\u8a5e|root": 10,
|
269 |
+
"ADV|\u526f\u8a5e|acl": 11,
|
270 |
+
"ADV|\u526f\u8a5e|advcl": 12,
|
271 |
+
"ADV|\u526f\u8a5e|advmod": 13,
|
272 |
+
"ADV|\u526f\u8a5e|amod": 14,
|
273 |
+
"ADV|\u526f\u8a5e|conj": 15,
|
274 |
+
"ADV|\u526f\u8a5e|discourse": 16,
|
275 |
+
"ADV|\u526f\u8a5e|parataxis": 17,
|
276 |
+
"ADV|\u526f\u8a5e|root": 18,
|
277 |
+
"AUX|\u30c7\u30a2\u30eb\u52d5\u8a5e|aux": 19,
|
278 |
+
"AUX|\u30c7\u30a2\u30eb\u52d5\u8a5e|case": 20,
|
279 |
+
"AUX|\u30c7\u30a2\u30eb\u52d5\u8a5e|cop": 21,
|
280 |
+
"AUX|\u52a9\u52d5\u8a5e|aux": 22,
|
281 |
+
"AUX|\u52a9\u52d5\u8a5e|orphan": 23,
|
282 |
+
"AUX|\u52d5\u8a5e-\u975e\u81ea\u7acb\u53ef\u80fd|aux": 24,
|
283 |
+
"CCONJ|\u63a5\u7d9a\u8a5e|cc": 25,
|
284 |
+
"CCONJ|\u63a5\u7d9a\u8a5e|mark": 26,
|
285 |
+
"CCONJ|\u63a5\u7d9a\u8a5e|root": 27,
|
286 |
+
"DET|_|det": 28,
|
287 |
+
"DET|\u9023\u4f53\u8a5e|advcl": 29,
|
288 |
+
"DET|\u9023\u4f53\u8a5e|advmod": 30,
|
289 |
+
"DET|\u9023\u4f53\u8a5e|amod": 31,
|
290 |
+
"DET|\u9023\u4f53\u8a5e|cop": 32,
|
291 |
+
"DET|\u9023\u4f53\u8a5e|det": 33,
|
292 |
+
"DET|\u9023\u4f53\u8a5e|nmod": 34,
|
293 |
+
"DET|\u9023\u4f53\u8a5e|nsubj": 35,
|
294 |
+
"DET|\u9023\u4f53\u8a5e|obj": 36,
|
295 |
+
"DET|\u9023\u4f53\u8a5e|obl": 37,
|
296 |
+
"DET|\u9023\u4f53\u8a5e|parataxis": 38,
|
297 |
+
"DET|\u9023\u4f53\u8a5e|root": 39,
|
298 |
+
"INTJ|\u9593\u6295\u8a5e|advcl": 40,
|
299 |
+
"INTJ|\u9593\u6295\u8a5e|conj": 41,
|
300 |
+
"INTJ|\u9593\u6295\u8a5e|discourse": 42,
|
301 |
+
"INTJ|\u9593\u6295\u8a5e|parataxis": 43,
|
302 |
+
"INTJ|\u9593\u6295\u8a5e|root": 44,
|
303 |
+
"NOUN|_|advcl": 45,
|
304 |
+
"NOUN|_|ccomp": 46,
|
305 |
+
"NOUN|_|nmod": 47,
|
306 |
+
"NOUN|\u4f4d\u7f6e\u540d\u8a5e|advcl": 48,
|
307 |
+
"NOUN|\u4f4d\u7f6e\u540d\u8a5e|conj": 49,
|
308 |
+
"NOUN|\u4f4d\u7f6e\u540d\u8a5e|discourse": 50,
|
309 |
+
"NOUN|\u4f4d\u7f6e\u540d\u8a5e|iobj": 51,
|
310 |
+
"NOUN|\u4f4d\u7f6e\u540d\u8a5e|nmod": 52,
|
311 |
+
"NOUN|\u4f4d\u7f6e\u540d\u8a5e|nsubj": 53,
|
312 |
+
"NOUN|\u4f4d\u7f6e\u540d\u8a5e|obj": 54,
|
313 |
+
"NOUN|\u4f4d\u7f6e\u540d\u8a5e|obl": 55,
|
314 |
+
"NOUN|\u4f4d\u7f6e\u540d\u8a5e|parataxis": 56,
|
315 |
+
"NOUN|\u4f4d\u7f6e\u540d\u8a5e|root": 57,
|
316 |
+
"NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u30b5\u5909\u53ef\u80fd|compound": 58,
|
317 |
+
"NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u30b5\u5909\u53ef\u80fd|obj": 59,
|
318 |
+
"NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u4e00\u822c|advcl": 60,
|
319 |
+
"NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u4e00\u822c|compound": 61,
|
320 |
+
"NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u4e00\u822c|conj": 62,
|
321 |
+
"NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u4e00\u822c|nmod": 63,
|
322 |
+
"NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u4e00\u822c|obl": 64,
|
323 |
+
"NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u4e00\u822c|parataxis": 65,
|
324 |
+
"NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u52a9\u6570\u8a5e\u53ef\u80fd|appos": 66,
|
325 |
+
"NOUN|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u52a9\u6570\u8a5e\u53ef\u80fd|clf": 67,
|
326 |
+
"NOUN|\u540d\u8a5e|acl": 68,
|
327 |
+
"NOUN|\u540d\u8a5e|advcl": 69,
|
328 |
+
"NOUN|\u540d\u8a5e|amod": 70,
|
329 |
+
"NOUN|\u540d\u8a5e|appos": 71,
|
330 |
+
"NOUN|\u540d\u8a5e|case": 72,
|
331 |
+
"NOUN|\u540d\u8a5e|ccomp": 73,
|
332 |
+
"NOUN|\u540d\u8a5e|clf": 74,
|
333 |
+
"NOUN|\u540d\u8a5e|compound": 75,
|
334 |
+
"NOUN|\u540d\u8a5e|conj": 76,
|
335 |
+
"NOUN|\u540d\u8a5e|dislocated": 77,
|
336 |
+
"NOUN|\u540d\u8a5e|expl": 78,
|
337 |
+
"NOUN|\u540d\u8a5e|iobj": 79,
|
338 |
+
"NOUN|\u540d\u8a5e|nmod": 80,
|
339 |
+
"NOUN|\u540d\u8a5e|nsubj": 81,
|
340 |
+
"NOUN|\u540d\u8a5e|obj": 82,
|
341 |
+
"NOUN|\u540d\u8a5e|obl": 83,
|
342 |
+
"NOUN|\u540d\u8a5e|parataxis": 84,
|
343 |
+
"NOUN|\u540d\u8a5e|root": 85,
|
344 |
+
"NOUN|\u540d\u8a5e|vocative": 86,
|
345 |
+
"NOUN|\u5f62\u5f0f\u540d\u8a5e|acl": 87,
|
346 |
+
"NOUN|\u5f62\u5f0f\u540d\u8a5e|advcl": 88,
|
347 |
+
"NOUN|\u5f62\u5f0f\u540d\u8a5e|conj": 89,
|
348 |
+
"NOUN|\u5f62\u5f0f\u540d\u8a5e|iobj": 90,
|
349 |
+
"NOUN|\u5f62\u5f0f\u540d\u8a5e|nmod": 91,
|
350 |
+
"NOUN|\u5f62\u5f0f\u540d\u8a5e|nsubj": 92,
|
351 |
+
"NOUN|\u5f62\u5f0f\u540d\u8a5e|obj": 93,
|
352 |
+
"NOUN|\u5f62\u5f0f\u540d\u8a5e|obl": 94,
|
353 |
+
"NOUN|\u5f62\u5f0f\u540d\u8a5e|parataxis": 95,
|
354 |
+
"NOUN|\u5f62\u5f0f\u540d\u8a5e|root": 96,
|
355 |
+
"NOUN|\u63a5\u5c3e\u8f9e-\u540d\u8a5e\u7684-\u4e00\u822c|compound": 97,
|
356 |
+
"NOUN|\u63a5\u5c3e\u8f9e-\u540d\u8a5e\u7684-\u4e00\u822c|nmod": 98,
|
357 |
+
"NOUN|\u63a5\u5c3e\u8f9e-\u540d\u8a5e\u7684-\u4e00\u822c|nsubj": 99,
|
358 |
+
"NOUN|\u63a5\u982d\u8f9e-\u4e00\u822c|compound": 100,
|
359 |
+
"NUM|\u540d\u8a5e-\u6570\u8a5e|nummod": 101,
|
360 |
+
"NUM|\u6570\u8a5e|acl": 102,
|
361 |
+
"NUM|\u6570\u8a5e|advcl": 103,
|
362 |
+
"NUM|\u6570\u8a5e|case": 104,
|
363 |
+
"NUM|\u6570\u8a5e|mark": 105,
|
364 |
+
"NUM|\u6570\u8a5e|nmod": 106,
|
365 |
+
"NUM|\u6570\u8a5e|nsubj": 107,
|
366 |
+
"NUM|\u6570\u8a5e|nummod": 108,
|
367 |
+
"NUM|\u6570\u8a5e|obj": 109,
|
368 |
+
"NUM|\u6570\u8a5e|root": 110,
|
369 |
+
"PART|\u4eba\u79f0\u63a5\u8f9e|advcl": 111,
|
370 |
+
"PART|\u4eba\u79f0\u63a5\u8f9e|appos": 112,
|
371 |
+
"PART|\u4eba\u79f0\u63a5\u8f9e|case": 113,
|
372 |
+
"PART|\u4eba\u79f0\u63a5\u8f9e|det": 114,
|
373 |
+
"PART|\u4eba\u79f0\u63a5\u8f9e|expl": 115,
|
374 |
+
"PART|\u4eba\u79f0\u63a5\u8f9e|iobj": 116,
|
375 |
+
"PART|\u4eba\u79f0\u63a5\u8f9e|mark": 117,
|
376 |
+
"PART|\u4eba\u79f0\u63a5\u8f9e|nmod": 118,
|
377 |
+
"PART|\u4eba\u79f0\u63a5\u8f9e|nsubj": 119,
|
378 |
+
"PART|\u4eba\u79f0\u63a5\u8f9e|obj": 120,
|
379 |
+
"PART|\u4eba\u79f0\u63a5\u8f9e|obl": 121,
|
380 |
+
"PART|\u4eba\u79f0\u63a5\u8f9e|parataxis": 122,
|
381 |
+
"PART|\u4eba\u79f0\u63a5\u8f9e|root": 123,
|
382 |
+
"PART|\u63a5\u5c3e\u8f9e|acl": 124,
|
383 |
+
"PART|\u63a5\u5c3e\u8f9e|advcl": 125,
|
384 |
+
"PART|\u63a5\u5c3e\u8f9e|conj": 126,
|
385 |
+
"PART|\u63a5\u5c3e\u8f9e|fixed": 127,
|
386 |
+
"PART|\u63a5\u5c3e\u8f9e|mark": 128,
|
387 |
+
"PART|\u63a5\u5c3e\u8f9e|nmod": 129,
|
388 |
+
"PART|\u63a5\u5c3e\u8f9e|nsubj": 130,
|
389 |
+
"PART|\u63a5\u5c3e\u8f9e|obj": 131,
|
390 |
+
"PART|\u63a5\u5c3e\u8f9e|obl": 132,
|
391 |
+
"PART|\u63a5\u5c3e\u8f9e|parataxis": 133,
|
392 |
+
"PART|\u63a5\u5c3e\u8f9e|root": 134,
|
393 |
+
"PART|\u63a5\u982d\u8f9e|compound": 135,
|
394 |
+
"PART|\u63a5\u982d\u8f9e|det": 136,
|
395 |
+
"PART|\u63a5\u982d\u8f9e|iobj": 137,
|
396 |
+
"PART|\u63a5\u982d\u8f9e|parataxis": 138,
|
397 |
+
"PART|\u7d42\u52a9\u8a5e|discourse": 139,
|
398 |
+
"PART|\u7d42\u52a9\u8a5e|expl": 140,
|
399 |
+
"PART|\u7d42\u52a9\u8a5e|mark": 141,
|
400 |
+
"PART|\u7d42\u52a9\u8a5e|nsubj": 142,
|
401 |
+
"PART|\u7d42\u52a9\u8a5e|root": 143,
|
402 |
+
"PRON|\u4ee3\u540d\u8a5e|advcl": 144,
|
403 |
+
"PRON|\u4ee3\u540d\u8a5e|case": 145,
|
404 |
+
"PRON|\u4ee3\u540d\u8a5e|compound": 146,
|
405 |
+
"PRON|\u4ee3\u540d\u8a5e|det": 147,
|
406 |
+
"PRON|\u4ee3\u540d\u8a5e|dislocated": 148,
|
407 |
+
"PRON|\u4ee3\u540d\u8a5e|nmod": 149,
|
408 |
+
"PRON|\u4ee3\u540d\u8a5e|nsubj": 150,
|
409 |
+
"PRON|\u4ee3\u540d\u8a5e|obj": 151,
|
410 |
+
"PRON|\u4ee3\u540d\u8a5e|obl": 152,
|
411 |
+
"PRON|\u4ee3\u540d\u8a5e|parataxis": 153,
|
412 |
+
"PRON|\u4ee3\u540d\u8a5e|root": 154,
|
413 |
+
"PRON|\u4ee3\u540d\u8a5e|vocative": 155,
|
414 |
+
"PROPN|_|nmod": 156,
|
415 |
+
"PROPN|\u540d\u8a5e-\u56fa\u6709\u540d\u8a5e-\u4e00\u822c|compound": 157,
|
416 |
+
"PROPN|\u540d\u8a5e-\u56fa\u6709\u540d\u8a5e-\u4eba\u540d-\u540d|conj": 158,
|
417 |
+
"PROPN|\u540d\u8a5e-\u56fa\u6709\u540d\u8a5e-\u4eba\u540d-\u540d|flat": 159,
|
418 |
+
"PROPN|\u540d\u8a5e-\u56fa\u6709\u540d\u8a5e-\u4eba\u540d-\u59d3|nsubj": 160,
|
419 |
+
"PROPN|\u540d\u8a5e-\u56fa\u6709\u540d\u8a5e-\u5730\u540d-\u56fd|compound": 161,
|
420 |
+
"PROPN|\u56fa\u6709\u540d\u8a5e|advcl": 162,
|
421 |
+
"PROPN|\u56fa\u6709\u540d\u8a5e|appos": 163,
|
422 |
+
"PROPN|\u56fa\u6709\u540d\u8a5e|conj": 164,
|
423 |
+
"PROPN|\u56fa\u6709\u540d\u8a5e|flat": 165,
|
424 |
+
"PROPN|\u56fa\u6709\u540d\u8a5e|nmod": 166,
|
425 |
+
"PROPN|\u56fa\u6709\u540d\u8a5e|nsubj": 167,
|
426 |
+
"PROPN|\u56fa\u6709\u540d\u8a5e|obj": 168,
|
427 |
+
"PROPN|\u56fa\u6709\u540d\u8a5e|obl": 169,
|
428 |
+
"PROPN|\u56fa\u6709\u540d\u8a5e|root": 170,
|
429 |
+
"PUNCT|\u88dc\u52a9\u8a18\u53f7-\u62ec\u5f27\u9589|punct": 171,
|
430 |
+
"PUNCT|\u88dc\u52a9\u8a18\u53f7-\u62ec\u5f27\u958b|punct": 172,
|
431 |
+
"PUNCT|\u8a18\u53f7|case": 173,
|
432 |
+
"PUNCT|\u8a18\u53f7|conj": 174,
|
433 |
+
"PUNCT|\u8a18\u53f7|nmod": 175,
|
434 |
+
"PUNCT|\u8a18\u53f7|nsubj": 176,
|
435 |
+
"PUNCT|\u8a18\u53f7|obj": 177,
|
436 |
+
"PUNCT|\u8a18\u53f7|obl": 178,
|
437 |
+
"PUNCT|\u8a18\u53f7|punct": 179,
|
438 |
+
"PUNCT|\u8a18\u53f7|root": 180,
|
439 |
+
"SCONJ|\u5f8c\u7f6e\u526f\u8a5e|acl": 181,
|
440 |
+
"SCONJ|\u5f8c\u7f6e\u526f\u8a5e|advcl": 182,
|
441 |
+
"SCONJ|\u5f8c\u7f6e\u526f\u8a5e|advmod": 183,
|
442 |
+
"SCONJ|\u5f8c\u7f6e\u526f\u8a5e|amod": 184,
|
443 |
+
"SCONJ|\u5f8c\u7f6e\u526f\u8a5e|case": 185,
|
444 |
+
"SCONJ|\u5f8c\u7f6e\u526f\u8a5e|mark": 186,
|
445 |
+
"SCONJ|\u5f8c\u7f6e\u526f\u8a5e|parataxis": 187,
|
446 |
+
"SCONJ|\u5f8c\u7f6e\u526f\u8a5e|root": 188,
|
447 |
+
"SCONJ|\u63a5\u7d9a\u52a9\u8a5e|advmod": 189,
|
448 |
+
"SCONJ|\u63a5\u7d9a\u52a9\u8a5e|case": 190,
|
449 |
+
"SCONJ|\u63a5\u7d9a\u52a9\u8a5e|mark": 191,
|
450 |
+
"SCONJ|\u63a5\u7d9a\u52a9\u8a5e|nsubj": 192,
|
451 |
+
"SCONJ|\u63a5\u7d9a\u52a9\u8a5e|obj": 193,
|
452 |
+
"SCONJ|\u63a5\u7d9a\u52a9\u8a5e|parataxis": 194,
|
453 |
+
"SCONJ|\u63a5\u7d9a\u52a9\u8a5e|punct": 195,
|
454 |
+
"SCONJ|\u63a5\u7d9a\u52a9\u8a5e|root": 196,
|
455 |
+
"SCONJ|\u63a5\u7d9a\u8a5e|advmod": 197,
|
456 |
+
"SCONJ|\u63a5\u7d9a\u8a5e|case": 198,
|
457 |
+
"SCONJ|\u63a5\u7d9a\u8a5e|mark": 199,
|
458 |
+
"SYM|_|conj": 200,
|
459 |
+
"SYM|_|obj": 201,
|
460 |
+
"VERB|\u4ed6\u52d5\u8a5e|acl": 202,
|
461 |
+
"VERB|\u4ed6\u52d5\u8a5e|advcl": 203,
|
462 |
+
"VERB|\u4ed6\u52d5\u8a5e|amod": 204,
|
463 |
+
"VERB|\u4ed6\u52d5\u8a5e|appos": 205,
|
464 |
+
"VERB|\u4ed6\u52d5\u8a5e|case": 206,
|
465 |
+
"VERB|\u4ed6\u52d5\u8a5e|ccomp": 207,
|
466 |
+
"VERB|\u4ed6\u52d5\u8a5e|conj": 208,
|
467 |
+
"VERB|\u4ed6\u52d5\u8a5e|csubj": 209,
|
468 |
+
"VERB|\u4ed6\u52d5\u8a5e|mark": 210,
|
469 |
+
"VERB|\u4ed6\u52d5\u8a5e|obj": 211,
|
470 |
+
"VERB|\u4ed6\u52d5\u8a5e|obl": 212,
|
471 |
+
"VERB|\u4ed6\u52d5\u8a5e|parataxis": 213,
|
472 |
+
"VERB|\u4ed6\u52d5\u8a5e|root": 214,
|
473 |
+
"VERB|\u540d\u8a5e-\u666e\u901a\u540d\u8a5e-\u30b5\u5909\u53ef\u80fd|acl": 215,
|
474 |
+
"VERB|\u5b8c\u5168\u52d5\u8a5e|acl": 216,
|
475 |
+
"VERB|\u5b8c\u5168\u52d5\u8a5e|advcl": 217,
|
476 |
+
"VERB|\u5b8c\u5168\u52d5\u8a5e|amod": 218,
|
477 |
+
"VERB|\u5b8c\u5168\u52d5\u8a5e|parataxis": 219,
|
478 |
+
"VERB|\u5b8c\u5168\u52d5\u8a5e|root": 220,
|
479 |
+
"VERB|\u81ea\u52d5\u8a5e|acl": 221,
|
480 |
+
"VERB|\u81ea\u52d5\u8a5e|advcl": 222,
|
481 |
+
"VERB|\u81ea\u52d5\u8a5e|amod": 223,
|
482 |
+
"VERB|\u81ea\u52d5\u8a5e|ccomp": 224,
|
483 |
+
"VERB|\u81ea\u52d5\u8a5e|conj": 225,
|
484 |
+
"VERB|\u81ea\u52d5\u8a5e|csubj": 226,
|
485 |
+
"VERB|\u81ea\u52d5\u8a5e|det": 227,
|
486 |
+
"VERB|\u81ea\u52d5\u8a5e|mark": 228,
|
487 |
+
"VERB|\u81ea\u52d5\u8a5e|nmod": 229,
|
488 |
+
"VERB|\u81ea\u52d5\u8a5e|nsubj": 230,
|
489 |
+
"VERB|\u81ea\u52d5\u8a5e|parataxis": 231,
|
490 |
+
"VERB|\u81ea\u52d5\u8a5e|root": 232,
|
491 |
+
"X|_|conj": 233,
|
492 |
+
"X|_|goeswith": 234,
|
493 |
+
"X|_|nsubj": 235
|
494 |
},
|
495 |
"layer_norm_eps": 1e-12,
|
496 |
"max_position_embeddings": 512,
|
|
|
501 |
"position_embedding_type": "absolute",
|
502 |
"tokenizer_class": "RemBertTokenizerFast",
|
503 |
"torch_dtype": "float32",
|
504 |
+
"transformers_version": "4.36.2",
|
505 |
"type_vocab_size": 2,
|
506 |
"use_cache": true,
|
507 |
"vocab_size": 6877
|
pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:39e7062ca2160bc7c3f2cd1b78c6f96e0bc56620b323232e0ef922bc9e1e42db
|
3 |
+
size 363721651
|
special_tokens_map.json
CHANGED
@@ -1,7 +1,25 @@
|
|
1 |
{
|
2 |
-
"bos_token":
|
3 |
-
|
4 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
"mask_token": {
|
6 |
"content": "[MASK]",
|
7 |
"lstrip": true,
|
@@ -9,7 +27,25 @@
|
|
9 |
"rstrip": false,
|
10 |
"single_word": false
|
11 |
},
|
12 |
-
"pad_token":
|
13 |
-
|
14 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
15 |
}
|
|
|
1 |
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "[CLS]",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": false,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"cls_token": {
|
10 |
+
"content": "[CLS]",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": false,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"eos_token": {
|
17 |
+
"content": "[SEP]",
|
18 |
+
"lstrip": false,
|
19 |
+
"normalized": false,
|
20 |
+
"rstrip": false,
|
21 |
+
"single_word": false
|
22 |
+
},
|
23 |
"mask_token": {
|
24 |
"content": "[MASK]",
|
25 |
"lstrip": true,
|
|
|
27 |
"rstrip": false,
|
28 |
"single_word": false
|
29 |
},
|
30 |
+
"pad_token": {
|
31 |
+
"content": "[PAD]",
|
32 |
+
"lstrip": false,
|
33 |
+
"normalized": false,
|
34 |
+
"rstrip": false,
|
35 |
+
"single_word": false
|
36 |
+
},
|
37 |
+
"sep_token": {
|
38 |
+
"content": "[SEP]",
|
39 |
+
"lstrip": false,
|
40 |
+
"normalized": false,
|
41 |
+
"rstrip": false,
|
42 |
+
"single_word": false
|
43 |
+
},
|
44 |
+
"unk_token": {
|
45 |
+
"content": "[UNK]",
|
46 |
+
"lstrip": false,
|
47 |
+
"normalized": false,
|
48 |
+
"rstrip": false,
|
49 |
+
"single_word": false
|
50 |
+
}
|
51 |
}
|
tokenizer.json
CHANGED
@@ -28755,6 +28755,7 @@
|
|
28755 |
"ciasn",
|
28756 |
-17.181486872491696
|
28757 |
]
|
28758 |
-
]
|
|
|
28759 |
}
|
28760 |
}
|
|
|
28755 |
"ciasn",
|
28756 |
-17.181486872491696
|
28757 |
]
|
28758 |
+
],
|
28759 |
+
"byte_fallback": false
|
28760 |
}
|
28761 |
}
|
tokenizer_config.json
CHANGED
@@ -1,22 +1,62 @@
|
|
1 |
{
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
"bos_token": "[CLS]",
|
|
|
3 |
"cls_token": "[CLS]",
|
4 |
"do_lower_case": true,
|
5 |
"eos_token": "[SEP]",
|
6 |
"keep_accents": false,
|
7 |
-
"mask_token":
|
8 |
-
|
9 |
-
"content": "[MASK]",
|
10 |
-
"lstrip": true,
|
11 |
-
"normalized": true,
|
12 |
-
"rstrip": false,
|
13 |
-
"single_word": false
|
14 |
-
},
|
15 |
"model_max_length": 512,
|
16 |
"pad_token": "[PAD]",
|
17 |
"remove_space": true,
|
18 |
"sep_token": "[SEP]",
|
19 |
"split_by_punct": true,
|
|
|
20 |
"tokenizer_class": "RemBertTokenizerFast",
|
|
|
|
|
21 |
"unk_token": "[UNK]"
|
22 |
}
|
|
|
1 |
{
|
2 |
+
"added_tokens_decoder": {
|
3 |
+
"0": {
|
4 |
+
"content": "[CLS]",
|
5 |
+
"lstrip": false,
|
6 |
+
"normalized": false,
|
7 |
+
"rstrip": false,
|
8 |
+
"single_word": false,
|
9 |
+
"special": true
|
10 |
+
},
|
11 |
+
"1": {
|
12 |
+
"content": "[PAD]",
|
13 |
+
"lstrip": false,
|
14 |
+
"normalized": false,
|
15 |
+
"rstrip": false,
|
16 |
+
"single_word": false,
|
17 |
+
"special": true
|
18 |
+
},
|
19 |
+
"2": {
|
20 |
+
"content": "[SEP]",
|
21 |
+
"lstrip": false,
|
22 |
+
"normalized": false,
|
23 |
+
"rstrip": false,
|
24 |
+
"single_word": false,
|
25 |
+
"special": true
|
26 |
+
},
|
27 |
+
"3": {
|
28 |
+
"content": "[UNK]",
|
29 |
+
"lstrip": false,
|
30 |
+
"normalized": false,
|
31 |
+
"rstrip": false,
|
32 |
+
"single_word": false,
|
33 |
+
"special": true
|
34 |
+
},
|
35 |
+
"4": {
|
36 |
+
"content": "[MASK]",
|
37 |
+
"lstrip": true,
|
38 |
+
"normalized": true,
|
39 |
+
"rstrip": false,
|
40 |
+
"single_word": false,
|
41 |
+
"special": true
|
42 |
+
}
|
43 |
+
},
|
44 |
"bos_token": "[CLS]",
|
45 |
+
"clean_up_tokenization_spaces": true,
|
46 |
"cls_token": "[CLS]",
|
47 |
"do_lower_case": true,
|
48 |
"eos_token": "[SEP]",
|
49 |
"keep_accents": false,
|
50 |
+
"mask_token": "[MASK]",
|
51 |
+
"max_length": 510,
|
|
|
|
|
|
|
|
|
|
|
|
|
52 |
"model_max_length": 512,
|
53 |
"pad_token": "[PAD]",
|
54 |
"remove_space": true,
|
55 |
"sep_token": "[SEP]",
|
56 |
"split_by_punct": true,
|
57 |
+
"stride": 0,
|
58 |
"tokenizer_class": "RemBertTokenizerFast",
|
59 |
+
"truncation_side": "right",
|
60 |
+
"truncation_strategy": "longest_first",
|
61 |
"unk_token": "[UNK]"
|
62 |
}
|