Hugging Face's logo

GeneZC
/
ofabert-base-minilm-6L-768H

Transformers
PyTorch
bert
Model card Files Community
GeneZC commited on
Commit
ccc00a1
·
1 Parent(s): 53fe7ea

Upload 5 files

Browse files
Files changed (5) hide show
  1. config.json +437 -0
  2. pytorch_model.bin +3 -0
  3. special_tokens_map.json +1 -0
  4. tokenizer_config.json +1 -0
  5. vocab.txt +0 -0
config.json ADDED
@@ -0,0 +1,437 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "../plms/bert-base-uncased",
3
+ "architectures": [
4
+ "SparseBertHiddenMiniLM"
5
+ ],
6
+ "attention_head_size": 64,
7
+ "attention_probs_dropout_prob": 0.1,
8
+ "classifier_dropout": null,
9
+ "gradient_checkpointing": false,
10
+ "hidden_act": "gelu",
11
+ "hidden_dropout_prob": 0.1,
12
+ "hidden_size": 768,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 3072,
15
+ "layer_norm_eps": 1e-12,
16
+ "layer_skip": 1,
17
+ "max_position_embeddings": 512,
18
+ "model_type": "bert",
19
+ "num_attention_heads": 12,
20
+ "num_hidden_layers": 6,
21
+ "num_relation_heads": 32,
22
+ "pad_token_id": 0,
23
+ "position_embedding_type": "absolute",
24
+ "sparsity": "6,768",
25
+ "sparsity_map": {
26
+ "3,192": {
27
+ "head": {
28
+ "0": 9,
29
+ "1": 9,
30
+ "2": 9,
31
+ "3": 12,
32
+ "4": 12,
33
+ "5": 12
34
+ },
35
+ "hidden": {
36
+ "-1": 576,
37
+ "0": 576,
38
+ "1": 576,
39
+ "2": 576,
40
+ "3": 576,
41
+ "4": 576,
42
+ "5": 576
43
+ },
44
+ "neuron": {
45
+ "0": 2304,
46
+ "1": 2304,
47
+ "2": 2304,
48
+ "3": 3072,
49
+ "4": 3072,
50
+ "5": 3072
51
+ }
52
+ },
53
+ "3,256": {
54
+ "head": {
55
+ "0": 8,
56
+ "1": 8,
57
+ "2": 8,
58
+ "3": 12,
59
+ "4": 12,
60
+ "5": 12
61
+ },
62
+ "hidden": {
63
+ "-1": 512,
64
+ "0": 512,
65
+ "1": 512,
66
+ "2": 512,
67
+ "3": 512,
68
+ "4": 512,
69
+ "5": 512
70
+ },
71
+ "neuron": {
72
+ "0": 2048,
73
+ "1": 2048,
74
+ "2": 2048,
75
+ "3": 3072,
76
+ "4": 3072,
77
+ "5": 3072
78
+ }
79
+ },
80
+ "3,384": {
81
+ "head": {
82
+ "0": 6,
83
+ "1": 6,
84
+ "2": 6,
85
+ "3": 12,
86
+ "4": 12,
87
+ "5": 12
88
+ },
89
+ "hidden": {
90
+ "-1": 384,
91
+ "0": 384,
92
+ "1": 384,
93
+ "2": 384,
94
+ "3": 384,
95
+ "4": 384,
96
+ "5": 384
97
+ },
98
+ "neuron": {
99
+ "0": 1536,
100
+ "1": 1536,
101
+ "2": 1536,
102
+ "3": 3072,
103
+ "4": 3072,
104
+ "5": 3072
105
+ }
106
+ },
107
+ "3,512": {
108
+ "head": {
109
+ "0": 4,
110
+ "1": 4,
111
+ "2": 4,
112
+ "3": 12,
113
+ "4": 12,
114
+ "5": 12
115
+ },
116
+ "hidden": {
117
+ "-1": 256,
118
+ "0": 256,
119
+ "1": 256,
120
+ "2": 256,
121
+ "3": 256,
122
+ "4": 256,
123
+ "5": 256
124
+ },
125
+ "neuron": {
126
+ "0": 1024,
127
+ "1": 1024,
128
+ "2": 1024,
129
+ "3": 3072,
130
+ "4": 3072,
131
+ "5": 3072
132
+ }
133
+ },
134
+ "3,768": {
135
+ "head": {
136
+ "0": 0,
137
+ "1": 0,
138
+ "2": 0,
139
+ "3": 12,
140
+ "4": 12,
141
+ "5": 12
142
+ },
143
+ "hidden": {
144
+ "-1": 0,
145
+ "0": 0,
146
+ "1": 0,
147
+ "2": 0,
148
+ "3": 0,
149
+ "4": 0,
150
+ "5": 0
151
+ },
152
+ "neuron": {
153
+ "0": 0,
154
+ "1": 0,
155
+ "2": 0,
156
+ "3": 3072,
157
+ "4": 3072,
158
+ "5": 3072
159
+ }
160
+ },
161
+ "4,192": {
162
+ "head": {
163
+ "0": 9,
164
+ "1": 9,
165
+ "2": 9,
166
+ "3": 9,
167
+ "4": 12,
168
+ "5": 12
169
+ },
170
+ "hidden": {
171
+ "-1": 576,
172
+ "0": 576,
173
+ "1": 576,
174
+ "2": 576,
175
+ "3": 576,
176
+ "4": 576,
177
+ "5": 576
178
+ },
179
+ "neuron": {
180
+ "0": 2304,
181
+ "1": 2304,
182
+ "2": 2304,
183
+ "3": 2304,
184
+ "4": 3072,
185
+ "5": 3072
186
+ }
187
+ },
188
+ "4,256": {
189
+ "head": {
190
+ "0": 8,
191
+ "1": 8,
192
+ "2": 8,
193
+ "3": 8,
194
+ "4": 12,
195
+ "5": 12
196
+ },
197
+ "hidden": {
198
+ "-1": 512,
199
+ "0": 512,
200
+ "1": 512,
201
+ "2": 512,
202
+ "3": 512,
203
+ "4": 512,
204
+ "5": 512
205
+ },
206
+ "neuron": {
207
+ "0": 2048,
208
+ "1": 2048,
209
+ "2": 2048,
210
+ "3": 2048,
211
+ "4": 3072,
212
+ "5": 3072
213
+ }
214
+ },
215
+ "4,384": {
216
+ "head": {
217
+ "0": 6,
218
+ "1": 6,
219
+ "2": 6,
220
+ "3": 6,
221
+ "4": 12,
222
+ "5": 12
223
+ },
224
+ "hidden": {
225
+ "-1": 384,
226
+ "0": 384,
227
+ "1": 384,
228
+ "2": 384,
229
+ "3": 384,
230
+ "4": 384,
231
+ "5": 384
232
+ },
233
+ "neuron": {
234
+ "0": 1536,
235
+ "1": 1536,
236
+ "2": 1536,
237
+ "3": 1536,
238
+ "4": 3072,
239
+ "5": 3072
240
+ }
241
+ },
242
+ "4,512": {
243
+ "head": {
244
+ "0": 4,
245
+ "1": 4,
246
+ "2": 4,
247
+ "3": 4,
248
+ "4": 12,
249
+ "5": 12
250
+ },
251
+ "hidden": {
252
+ "-1": 256,
253
+ "0": 256,
254
+ "1": 256,
255
+ "2": 256,
256
+ "3": 256,
257
+ "4": 256,
258
+ "5": 256
259
+ },
260
+ "neuron": {
261
+ "0": 1024,
262
+ "1": 1024,
263
+ "2": 1024,
264
+ "3": 1024,
265
+ "4": 3072,
266
+ "5": 3072
267
+ }
268
+ },
269
+ "4,768": {
270
+ "head": {
271
+ "0": 0,
272
+ "1": 0,
273
+ "2": 0,
274
+ "3": 0,
275
+ "4": 12,
276
+ "5": 12
277
+ },
278
+ "hidden": {
279
+ "-1": 0,
280
+ "0": 0,
281
+ "1": 0,
282
+ "2": 0,
283
+ "3": 0,
284
+ "4": 0,
285
+ "5": 0
286
+ },
287
+ "neuron": {
288
+ "0": 0,
289
+ "1": 0,
290
+ "2": 0,
291
+ "3": 0,
292
+ "4": 3072,
293
+ "5": 3072
294
+ }
295
+ },
296
+ "6,192": {
297
+ "head": {
298
+ "0": 9,
299
+ "1": 9,
300
+ "2": 9,
301
+ "3": 9,
302
+ "4": 9,
303
+ "5": 9
304
+ },
305
+ "hidden": {
306
+ "-1": 576,
307
+ "0": 576,
308
+ "1": 576,
309
+ "2": 576,
310
+ "3": 576,
311
+ "4": 576,
312
+ "5": 576
313
+ },
314
+ "neuron": {
315
+ "0": 2304,
316
+ "1": 2304,
317
+ "2": 2304,
318
+ "3": 2304,
319
+ "4": 2304,
320
+ "5": 2304
321
+ }
322
+ },
323
+ "6,256": {
324
+ "head": {
325
+ "0": 8,
326
+ "1": 8,
327
+ "2": 8,
328
+ "3": 8,
329
+ "4": 8,
330
+ "5": 8
331
+ },
332
+ "hidden": {
333
+ "-1": 512,
334
+ "0": 512,
335
+ "1": 512,
336
+ "2": 512,
337
+ "3": 512,
338
+ "4": 512,
339
+ "5": 512
340
+ },
341
+ "neuron": {
342
+ "0": 2048,
343
+ "1": 2048,
344
+ "2": 2048,
345
+ "3": 2048,
346
+ "4": 2048,
347
+ "5": 2048
348
+ }
349
+ },
350
+ "6,384": {
351
+ "head": {
352
+ "0": 6,
353
+ "1": 6,
354
+ "2": 6,
355
+ "3": 6,
356
+ "4": 6,
357
+ "5": 6
358
+ },
359
+ "hidden": {
360
+ "-1": 384,
361
+ "0": 384,
362
+ "1": 384,
363
+ "2": 384,
364
+ "3": 384,
365
+ "4": 384,
366
+ "5": 384
367
+ },
368
+ "neuron": {
369
+ "0": 1536,
370
+ "1": 1536,
371
+ "2": 1536,
372
+ "3": 1536,
373
+ "4": 1536,
374
+ "5": 1536
375
+ }
376
+ },
377
+ "6,512": {
378
+ "head": {
379
+ "0": 4,
380
+ "1": 4,
381
+ "2": 4,
382
+ "3": 4,
383
+ "4": 4,
384
+ "5": 4
385
+ },
386
+ "hidden": {
387
+ "-1": 256,
388
+ "0": 256,
389
+ "1": 256,
390
+ "2": 256,
391
+ "3": 256,
392
+ "4": 256,
393
+ "5": 256
394
+ },
395
+ "neuron": {
396
+ "0": 1024,
397
+ "1": 1024,
398
+ "2": 1024,
399
+ "3": 1024,
400
+ "4": 1024,
401
+ "5": 1024
402
+ }
403
+ },
404
+ "6,768": {
405
+ "head": {
406
+ "0": 0,
407
+ "1": 0,
408
+ "2": 0,
409
+ "3": 0,
410
+ "4": 0,
411
+ "5": 0
412
+ },
413
+ "hidden": {
414
+ "-1": 0,
415
+ "0": 0,
416
+ "1": 0,
417
+ "2": 0,
418
+ "3": 0,
419
+ "4": 0,
420
+ "5": 0
421
+ },
422
+ "neuron": {
423
+ "0": 0,
424
+ "1": 0,
425
+ "2": 0,
426
+ "3": 0,
427
+ "4": 0,
428
+ "5": 0
429
+ }
430
+ }
431
+ },
432
+ "torch_dtype": "float32",
433
+ "transformers_version": "4.9.2",
434
+ "type_vocab_size": 2,
435
+ "use_cache": true,
436
+ "vocab_size": 30522
437
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:27737fd76295e9b90201828951484e76b4e9cc7f48b6f411134cfb966c5668be
3
+ size 274959706
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"do_lower_case": true, "do_basic_tokenize": true, "never_split": null, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "/mnt/dolphinfs/ssd_pool/docker/user/hadoop-aipnlp/zhangchen76/mosification/sparsebert_hidden-base", "use_fast": true, "tokenizer_file": null, "tokenizer_class": "BertTokenizer"}
vocab.txt ADDED
The diff for this file is too large to render. See raw diff