李登杰 commited on
Commit
1abeac5
·
1 Parent(s): cc41344
config.json ADDED
@@ -0,0 +1,1376 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/mnt/dolphinfs/ssd_pool/docker/user/hadoop-vacv/lidengjie/projects/fsvd-llm/cache/llama-7b-hf/models/stage1-bf16-usv",
3
+ "architectures": [
4
+ "LlamaForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "bos_token_id": 1,
9
+ "eos_token_id": 2,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 4096,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 11008,
14
+ "linear_info": {
15
+ "model.layers.0.mlp.down_proj": {
16
+ "bias": true,
17
+ "in_features": 11008,
18
+ "out_features": 4096,
19
+ "rank": 2023
20
+ },
21
+ "model.layers.0.mlp.gate_proj": {
22
+ "bias": true,
23
+ "in_features": 4096,
24
+ "out_features": 11008,
25
+ "rank": 1124
26
+ },
27
+ "model.layers.0.mlp.up_proj": {
28
+ "bias": true,
29
+ "in_features": 4096,
30
+ "out_features": 11008,
31
+ "rank": 1139
32
+ },
33
+ "model.layers.0.self_attn.k_proj": {
34
+ "bias": true,
35
+ "in_features": 4096,
36
+ "out_features": 4096,
37
+ "rank": 94
38
+ },
39
+ "model.layers.0.self_attn.o_proj": {
40
+ "bias": true,
41
+ "in_features": 4096,
42
+ "out_features": 4096,
43
+ "rank": 538
44
+ },
45
+ "model.layers.0.self_attn.q_proj": {
46
+ "bias": true,
47
+ "in_features": 4096,
48
+ "out_features": 4096,
49
+ "rank": 83
50
+ },
51
+ "model.layers.0.self_attn.v_proj": {
52
+ "bias": true,
53
+ "in_features": 4096,
54
+ "out_features": 4096,
55
+ "rank": 1063
56
+ },
57
+ "model.layers.1.mlp.down_proj": {
58
+ "bias": true,
59
+ "in_features": 11008,
60
+ "out_features": 4096,
61
+ "rank": 1644
62
+ },
63
+ "model.layers.1.mlp.gate_proj": {
64
+ "bias": true,
65
+ "in_features": 4096,
66
+ "out_features": 11008,
67
+ "rank": 1288
68
+ },
69
+ "model.layers.1.mlp.up_proj": {
70
+ "bias": true,
71
+ "in_features": 4096,
72
+ "out_features": 11008,
73
+ "rank": 1319
74
+ },
75
+ "model.layers.1.self_attn.k_proj": {
76
+ "bias": true,
77
+ "in_features": 4096,
78
+ "out_features": 4096,
79
+ "rank": 207
80
+ },
81
+ "model.layers.1.self_attn.o_proj": {
82
+ "bias": true,
83
+ "in_features": 4096,
84
+ "out_features": 4096,
85
+ "rank": 945
86
+ },
87
+ "model.layers.1.self_attn.q_proj": {
88
+ "bias": true,
89
+ "in_features": 4096,
90
+ "out_features": 4096,
91
+ "rank": 211
92
+ },
93
+ "model.layers.1.self_attn.v_proj": {
94
+ "bias": true,
95
+ "in_features": 4096,
96
+ "out_features": 4096,
97
+ "rank": 762
98
+ },
99
+ "model.layers.10.mlp.down_proj": {
100
+ "bias": true,
101
+ "in_features": 11008,
102
+ "out_features": 4096,
103
+ "rank": 193
104
+ },
105
+ "model.layers.10.mlp.gate_proj": {
106
+ "bias": true,
107
+ "in_features": 4096,
108
+ "out_features": 11008,
109
+ "rank": 162
110
+ },
111
+ "model.layers.10.mlp.up_proj": {
112
+ "bias": true,
113
+ "in_features": 4096,
114
+ "out_features": 11008,
115
+ "rank": 196
116
+ },
117
+ "model.layers.10.self_attn.k_proj": {
118
+ "bias": true,
119
+ "in_features": 4096,
120
+ "out_features": 4096,
121
+ "rank": 94
122
+ },
123
+ "model.layers.10.self_attn.o_proj": {
124
+ "bias": true,
125
+ "in_features": 4096,
126
+ "out_features": 4096,
127
+ "rank": 32
128
+ },
129
+ "model.layers.10.self_attn.q_proj": {
130
+ "bias": true,
131
+ "in_features": 4096,
132
+ "out_features": 4096,
133
+ "rank": 113
134
+ },
135
+ "model.layers.10.self_attn.v_proj": {
136
+ "bias": true,
137
+ "in_features": 4096,
138
+ "out_features": 4096,
139
+ "rank": 99
140
+ },
141
+ "model.layers.11.mlp.down_proj": {
142
+ "bias": true,
143
+ "in_features": 11008,
144
+ "out_features": 4096,
145
+ "rank": 273
146
+ },
147
+ "model.layers.11.mlp.gate_proj": {
148
+ "bias": true,
149
+ "in_features": 4096,
150
+ "out_features": 11008,
151
+ "rank": 219
152
+ },
153
+ "model.layers.11.mlp.up_proj": {
154
+ "bias": true,
155
+ "in_features": 4096,
156
+ "out_features": 11008,
157
+ "rank": 268
158
+ },
159
+ "model.layers.11.self_attn.k_proj": {
160
+ "bias": true,
161
+ "in_features": 4096,
162
+ "out_features": 4096,
163
+ "rank": 54
164
+ },
165
+ "model.layers.11.self_attn.o_proj": {
166
+ "bias": true,
167
+ "in_features": 4096,
168
+ "out_features": 4096,
169
+ "rank": 34
170
+ },
171
+ "model.layers.11.self_attn.q_proj": {
172
+ "bias": true,
173
+ "in_features": 4096,
174
+ "out_features": 4096,
175
+ "rank": 62
176
+ },
177
+ "model.layers.11.self_attn.v_proj": {
178
+ "bias": true,
179
+ "in_features": 4096,
180
+ "out_features": 4096,
181
+ "rank": 133
182
+ },
183
+ "model.layers.12.mlp.down_proj": {
184
+ "bias": true,
185
+ "in_features": 11008,
186
+ "out_features": 4096,
187
+ "rank": 46
188
+ },
189
+ "model.layers.12.mlp.gate_proj": {
190
+ "bias": true,
191
+ "in_features": 4096,
192
+ "out_features": 11008,
193
+ "rank": 84
194
+ },
195
+ "model.layers.12.mlp.up_proj": {
196
+ "bias": true,
197
+ "in_features": 4096,
198
+ "out_features": 11008,
199
+ "rank": 66
200
+ },
201
+ "model.layers.12.self_attn.k_proj": {
202
+ "bias": true,
203
+ "in_features": 4096,
204
+ "out_features": 4096,
205
+ "rank": 100
206
+ },
207
+ "model.layers.12.self_attn.o_proj": {
208
+ "bias": true,
209
+ "in_features": 4096,
210
+ "out_features": 4096,
211
+ "rank": 63
212
+ },
213
+ "model.layers.12.self_attn.q_proj": {
214
+ "bias": true,
215
+ "in_features": 4096,
216
+ "out_features": 4096,
217
+ "rank": 115
218
+ },
219
+ "model.layers.12.self_attn.v_proj": {
220
+ "bias": true,
221
+ "in_features": 4096,
222
+ "out_features": 4096,
223
+ "rank": 156
224
+ },
225
+ "model.layers.13.mlp.down_proj": {
226
+ "bias": true,
227
+ "in_features": 11008,
228
+ "out_features": 4096,
229
+ "rank": 270
230
+ },
231
+ "model.layers.13.mlp.gate_proj": {
232
+ "bias": true,
233
+ "in_features": 4096,
234
+ "out_features": 11008,
235
+ "rank": 247
236
+ },
237
+ "model.layers.13.mlp.up_proj": {
238
+ "bias": true,
239
+ "in_features": 4096,
240
+ "out_features": 11008,
241
+ "rank": 278
242
+ },
243
+ "model.layers.13.self_attn.k_proj": {
244
+ "bias": true,
245
+ "in_features": 4096,
246
+ "out_features": 4096,
247
+ "rank": 40
248
+ },
249
+ "model.layers.13.self_attn.o_proj": {
250
+ "bias": true,
251
+ "in_features": 4096,
252
+ "out_features": 4096,
253
+ "rank": 1
254
+ },
255
+ "model.layers.13.self_attn.q_proj": {
256
+ "bias": true,
257
+ "in_features": 4096,
258
+ "out_features": 4096,
259
+ "rank": 42
260
+ },
261
+ "model.layers.13.self_attn.v_proj": {
262
+ "bias": true,
263
+ "in_features": 4096,
264
+ "out_features": 4096,
265
+ "rank": 55
266
+ },
267
+ "model.layers.14.mlp.down_proj": {
268
+ "bias": true,
269
+ "in_features": 11008,
270
+ "out_features": 4096,
271
+ "rank": 678
272
+ },
273
+ "model.layers.14.mlp.gate_proj": {
274
+ "bias": true,
275
+ "in_features": 4096,
276
+ "out_features": 11008,
277
+ "rank": 467
278
+ },
279
+ "model.layers.14.mlp.up_proj": {
280
+ "bias": true,
281
+ "in_features": 4096,
282
+ "out_features": 11008,
283
+ "rank": 613
284
+ },
285
+ "model.layers.14.self_attn.k_proj": {
286
+ "bias": true,
287
+ "in_features": 4096,
288
+ "out_features": 4096,
289
+ "rank": 271
290
+ },
291
+ "model.layers.14.self_attn.o_proj": {
292
+ "bias": true,
293
+ "in_features": 4096,
294
+ "out_features": 4096,
295
+ "rank": 702
296
+ },
297
+ "model.layers.14.self_attn.q_proj": {
298
+ "bias": true,
299
+ "in_features": 4096,
300
+ "out_features": 4096,
301
+ "rank": 265
302
+ },
303
+ "model.layers.14.self_attn.v_proj": {
304
+ "bias": true,
305
+ "in_features": 4096,
306
+ "out_features": 4096,
307
+ "rank": 682
308
+ },
309
+ "model.layers.15.mlp.down_proj": {
310
+ "bias": true,
311
+ "in_features": 11008,
312
+ "out_features": 4096,
313
+ "rank": 996
314
+ },
315
+ "model.layers.15.mlp.gate_proj": {
316
+ "bias": true,
317
+ "in_features": 4096,
318
+ "out_features": 11008,
319
+ "rank": 665
320
+ },
321
+ "model.layers.15.mlp.up_proj": {
322
+ "bias": true,
323
+ "in_features": 4096,
324
+ "out_features": 11008,
325
+ "rank": 826
326
+ },
327
+ "model.layers.15.self_attn.k_proj": {
328
+ "bias": true,
329
+ "in_features": 4096,
330
+ "out_features": 4096,
331
+ "rank": 126
332
+ },
333
+ "model.layers.15.self_attn.o_proj": {
334
+ "bias": true,
335
+ "in_features": 4096,
336
+ "out_features": 4096,
337
+ "rank": 359
338
+ },
339
+ "model.layers.15.self_attn.q_proj": {
340
+ "bias": true,
341
+ "in_features": 4096,
342
+ "out_features": 4096,
343
+ "rank": 152
344
+ },
345
+ "model.layers.15.self_attn.v_proj": {
346
+ "bias": true,
347
+ "in_features": 4096,
348
+ "out_features": 4096,
349
+ "rank": 455
350
+ },
351
+ "model.layers.16.mlp.down_proj": {
352
+ "bias": true,
353
+ "in_features": 11008,
354
+ "out_features": 4096,
355
+ "rank": 1673
356
+ },
357
+ "model.layers.16.mlp.gate_proj": {
358
+ "bias": true,
359
+ "in_features": 4096,
360
+ "out_features": 11008,
361
+ "rank": 978
362
+ },
363
+ "model.layers.16.mlp.up_proj": {
364
+ "bias": true,
365
+ "in_features": 4096,
366
+ "out_features": 11008,
367
+ "rank": 1109
368
+ },
369
+ "model.layers.16.self_attn.k_proj": {
370
+ "bias": true,
371
+ "in_features": 4096,
372
+ "out_features": 4096,
373
+ "rank": 128
374
+ },
375
+ "model.layers.16.self_attn.o_proj": {
376
+ "bias": true,
377
+ "in_features": 4096,
378
+ "out_features": 4096,
379
+ "rank": 504
380
+ },
381
+ "model.layers.16.self_attn.q_proj": {
382
+ "bias": true,
383
+ "in_features": 4096,
384
+ "out_features": 4096,
385
+ "rank": 159
386
+ },
387
+ "model.layers.16.self_attn.v_proj": {
388
+ "bias": true,
389
+ "in_features": 4096,
390
+ "out_features": 4096,
391
+ "rank": 469
392
+ },
393
+ "model.layers.17.mlp.down_proj": {
394
+ "bias": true,
395
+ "in_features": 11008,
396
+ "out_features": 4096,
397
+ "rank": 1545
398
+ },
399
+ "model.layers.17.mlp.gate_proj": {
400
+ "bias": true,
401
+ "in_features": 4096,
402
+ "out_features": 11008,
403
+ "rank": 905
404
+ },
405
+ "model.layers.17.mlp.up_proj": {
406
+ "bias": true,
407
+ "in_features": 4096,
408
+ "out_features": 11008,
409
+ "rank": 1087
410
+ },
411
+ "model.layers.17.self_attn.k_proj": {
412
+ "bias": true,
413
+ "in_features": 4096,
414
+ "out_features": 4096,
415
+ "rank": 262
416
+ },
417
+ "model.layers.17.self_attn.o_proj": {
418
+ "bias": true,
419
+ "in_features": 4096,
420
+ "out_features": 4096,
421
+ "rank": 1038
422
+ },
423
+ "model.layers.17.self_attn.q_proj": {
424
+ "bias": true,
425
+ "in_features": 4096,
426
+ "out_features": 4096,
427
+ "rank": 327
428
+ },
429
+ "model.layers.17.self_attn.v_proj": {
430
+ "bias": true,
431
+ "in_features": 4096,
432
+ "out_features": 4096,
433
+ "rank": 785
434
+ },
435
+ "model.layers.18.mlp.down_proj": {
436
+ "bias": true,
437
+ "in_features": 11008,
438
+ "out_features": 4096,
439
+ "rank": 1414
440
+ },
441
+ "model.layers.18.mlp.gate_proj": {
442
+ "bias": true,
443
+ "in_features": 4096,
444
+ "out_features": 11008,
445
+ "rank": 851
446
+ },
447
+ "model.layers.18.mlp.up_proj": {
448
+ "bias": true,
449
+ "in_features": 4096,
450
+ "out_features": 11008,
451
+ "rank": 984
452
+ },
453
+ "model.layers.18.self_attn.k_proj": {
454
+ "bias": true,
455
+ "in_features": 4096,
456
+ "out_features": 4096,
457
+ "rank": 298
458
+ },
459
+ "model.layers.18.self_attn.o_proj": {
460
+ "bias": true,
461
+ "in_features": 4096,
462
+ "out_features": 4096,
463
+ "rank": 1130
464
+ },
465
+ "model.layers.18.self_attn.q_proj": {
466
+ "bias": true,
467
+ "in_features": 4096,
468
+ "out_features": 4096,
469
+ "rank": 285
470
+ },
471
+ "model.layers.18.self_attn.v_proj": {
472
+ "bias": true,
473
+ "in_features": 4096,
474
+ "out_features": 4096,
475
+ "rank": 857
476
+ },
477
+ "model.layers.19.mlp.down_proj": {
478
+ "bias": true,
479
+ "in_features": 11008,
480
+ "out_features": 4096,
481
+ "rank": 1348
482
+ },
483
+ "model.layers.19.mlp.gate_proj": {
484
+ "bias": true,
485
+ "in_features": 4096,
486
+ "out_features": 11008,
487
+ "rank": 819
488
+ },
489
+ "model.layers.19.mlp.up_proj": {
490
+ "bias": true,
491
+ "in_features": 4096,
492
+ "out_features": 11008,
493
+ "rank": 999
494
+ },
495
+ "model.layers.19.self_attn.k_proj": {
496
+ "bias": true,
497
+ "in_features": 4096,
498
+ "out_features": 4096,
499
+ "rank": 160
500
+ },
501
+ "model.layers.19.self_attn.o_proj": {
502
+ "bias": true,
503
+ "in_features": 4096,
504
+ "out_features": 4096,
505
+ "rank": 648
506
+ },
507
+ "model.layers.19.self_attn.q_proj": {
508
+ "bias": true,
509
+ "in_features": 4096,
510
+ "out_features": 4096,
511
+ "rank": 151
512
+ },
513
+ "model.layers.19.self_attn.v_proj": {
514
+ "bias": true,
515
+ "in_features": 4096,
516
+ "out_features": 4096,
517
+ "rank": 497
518
+ },
519
+ "model.layers.2.mlp.down_proj": {
520
+ "bias": true,
521
+ "in_features": 11008,
522
+ "out_features": 4096,
523
+ "rank": 1450
524
+ },
525
+ "model.layers.2.mlp.gate_proj": {
526
+ "bias": true,
527
+ "in_features": 4096,
528
+ "out_features": 11008,
529
+ "rank": 1014
530
+ },
531
+ "model.layers.2.mlp.up_proj": {
532
+ "bias": true,
533
+ "in_features": 4096,
534
+ "out_features": 11008,
535
+ "rank": 1167
536
+ },
537
+ "model.layers.2.self_attn.k_proj": {
538
+ "bias": true,
539
+ "in_features": 4096,
540
+ "out_features": 4096,
541
+ "rank": 257
542
+ },
543
+ "model.layers.2.self_attn.o_proj": {
544
+ "bias": true,
545
+ "in_features": 4096,
546
+ "out_features": 4096,
547
+ "rank": 1176
548
+ },
549
+ "model.layers.2.self_attn.q_proj": {
550
+ "bias": true,
551
+ "in_features": 4096,
552
+ "out_features": 4096,
553
+ "rank": 205
554
+ },
555
+ "model.layers.2.self_attn.v_proj": {
556
+ "bias": true,
557
+ "in_features": 4096,
558
+ "out_features": 4096,
559
+ "rank": 862
560
+ },
561
+ "model.layers.20.mlp.down_proj": {
562
+ "bias": true,
563
+ "in_features": 11008,
564
+ "out_features": 4096,
565
+ "rank": 1412
566
+ },
567
+ "model.layers.20.mlp.gate_proj": {
568
+ "bias": true,
569
+ "in_features": 4096,
570
+ "out_features": 11008,
571
+ "rank": 844
572
+ },
573
+ "model.layers.20.mlp.up_proj": {
574
+ "bias": true,
575
+ "in_features": 4096,
576
+ "out_features": 11008,
577
+ "rank": 965
578
+ },
579
+ "model.layers.20.self_attn.k_proj": {
580
+ "bias": true,
581
+ "in_features": 4096,
582
+ "out_features": 4096,
583
+ "rank": 151
584
+ },
585
+ "model.layers.20.self_attn.o_proj": {
586
+ "bias": true,
587
+ "in_features": 4096,
588
+ "out_features": 4096,
589
+ "rank": 685
590
+ },
591
+ "model.layers.20.self_attn.q_proj": {
592
+ "bias": true,
593
+ "in_features": 4096,
594
+ "out_features": 4096,
595
+ "rank": 166
596
+ },
597
+ "model.layers.20.self_attn.v_proj": {
598
+ "bias": true,
599
+ "in_features": 4096,
600
+ "out_features": 4096,
601
+ "rank": 587
602
+ },
603
+ "model.layers.21.mlp.down_proj": {
604
+ "bias": true,
605
+ "in_features": 11008,
606
+ "out_features": 4096,
607
+ "rank": 1355
608
+ },
609
+ "model.layers.21.mlp.gate_proj": {
610
+ "bias": true,
611
+ "in_features": 4096,
612
+ "out_features": 11008,
613
+ "rank": 833
614
+ },
615
+ "model.layers.21.mlp.up_proj": {
616
+ "bias": true,
617
+ "in_features": 4096,
618
+ "out_features": 11008,
619
+ "rank": 908
620
+ },
621
+ "model.layers.21.self_attn.k_proj": {
622
+ "bias": true,
623
+ "in_features": 4096,
624
+ "out_features": 4096,
625
+ "rank": 137
626
+ },
627
+ "model.layers.21.self_attn.o_proj": {
628
+ "bias": true,
629
+ "in_features": 4096,
630
+ "out_features": 4096,
631
+ "rank": 472
632
+ },
633
+ "model.layers.21.self_attn.q_proj": {
634
+ "bias": true,
635
+ "in_features": 4096,
636
+ "out_features": 4096,
637
+ "rank": 124
638
+ },
639
+ "model.layers.21.self_attn.v_proj": {
640
+ "bias": true,
641
+ "in_features": 4096,
642
+ "out_features": 4096,
643
+ "rank": 393
644
+ },
645
+ "model.layers.22.mlp.down_proj": {
646
+ "bias": true,
647
+ "in_features": 11008,
648
+ "out_features": 4096,
649
+ "rank": 1113
650
+ },
651
+ "model.layers.22.mlp.gate_proj": {
652
+ "bias": true,
653
+ "in_features": 4096,
654
+ "out_features": 11008,
655
+ "rank": 667
656
+ },
657
+ "model.layers.22.mlp.up_proj": {
658
+ "bias": true,
659
+ "in_features": 4096,
660
+ "out_features": 11008,
661
+ "rank": 737
662
+ },
663
+ "model.layers.22.self_attn.k_proj": {
664
+ "bias": true,
665
+ "in_features": 4096,
666
+ "out_features": 4096,
667
+ "rank": 92
668
+ },
669
+ "model.layers.22.self_attn.o_proj": {
670
+ "bias": true,
671
+ "in_features": 4096,
672
+ "out_features": 4096,
673
+ "rank": 130
674
+ },
675
+ "model.layers.22.self_attn.q_proj": {
676
+ "bias": true,
677
+ "in_features": 4096,
678
+ "out_features": 4096,
679
+ "rank": 89
680
+ },
681
+ "model.layers.22.self_attn.v_proj": {
682
+ "bias": true,
683
+ "in_features": 4096,
684
+ "out_features": 4096,
685
+ "rank": 150
686
+ },
687
+ "model.layers.23.mlp.down_proj": {
688
+ "bias": true,
689
+ "in_features": 11008,
690
+ "out_features": 4096,
691
+ "rank": 856
692
+ },
693
+ "model.layers.23.mlp.gate_proj": {
694
+ "bias": true,
695
+ "in_features": 4096,
696
+ "out_features": 11008,
697
+ "rank": 478
698
+ },
699
+ "model.layers.23.mlp.up_proj": {
700
+ "bias": true,
701
+ "in_features": 4096,
702
+ "out_features": 11008,
703
+ "rank": 586
704
+ },
705
+ "model.layers.23.self_attn.k_proj": {
706
+ "bias": true,
707
+ "in_features": 4096,
708
+ "out_features": 4096,
709
+ "rank": 168
710
+ },
711
+ "model.layers.23.self_attn.o_proj": {
712
+ "bias": true,
713
+ "in_features": 4096,
714
+ "out_features": 4096,
715
+ "rank": 516
716
+ },
717
+ "model.layers.23.self_attn.q_proj": {
718
+ "bias": true,
719
+ "in_features": 4096,
720
+ "out_features": 4096,
721
+ "rank": 144
722
+ },
723
+ "model.layers.23.self_attn.v_proj": {
724
+ "bias": true,
725
+ "in_features": 4096,
726
+ "out_features": 4096,
727
+ "rank": 356
728
+ },
729
+ "model.layers.24.mlp.down_proj": {
730
+ "bias": true,
731
+ "in_features": 11008,
732
+ "out_features": 4096,
733
+ "rank": 435
734
+ },
735
+ "model.layers.24.mlp.gate_proj": {
736
+ "bias": true,
737
+ "in_features": 4096,
738
+ "out_features": 11008,
739
+ "rank": 242
740
+ },
741
+ "model.layers.24.mlp.up_proj": {
742
+ "bias": true,
743
+ "in_features": 4096,
744
+ "out_features": 11008,
745
+ "rank": 319
746
+ },
747
+ "model.layers.24.self_attn.k_proj": {
748
+ "bias": true,
749
+ "in_features": 4096,
750
+ "out_features": 4096,
751
+ "rank": 96
752
+ },
753
+ "model.layers.24.self_attn.o_proj": {
754
+ "bias": true,
755
+ "in_features": 4096,
756
+ "out_features": 4096,
757
+ "rank": 440
758
+ },
759
+ "model.layers.24.self_attn.q_proj": {
760
+ "bias": true,
761
+ "in_features": 4096,
762
+ "out_features": 4096,
763
+ "rank": 93
764
+ },
765
+ "model.layers.24.self_attn.v_proj": {
766
+ "bias": true,
767
+ "in_features": 4096,
768
+ "out_features": 4096,
769
+ "rank": 297
770
+ },
771
+ "model.layers.25.mlp.down_proj": {
772
+ "bias": true,
773
+ "in_features": 11008,
774
+ "out_features": 4096,
775
+ "rank": 147
776
+ },
777
+ "model.layers.25.mlp.gate_proj": {
778
+ "bias": true,
779
+ "in_features": 4096,
780
+ "out_features": 11008,
781
+ "rank": 121
782
+ },
783
+ "model.layers.25.mlp.up_proj": {
784
+ "bias": true,
785
+ "in_features": 4096,
786
+ "out_features": 11008,
787
+ "rank": 122
788
+ },
789
+ "model.layers.25.self_attn.k_proj": {
790
+ "bias": true,
791
+ "in_features": 4096,
792
+ "out_features": 4096,
793
+ "rank": 11
794
+ },
795
+ "model.layers.25.self_attn.o_proj": {
796
+ "bias": true,
797
+ "in_features": 4096,
798
+ "out_features": 4096,
799
+ "rank": 7
800
+ },
801
+ "model.layers.25.self_attn.q_proj": {
802
+ "bias": true,
803
+ "in_features": 4096,
804
+ "out_features": 4096,
805
+ "rank": 14
806
+ },
807
+ "model.layers.25.self_attn.v_proj": {
808
+ "bias": true,
809
+ "in_features": 4096,
810
+ "out_features": 4096,
811
+ "rank": 13
812
+ },
813
+ "model.layers.26.mlp.down_proj": {
814
+ "bias": true,
815
+ "in_features": 11008,
816
+ "out_features": 4096,
817
+ "rank": 17
818
+ },
819
+ "model.layers.26.mlp.gate_proj": {
820
+ "bias": true,
821
+ "in_features": 4096,
822
+ "out_features": 11008,
823
+ "rank": 36
824
+ },
825
+ "model.layers.26.mlp.up_proj": {
826
+ "bias": true,
827
+ "in_features": 4096,
828
+ "out_features": 11008,
829
+ "rank": 20
830
+ },
831
+ "model.layers.26.self_attn.k_proj": {
832
+ "bias": true,
833
+ "in_features": 4096,
834
+ "out_features": 4096,
835
+ "rank": 2
836
+ },
837
+ "model.layers.26.self_attn.o_proj": {
838
+ "bias": true,
839
+ "in_features": 4096,
840
+ "out_features": 4096,
841
+ "rank": 1
842
+ },
843
+ "model.layers.26.self_attn.q_proj": {
844
+ "bias": true,
845
+ "in_features": 4096,
846
+ "out_features": 4096,
847
+ "rank": 4
848
+ },
849
+ "model.layers.26.self_attn.v_proj": {
850
+ "bias": true,
851
+ "in_features": 4096,
852
+ "out_features": 4096,
853
+ "rank": 1
854
+ },
855
+ "model.layers.27.mlp.down_proj": {
856
+ "bias": true,
857
+ "in_features": 11008,
858
+ "out_features": 4096,
859
+ "rank": 1
860
+ },
861
+ "model.layers.27.mlp.gate_proj": {
862
+ "bias": true,
863
+ "in_features": 4096,
864
+ "out_features": 11008,
865
+ "rank": 15
866
+ },
867
+ "model.layers.27.mlp.up_proj": {
868
+ "bias": true,
869
+ "in_features": 4096,
870
+ "out_features": 11008,
871
+ "rank": 6
872
+ },
873
+ "model.layers.27.self_attn.k_proj": {
874
+ "bias": true,
875
+ "in_features": 4096,
876
+ "out_features": 4096,
877
+ "rank": 5
878
+ },
879
+ "model.layers.27.self_attn.o_proj": {
880
+ "bias": true,
881
+ "in_features": 4096,
882
+ "out_features": 4096,
883
+ "rank": 1
884
+ },
885
+ "model.layers.27.self_attn.q_proj": {
886
+ "bias": true,
887
+ "in_features": 4096,
888
+ "out_features": 4096,
889
+ "rank": 4
890
+ },
891
+ "model.layers.27.self_attn.v_proj": {
892
+ "bias": true,
893
+ "in_features": 4096,
894
+ "out_features": 4096,
895
+ "rank": 8
896
+ },
897
+ "model.layers.28.mlp.down_proj": {
898
+ "bias": true,
899
+ "in_features": 11008,
900
+ "out_features": 4096,
901
+ "rank": 304
902
+ },
903
+ "model.layers.28.mlp.gate_proj": {
904
+ "bias": true,
905
+ "in_features": 4096,
906
+ "out_features": 11008,
907
+ "rank": 248
908
+ },
909
+ "model.layers.28.mlp.up_proj": {
910
+ "bias": true,
911
+ "in_features": 4096,
912
+ "out_features": 11008,
913
+ "rank": 213
914
+ },
915
+ "model.layers.28.self_attn.k_proj": {
916
+ "bias": true,
917
+ "in_features": 4096,
918
+ "out_features": 4096,
919
+ "rank": 65
920
+ },
921
+ "model.layers.28.self_attn.o_proj": {
922
+ "bias": true,
923
+ "in_features": 4096,
924
+ "out_features": 4096,
925
+ "rank": 174
926
+ },
927
+ "model.layers.28.self_attn.q_proj": {
928
+ "bias": true,
929
+ "in_features": 4096,
930
+ "out_features": 4096,
931
+ "rank": 80
932
+ },
933
+ "model.layers.28.self_attn.v_proj": {
934
+ "bias": true,
935
+ "in_features": 4096,
936
+ "out_features": 4096,
937
+ "rank": 238
938
+ },
939
+ "model.layers.29.mlp.down_proj": {
940
+ "bias": true,
941
+ "in_features": 11008,
942
+ "out_features": 4096,
943
+ "rank": 436
944
+ },
945
+ "model.layers.29.mlp.gate_proj": {
946
+ "bias": true,
947
+ "in_features": 4096,
948
+ "out_features": 11008,
949
+ "rank": 430
950
+ },
951
+ "model.layers.29.mlp.up_proj": {
952
+ "bias": true,
953
+ "in_features": 4096,
954
+ "out_features": 11008,
955
+ "rank": 334
956
+ },
957
+ "model.layers.29.self_attn.k_proj": {
958
+ "bias": true,
959
+ "in_features": 4096,
960
+ "out_features": 4096,
961
+ "rank": 1
962
+ },
963
+ "model.layers.29.self_attn.o_proj": {
964
+ "bias": true,
965
+ "in_features": 4096,
966
+ "out_features": 4096,
967
+ "rank": 1
968
+ },
969
+ "model.layers.29.self_attn.q_proj": {
970
+ "bias": true,
971
+ "in_features": 4096,
972
+ "out_features": 4096,
973
+ "rank": 2
974
+ },
975
+ "model.layers.29.self_attn.v_proj": {
976
+ "bias": true,
977
+ "in_features": 4096,
978
+ "out_features": 4096,
979
+ "rank": 2
980
+ },
981
+ "model.layers.3.mlp.down_proj": {
982
+ "bias": true,
983
+ "in_features": 11008,
984
+ "out_features": 4096,
985
+ "rank": 1216
986
+ },
987
+ "model.layers.3.mlp.gate_proj": {
988
+ "bias": true,
989
+ "in_features": 4096,
990
+ "out_features": 11008,
991
+ "rank": 1042
992
+ },
993
+ "model.layers.3.mlp.up_proj": {
994
+ "bias": true,
995
+ "in_features": 4096,
996
+ "out_features": 11008,
997
+ "rank": 1217
998
+ },
999
+ "model.layers.3.self_attn.k_proj": {
1000
+ "bias": true,
1001
+ "in_features": 4096,
1002
+ "out_features": 4096,
1003
+ "rank": 228
1004
+ },
1005
+ "model.layers.3.self_attn.o_proj": {
1006
+ "bias": true,
1007
+ "in_features": 4096,
1008
+ "out_features": 4096,
1009
+ "rank": 627
1010
+ },
1011
+ "model.layers.3.self_attn.q_proj": {
1012
+ "bias": true,
1013
+ "in_features": 4096,
1014
+ "out_features": 4096,
1015
+ "rank": 203
1016
+ },
1017
+ "model.layers.3.self_attn.v_proj": {
1018
+ "bias": true,
1019
+ "in_features": 4096,
1020
+ "out_features": 4096,
1021
+ "rank": 574
1022
+ },
1023
+ "model.layers.30.mlp.down_proj": {
1024
+ "bias": true,
1025
+ "in_features": 11008,
1026
+ "out_features": 4096,
1027
+ "rank": 973
1028
+ },
1029
+ "model.layers.30.mlp.gate_proj": {
1030
+ "bias": true,
1031
+ "in_features": 4096,
1032
+ "out_features": 11008,
1033
+ "rank": 747
1034
+ },
1035
+ "model.layers.30.mlp.up_proj": {
1036
+ "bias": true,
1037
+ "in_features": 4096,
1038
+ "out_features": 11008,
1039
+ "rank": 683
1040
+ },
1041
+ "model.layers.30.self_attn.k_proj": {
1042
+ "bias": true,
1043
+ "in_features": 4096,
1044
+ "out_features": 4096,
1045
+ "rank": 14
1046
+ },
1047
+ "model.layers.30.self_attn.o_proj": {
1048
+ "bias": true,
1049
+ "in_features": 4096,
1050
+ "out_features": 4096,
1051
+ "rank": 10
1052
+ },
1053
+ "model.layers.30.self_attn.q_proj": {
1054
+ "bias": true,
1055
+ "in_features": 4096,
1056
+ "out_features": 4096,
1057
+ "rank": 14
1058
+ },
1059
+ "model.layers.30.self_attn.v_proj": {
1060
+ "bias": true,
1061
+ "in_features": 4096,
1062
+ "out_features": 4096,
1063
+ "rank": 22
1064
+ },
1065
+ "model.layers.31.mlp.down_proj": {
1066
+ "bias": true,
1067
+ "in_features": 11008,
1068
+ "out_features": 4096,
1069
+ "rank": 2035
1070
+ },
1071
+ "model.layers.31.mlp.gate_proj": {
1072
+ "bias": true,
1073
+ "in_features": 4096,
1074
+ "out_features": 11008,
1075
+ "rank": 1270
1076
+ },
1077
+ "model.layers.31.mlp.up_proj": {
1078
+ "bias": true,
1079
+ "in_features": 4096,
1080
+ "out_features": 11008,
1081
+ "rank": 1335
1082
+ },
1083
+ "model.layers.31.self_attn.k_proj": {
1084
+ "bias": true,
1085
+ "in_features": 4096,
1086
+ "out_features": 4096,
1087
+ "rank": 59
1088
+ },
1089
+ "model.layers.31.self_attn.o_proj": {
1090
+ "bias": true,
1091
+ "in_features": 4096,
1092
+ "out_features": 4096,
1093
+ "rank": 214
1094
+ },
1095
+ "model.layers.31.self_attn.q_proj": {
1096
+ "bias": true,
1097
+ "in_features": 4096,
1098
+ "out_features": 4096,
1099
+ "rank": 60
1100
+ },
1101
+ "model.layers.31.self_attn.v_proj": {
1102
+ "bias": true,
1103
+ "in_features": 4096,
1104
+ "out_features": 4096,
1105
+ "rank": 284
1106
+ },
1107
+ "model.layers.4.mlp.down_proj": {
1108
+ "bias": true,
1109
+ "in_features": 11008,
1110
+ "out_features": 4096,
1111
+ "rank": 960
1112
+ },
1113
+ "model.layers.4.mlp.gate_proj": {
1114
+ "bias": true,
1115
+ "in_features": 4096,
1116
+ "out_features": 11008,
1117
+ "rank": 730
1118
+ },
1119
+ "model.layers.4.mlp.up_proj": {
1120
+ "bias": true,
1121
+ "in_features": 4096,
1122
+ "out_features": 11008,
1123
+ "rank": 962
1124
+ },
1125
+ "model.layers.4.self_attn.k_proj": {
1126
+ "bias": true,
1127
+ "in_features": 4096,
1128
+ "out_features": 4096,
1129
+ "rank": 271
1130
+ },
1131
+ "model.layers.4.self_attn.o_proj": {
1132
+ "bias": true,
1133
+ "in_features": 4096,
1134
+ "out_features": 4096,
1135
+ "rank": 966
1136
+ },
1137
+ "model.layers.4.self_attn.q_proj": {
1138
+ "bias": true,
1139
+ "in_features": 4096,
1140
+ "out_features": 4096,
1141
+ "rank": 266
1142
+ },
1143
+ "model.layers.4.self_attn.v_proj": {
1144
+ "bias": true,
1145
+ "in_features": 4096,
1146
+ "out_features": 4096,
1147
+ "rank": 699
1148
+ },
1149
+ "model.layers.5.mlp.down_proj": {
1150
+ "bias": true,
1151
+ "in_features": 11008,
1152
+ "out_features": 4096,
1153
+ "rank": 906
1154
+ },
1155
+ "model.layers.5.mlp.gate_proj": {
1156
+ "bias": true,
1157
+ "in_features": 4096,
1158
+ "out_features": 11008,
1159
+ "rank": 654
1160
+ },
1161
+ "model.layers.5.mlp.up_proj": {
1162
+ "bias": true,
1163
+ "in_features": 4096,
1164
+ "out_features": 11008,
1165
+ "rank": 1019
1166
+ },
1167
+ "model.layers.5.self_attn.k_proj": {
1168
+ "bias": true,
1169
+ "in_features": 4096,
1170
+ "out_features": 4096,
1171
+ "rank": 243
1172
+ },
1173
+ "model.layers.5.self_attn.o_proj": {
1174
+ "bias": true,
1175
+ "in_features": 4096,
1176
+ "out_features": 4096,
1177
+ "rank": 503
1178
+ },
1179
+ "model.layers.5.self_attn.q_proj": {
1180
+ "bias": true,
1181
+ "in_features": 4096,
1182
+ "out_features": 4096,
1183
+ "rank": 233
1184
+ },
1185
+ "model.layers.5.self_attn.v_proj": {
1186
+ "bias": true,
1187
+ "in_features": 4096,
1188
+ "out_features": 4096,
1189
+ "rank": 470
1190
+ },
1191
+ "model.layers.6.mlp.down_proj": {
1192
+ "bias": true,
1193
+ "in_features": 11008,
1194
+ "out_features": 4096,
1195
+ "rank": 740
1196
+ },
1197
+ "model.layers.6.mlp.gate_proj": {
1198
+ "bias": true,
1199
+ "in_features": 4096,
1200
+ "out_features": 11008,
1201
+ "rank": 535
1202
+ },
1203
+ "model.layers.6.mlp.up_proj": {
1204
+ "bias": true,
1205
+ "in_features": 4096,
1206
+ "out_features": 11008,
1207
+ "rank": 826
1208
+ },
1209
+ "model.layers.6.self_attn.k_proj": {
1210
+ "bias": true,
1211
+ "in_features": 4096,
1212
+ "out_features": 4096,
1213
+ "rank": 248
1214
+ },
1215
+ "model.layers.6.self_attn.o_proj": {
1216
+ "bias": true,
1217
+ "in_features": 4096,
1218
+ "out_features": 4096,
1219
+ "rank": 456
1220
+ },
1221
+ "model.layers.6.self_attn.q_proj": {
1222
+ "bias": true,
1223
+ "in_features": 4096,
1224
+ "out_features": 4096,
1225
+ "rank": 237
1226
+ },
1227
+ "model.layers.6.self_attn.v_proj": {
1228
+ "bias": true,
1229
+ "in_features": 4096,
1230
+ "out_features": 4096,
1231
+ "rank": 370
1232
+ },
1233
+ "model.layers.7.mlp.down_proj": {
1234
+ "bias": true,
1235
+ "in_features": 11008,
1236
+ "out_features": 4096,
1237
+ "rank": 437
1238
+ },
1239
+ "model.layers.7.mlp.gate_proj": {
1240
+ "bias": true,
1241
+ "in_features": 4096,
1242
+ "out_features": 11008,
1243
+ "rank": 317
1244
+ },
1245
+ "model.layers.7.mlp.up_proj": {
1246
+ "bias": true,
1247
+ "in_features": 4096,
1248
+ "out_features": 11008,
1249
+ "rank": 494
1250
+ },
1251
+ "model.layers.7.self_attn.k_proj": {
1252
+ "bias": true,
1253
+ "in_features": 4096,
1254
+ "out_features": 4096,
1255
+ "rank": 244
1256
+ },
1257
+ "model.layers.7.self_attn.o_proj": {
1258
+ "bias": true,
1259
+ "in_features": 4096,
1260
+ "out_features": 4096,
1261
+ "rank": 625
1262
+ },
1263
+ "model.layers.7.self_attn.q_proj": {
1264
+ "bias": true,
1265
+ "in_features": 4096,
1266
+ "out_features": 4096,
1267
+ "rank": 251
1268
+ },
1269
+ "model.layers.7.self_attn.v_proj": {
1270
+ "bias": true,
1271
+ "in_features": 4096,
1272
+ "out_features": 4096,
1273
+ "rank": 545
1274
+ },
1275
+ "model.layers.8.mlp.down_proj": {
1276
+ "bias": true,
1277
+ "in_features": 11008,
1278
+ "out_features": 4096,
1279
+ "rank": 201
1280
+ },
1281
+ "model.layers.8.mlp.gate_proj": {
1282
+ "bias": true,
1283
+ "in_features": 4096,
1284
+ "out_features": 11008,
1285
+ "rank": 164
1286
+ },
1287
+ "model.layers.8.mlp.up_proj": {
1288
+ "bias": true,
1289
+ "in_features": 4096,
1290
+ "out_features": 11008,
1291
+ "rank": 203
1292
+ },
1293
+ "model.layers.8.self_attn.k_proj": {
1294
+ "bias": true,
1295
+ "in_features": 4096,
1296
+ "out_features": 4096,
1297
+ "rank": 167
1298
+ },
1299
+ "model.layers.8.self_attn.o_proj": {
1300
+ "bias": true,
1301
+ "in_features": 4096,
1302
+ "out_features": 4096,
1303
+ "rank": 195
1304
+ },
1305
+ "model.layers.8.self_attn.q_proj": {
1306
+ "bias": true,
1307
+ "in_features": 4096,
1308
+ "out_features": 4096,
1309
+ "rank": 149
1310
+ },
1311
+ "model.layers.8.self_attn.v_proj": {
1312
+ "bias": true,
1313
+ "in_features": 4096,
1314
+ "out_features": 4096,
1315
+ "rank": 287
1316
+ },
1317
+ "model.layers.9.mlp.down_proj": {
1318
+ "bias": true,
1319
+ "in_features": 11008,
1320
+ "out_features": 4096,
1321
+ "rank": 93
1322
+ },
1323
+ "model.layers.9.mlp.gate_proj": {
1324
+ "bias": true,
1325
+ "in_features": 4096,
1326
+ "out_features": 11008,
1327
+ "rank": 118
1328
+ },
1329
+ "model.layers.9.mlp.up_proj": {
1330
+ "bias": true,
1331
+ "in_features": 4096,
1332
+ "out_features": 11008,
1333
+ "rank": 89
1334
+ },
1335
+ "model.layers.9.self_attn.k_proj": {
1336
+ "bias": true,
1337
+ "in_features": 4096,
1338
+ "out_features": 4096,
1339
+ "rank": 212
1340
+ },
1341
+ "model.layers.9.self_attn.o_proj": {
1342
+ "bias": true,
1343
+ "in_features": 4096,
1344
+ "out_features": 4096,
1345
+ "rank": 284
1346
+ },
1347
+ "model.layers.9.self_attn.q_proj": {
1348
+ "bias": true,
1349
+ "in_features": 4096,
1350
+ "out_features": 4096,
1351
+ "rank": 199
1352
+ },
1353
+ "model.layers.9.self_attn.v_proj": {
1354
+ "bias": true,
1355
+ "in_features": 4096,
1356
+ "out_features": 4096,
1357
+ "rank": 235
1358
+ }
1359
+ },
1360
+ "max_position_embeddings": 2048,
1361
+ "mlp_bias": false,
1362
+ "model_type": "llama",
1363
+ "num_attention_heads": 32,
1364
+ "num_hidden_layers": 32,
1365
+ "num_key_value_heads": 32,
1366
+ "pad_token_id": 0,
1367
+ "pretraining_tp": 1,
1368
+ "rms_norm_eps": 1e-06,
1369
+ "rope_scaling": null,
1370
+ "rope_theta": 10000.0,
1371
+ "tie_word_embeddings": false,
1372
+ "torch_dtype": "bfloat16",
1373
+ "transformers_version": "4.43.2",
1374
+ "use_cache": true,
1375
+ "vocab_size": 32000
1376
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.43.2"
7
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cb06b02ef8441db6bf60c6839c4d21dcc3a146bdcafd18424ace89c2420044a1
3
+ size 3110904216
special_tokens_map.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "<unk>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "unk_token": {
24
+ "content": "<unk>",
25
+ "lstrip": false,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ }
30
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
tokenizer_config.json ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": false,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<unk>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<s>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "2": {
23
+ "content": "</s>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": true
29
+ }
30
+ },
31
+ "bos_token": "<s>",
32
+ "clean_up_tokenization_spaces": false,
33
+ "eos_token": "</s>",
34
+ "legacy": true,
35
+ "model_max_length": 1000000000000000019884624838656,
36
+ "pad_token": "<unk>",
37
+ "padding_side": "right",
38
+ "sp_model_kwargs": {},
39
+ "spaces_between_special_tokens": false,
40
+ "tokenizer_class": "LlamaTokenizer",
41
+ "unk_token": "<unk>",
42
+ "use_default_system_prompt": false
43
+ }