ruihanglai commited on
Commit
3e49c0b
·
verified ·
1 Parent(s): 37a515a

Upload tensor-cache.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. tensor-cache.json +1614 -0
tensor-cache.json ADDED
@@ -0,0 +1,1614 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "ParamSize": 149,
4
+ "ParamBytes": 326074368.0,
5
+ "BitsPerParam": 16.0
6
+ },
7
+ "records": [
8
+ {
9
+ "dataPath": "params_shard_0.bin",
10
+ "format": "raw-shard",
11
+ "nbytes": 77194752,
12
+ "records": [
13
+ {
14
+ "name": "lm_head.weight",
15
+ "shape": [
16
+ 50257,
17
+ 768
18
+ ],
19
+ "dtype": "float16",
20
+ "format": "f32-to-bf16",
21
+ "nbytes": 77194752,
22
+ "byteOffset": 0
23
+ }
24
+ ],
25
+ "md5sum": "8aa629e8739ff337f4983e485d0a145e"
26
+ },
27
+ {
28
+ "dataPath": "params_shard_1.bin",
29
+ "format": "raw-shard",
30
+ "nbytes": 77194752,
31
+ "records": [
32
+ {
33
+ "name": "transformer.wte.weight",
34
+ "shape": [
35
+ 50257,
36
+ 768
37
+ ],
38
+ "dtype": "float16",
39
+ "format": "f32-to-bf16",
40
+ "nbytes": 77194752,
41
+ "byteOffset": 0
42
+ }
43
+ ],
44
+ "md5sum": "8aa629e8739ff337f4983e485d0a145e"
45
+ },
46
+ {
47
+ "dataPath": "params_shard_2.bin",
48
+ "format": "raw-shard",
49
+ "nbytes": 33470976,
50
+ "records": [
51
+ {
52
+ "name": "transformer.wpe.weight",
53
+ "shape": [
54
+ 1024,
55
+ 768
56
+ ],
57
+ "dtype": "float16",
58
+ "format": "f32-to-bf16",
59
+ "nbytes": 1572864,
60
+ "byteOffset": 0
61
+ },
62
+ {
63
+ "name": "transformer.h.0.ln_1.weight",
64
+ "shape": [
65
+ 768
66
+ ],
67
+ "dtype": "float16",
68
+ "format": "f32-to-bf16",
69
+ "nbytes": 1536,
70
+ "byteOffset": 1572864
71
+ },
72
+ {
73
+ "name": "transformer.h.0.ln_1.bias",
74
+ "shape": [
75
+ 768
76
+ ],
77
+ "dtype": "float16",
78
+ "format": "f32-to-bf16",
79
+ "nbytes": 1536,
80
+ "byteOffset": 1574400
81
+ },
82
+ {
83
+ "name": "transformer.h.0.attn.c_attn.weight",
84
+ "shape": [
85
+ 2304,
86
+ 768
87
+ ],
88
+ "dtype": "float16",
89
+ "format": "f32-to-bf16",
90
+ "nbytes": 3538944,
91
+ "byteOffset": 1575936
92
+ },
93
+ {
94
+ "name": "transformer.h.0.attn.c_attn.bias",
95
+ "shape": [
96
+ 2304
97
+ ],
98
+ "dtype": "float16",
99
+ "format": "f32-to-bf16",
100
+ "nbytes": 4608,
101
+ "byteOffset": 5114880
102
+ },
103
+ {
104
+ "name": "transformer.h.0.attn.c_proj.weight",
105
+ "shape": [
106
+ 768,
107
+ 768
108
+ ],
109
+ "dtype": "float16",
110
+ "format": "f32-to-bf16",
111
+ "nbytes": 1179648,
112
+ "byteOffset": 5119488
113
+ },
114
+ {
115
+ "name": "transformer.h.0.attn.c_proj.bias",
116
+ "shape": [
117
+ 768
118
+ ],
119
+ "dtype": "float16",
120
+ "format": "f32-to-bf16",
121
+ "nbytes": 1536,
122
+ "byteOffset": 6299136
123
+ },
124
+ {
125
+ "name": "transformer.h.0.ln_2.weight",
126
+ "shape": [
127
+ 768
128
+ ],
129
+ "dtype": "float16",
130
+ "format": "f32-to-bf16",
131
+ "nbytes": 1536,
132
+ "byteOffset": 6300672
133
+ },
134
+ {
135
+ "name": "transformer.h.0.ln_2.bias",
136
+ "shape": [
137
+ 768
138
+ ],
139
+ "dtype": "float16",
140
+ "format": "f32-to-bf16",
141
+ "nbytes": 1536,
142
+ "byteOffset": 6302208
143
+ },
144
+ {
145
+ "name": "transformer.h.0.mlp.c_fc.weight",
146
+ "shape": [
147
+ 3072,
148
+ 768
149
+ ],
150
+ "dtype": "float16",
151
+ "format": "f32-to-bf16",
152
+ "nbytes": 4718592,
153
+ "byteOffset": 6303744
154
+ },
155
+ {
156
+ "name": "transformer.h.0.mlp.c_fc.bias",
157
+ "shape": [
158
+ 3072
159
+ ],
160
+ "dtype": "float16",
161
+ "format": "f32-to-bf16",
162
+ "nbytes": 6144,
163
+ "byteOffset": 11022336
164
+ },
165
+ {
166
+ "name": "transformer.h.0.mlp.c_proj.weight",
167
+ "shape": [
168
+ 768,
169
+ 3072
170
+ ],
171
+ "dtype": "float16",
172
+ "format": "f32-to-bf16",
173
+ "nbytes": 4718592,
174
+ "byteOffset": 11028480
175
+ },
176
+ {
177
+ "name": "transformer.h.0.mlp.c_proj.bias",
178
+ "shape": [
179
+ 768
180
+ ],
181
+ "dtype": "float16",
182
+ "format": "f32-to-bf16",
183
+ "nbytes": 1536,
184
+ "byteOffset": 15747072
185
+ },
186
+ {
187
+ "name": "transformer.h.1.ln_1.weight",
188
+ "shape": [
189
+ 768
190
+ ],
191
+ "dtype": "float16",
192
+ "format": "f32-to-bf16",
193
+ "nbytes": 1536,
194
+ "byteOffset": 15748608
195
+ },
196
+ {
197
+ "name": "transformer.h.1.ln_1.bias",
198
+ "shape": [
199
+ 768
200
+ ],
201
+ "dtype": "float16",
202
+ "format": "f32-to-bf16",
203
+ "nbytes": 1536,
204
+ "byteOffset": 15750144
205
+ },
206
+ {
207
+ "name": "transformer.h.1.attn.c_attn.weight",
208
+ "shape": [
209
+ 2304,
210
+ 768
211
+ ],
212
+ "dtype": "float16",
213
+ "format": "f32-to-bf16",
214
+ "nbytes": 3538944,
215
+ "byteOffset": 15751680
216
+ },
217
+ {
218
+ "name": "transformer.h.1.attn.c_attn.bias",
219
+ "shape": [
220
+ 2304
221
+ ],
222
+ "dtype": "float16",
223
+ "format": "f32-to-bf16",
224
+ "nbytes": 4608,
225
+ "byteOffset": 19290624
226
+ },
227
+ {
228
+ "name": "transformer.h.1.attn.c_proj.weight",
229
+ "shape": [
230
+ 768,
231
+ 768
232
+ ],
233
+ "dtype": "float16",
234
+ "format": "f32-to-bf16",
235
+ "nbytes": 1179648,
236
+ "byteOffset": 19295232
237
+ },
238
+ {
239
+ "name": "transformer.h.1.attn.c_proj.bias",
240
+ "shape": [
241
+ 768
242
+ ],
243
+ "dtype": "float16",
244
+ "format": "f32-to-bf16",
245
+ "nbytes": 1536,
246
+ "byteOffset": 20474880
247
+ },
248
+ {
249
+ "name": "transformer.h.1.ln_2.weight",
250
+ "shape": [
251
+ 768
252
+ ],
253
+ "dtype": "float16",
254
+ "format": "f32-to-bf16",
255
+ "nbytes": 1536,
256
+ "byteOffset": 20476416
257
+ },
258
+ {
259
+ "name": "transformer.h.1.ln_2.bias",
260
+ "shape": [
261
+ 768
262
+ ],
263
+ "dtype": "float16",
264
+ "format": "f32-to-bf16",
265
+ "nbytes": 1536,
266
+ "byteOffset": 20477952
267
+ },
268
+ {
269
+ "name": "transformer.h.1.mlp.c_fc.weight",
270
+ "shape": [
271
+ 3072,
272
+ 768
273
+ ],
274
+ "dtype": "float16",
275
+ "format": "f32-to-bf16",
276
+ "nbytes": 4718592,
277
+ "byteOffset": 20479488
278
+ },
279
+ {
280
+ "name": "transformer.h.1.mlp.c_fc.bias",
281
+ "shape": [
282
+ 3072
283
+ ],
284
+ "dtype": "float16",
285
+ "format": "f32-to-bf16",
286
+ "nbytes": 6144,
287
+ "byteOffset": 25198080
288
+ },
289
+ {
290
+ "name": "transformer.h.1.mlp.c_proj.weight",
291
+ "shape": [
292
+ 768,
293
+ 3072
294
+ ],
295
+ "dtype": "float16",
296
+ "format": "f32-to-bf16",
297
+ "nbytes": 4718592,
298
+ "byteOffset": 25204224
299
+ },
300
+ {
301
+ "name": "transformer.h.1.mlp.c_proj.bias",
302
+ "shape": [
303
+ 768
304
+ ],
305
+ "dtype": "float16",
306
+ "format": "f32-to-bf16",
307
+ "nbytes": 1536,
308
+ "byteOffset": 29922816
309
+ },
310
+ {
311
+ "name": "transformer.h.2.ln_1.weight",
312
+ "shape": [
313
+ 768
314
+ ],
315
+ "dtype": "float16",
316
+ "format": "f32-to-bf16",
317
+ "nbytes": 1536,
318
+ "byteOffset": 29924352
319
+ },
320
+ {
321
+ "name": "transformer.h.2.ln_1.bias",
322
+ "shape": [
323
+ 768
324
+ ],
325
+ "dtype": "float16",
326
+ "format": "f32-to-bf16",
327
+ "nbytes": 1536,
328
+ "byteOffset": 29925888
329
+ },
330
+ {
331
+ "name": "transformer.h.2.attn.c_attn.weight",
332
+ "shape": [
333
+ 2304,
334
+ 768
335
+ ],
336
+ "dtype": "float16",
337
+ "format": "f32-to-bf16",
338
+ "nbytes": 3538944,
339
+ "byteOffset": 29927424
340
+ },
341
+ {
342
+ "name": "transformer.h.2.attn.c_attn.bias",
343
+ "shape": [
344
+ 2304
345
+ ],
346
+ "dtype": "float16",
347
+ "format": "f32-to-bf16",
348
+ "nbytes": 4608,
349
+ "byteOffset": 33466368
350
+ }
351
+ ],
352
+ "md5sum": "748d75aca928d18aac58d8f9301c3bcb"
353
+ },
354
+ {
355
+ "dataPath": "params_shard_3.bin",
356
+ "format": "raw-shard",
357
+ "nbytes": 29535744,
358
+ "records": [
359
+ {
360
+ "name": "transformer.h.2.attn.c_proj.weight",
361
+ "shape": [
362
+ 768,
363
+ 768
364
+ ],
365
+ "dtype": "float16",
366
+ "format": "f32-to-bf16",
367
+ "nbytes": 1179648,
368
+ "byteOffset": 0
369
+ },
370
+ {
371
+ "name": "transformer.h.2.attn.c_proj.bias",
372
+ "shape": [
373
+ 768
374
+ ],
375
+ "dtype": "float16",
376
+ "format": "f32-to-bf16",
377
+ "nbytes": 1536,
378
+ "byteOffset": 1179648
379
+ },
380
+ {
381
+ "name": "transformer.h.2.ln_2.weight",
382
+ "shape": [
383
+ 768
384
+ ],
385
+ "dtype": "float16",
386
+ "format": "f32-to-bf16",
387
+ "nbytes": 1536,
388
+ "byteOffset": 1181184
389
+ },
390
+ {
391
+ "name": "transformer.h.2.ln_2.bias",
392
+ "shape": [
393
+ 768
394
+ ],
395
+ "dtype": "float16",
396
+ "format": "f32-to-bf16",
397
+ "nbytes": 1536,
398
+ "byteOffset": 1182720
399
+ },
400
+ {
401
+ "name": "transformer.h.2.mlp.c_fc.weight",
402
+ "shape": [
403
+ 3072,
404
+ 768
405
+ ],
406
+ "dtype": "float16",
407
+ "format": "f32-to-bf16",
408
+ "nbytes": 4718592,
409
+ "byteOffset": 1184256
410
+ },
411
+ {
412
+ "name": "transformer.h.2.mlp.c_fc.bias",
413
+ "shape": [
414
+ 3072
415
+ ],
416
+ "dtype": "float16",
417
+ "format": "f32-to-bf16",
418
+ "nbytes": 6144,
419
+ "byteOffset": 5902848
420
+ },
421
+ {
422
+ "name": "transformer.h.2.mlp.c_proj.weight",
423
+ "shape": [
424
+ 768,
425
+ 3072
426
+ ],
427
+ "dtype": "float16",
428
+ "format": "f32-to-bf16",
429
+ "nbytes": 4718592,
430
+ "byteOffset": 5908992
431
+ },
432
+ {
433
+ "name": "transformer.h.2.mlp.c_proj.bias",
434
+ "shape": [
435
+ 768
436
+ ],
437
+ "dtype": "float16",
438
+ "format": "f32-to-bf16",
439
+ "nbytes": 1536,
440
+ "byteOffset": 10627584
441
+ },
442
+ {
443
+ "name": "transformer.h.3.ln_1.weight",
444
+ "shape": [
445
+ 768
446
+ ],
447
+ "dtype": "float16",
448
+ "format": "f32-to-bf16",
449
+ "nbytes": 1536,
450
+ "byteOffset": 10629120
451
+ },
452
+ {
453
+ "name": "transformer.h.3.ln_1.bias",
454
+ "shape": [
455
+ 768
456
+ ],
457
+ "dtype": "float16",
458
+ "format": "f32-to-bf16",
459
+ "nbytes": 1536,
460
+ "byteOffset": 10630656
461
+ },
462
+ {
463
+ "name": "transformer.h.3.attn.c_attn.weight",
464
+ "shape": [
465
+ 2304,
466
+ 768
467
+ ],
468
+ "dtype": "float16",
469
+ "format": "f32-to-bf16",
470
+ "nbytes": 3538944,
471
+ "byteOffset": 10632192
472
+ },
473
+ {
474
+ "name": "transformer.h.3.attn.c_attn.bias",
475
+ "shape": [
476
+ 2304
477
+ ],
478
+ "dtype": "float16",
479
+ "format": "f32-to-bf16",
480
+ "nbytes": 4608,
481
+ "byteOffset": 14171136
482
+ },
483
+ {
484
+ "name": "transformer.h.3.attn.c_proj.weight",
485
+ "shape": [
486
+ 768,
487
+ 768
488
+ ],
489
+ "dtype": "float16",
490
+ "format": "f32-to-bf16",
491
+ "nbytes": 1179648,
492
+ "byteOffset": 14175744
493
+ },
494
+ {
495
+ "name": "transformer.h.3.attn.c_proj.bias",
496
+ "shape": [
497
+ 768
498
+ ],
499
+ "dtype": "float16",
500
+ "format": "f32-to-bf16",
501
+ "nbytes": 1536,
502
+ "byteOffset": 15355392
503
+ },
504
+ {
505
+ "name": "transformer.h.3.ln_2.weight",
506
+ "shape": [
507
+ 768
508
+ ],
509
+ "dtype": "float16",
510
+ "format": "f32-to-bf16",
511
+ "nbytes": 1536,
512
+ "byteOffset": 15356928
513
+ },
514
+ {
515
+ "name": "transformer.h.3.ln_2.bias",
516
+ "shape": [
517
+ 768
518
+ ],
519
+ "dtype": "float16",
520
+ "format": "f32-to-bf16",
521
+ "nbytes": 1536,
522
+ "byteOffset": 15358464
523
+ },
524
+ {
525
+ "name": "transformer.h.3.mlp.c_fc.weight",
526
+ "shape": [
527
+ 3072,
528
+ 768
529
+ ],
530
+ "dtype": "float16",
531
+ "format": "f32-to-bf16",
532
+ "nbytes": 4718592,
533
+ "byteOffset": 15360000
534
+ },
535
+ {
536
+ "name": "transformer.h.3.mlp.c_fc.bias",
537
+ "shape": [
538
+ 3072
539
+ ],
540
+ "dtype": "float16",
541
+ "format": "f32-to-bf16",
542
+ "nbytes": 6144,
543
+ "byteOffset": 20078592
544
+ },
545
+ {
546
+ "name": "transformer.h.3.mlp.c_proj.weight",
547
+ "shape": [
548
+ 768,
549
+ 3072
550
+ ],
551
+ "dtype": "float16",
552
+ "format": "f32-to-bf16",
553
+ "nbytes": 4718592,
554
+ "byteOffset": 20084736
555
+ },
556
+ {
557
+ "name": "transformer.h.3.mlp.c_proj.bias",
558
+ "shape": [
559
+ 768
560
+ ],
561
+ "dtype": "float16",
562
+ "format": "f32-to-bf16",
563
+ "nbytes": 1536,
564
+ "byteOffset": 24803328
565
+ },
566
+ {
567
+ "name": "transformer.h.4.ln_1.weight",
568
+ "shape": [
569
+ 768
570
+ ],
571
+ "dtype": "float16",
572
+ "format": "f32-to-bf16",
573
+ "nbytes": 1536,
574
+ "byteOffset": 24804864
575
+ },
576
+ {
577
+ "name": "transformer.h.4.ln_1.bias",
578
+ "shape": [
579
+ 768
580
+ ],
581
+ "dtype": "float16",
582
+ "format": "f32-to-bf16",
583
+ "nbytes": 1536,
584
+ "byteOffset": 24806400
585
+ },
586
+ {
587
+ "name": "transformer.h.4.attn.c_attn.weight",
588
+ "shape": [
589
+ 2304,
590
+ 768
591
+ ],
592
+ "dtype": "float16",
593
+ "format": "f32-to-bf16",
594
+ "nbytes": 3538944,
595
+ "byteOffset": 24807936
596
+ },
597
+ {
598
+ "name": "transformer.h.4.attn.c_attn.bias",
599
+ "shape": [
600
+ 2304
601
+ ],
602
+ "dtype": "float16",
603
+ "format": "f32-to-bf16",
604
+ "nbytes": 4608,
605
+ "byteOffset": 28346880
606
+ },
607
+ {
608
+ "name": "transformer.h.4.attn.c_proj.weight",
609
+ "shape": [
610
+ 768,
611
+ 768
612
+ ],
613
+ "dtype": "float16",
614
+ "format": "f32-to-bf16",
615
+ "nbytes": 1179648,
616
+ "byteOffset": 28351488
617
+ },
618
+ {
619
+ "name": "transformer.h.4.attn.c_proj.bias",
620
+ "shape": [
621
+ 768
622
+ ],
623
+ "dtype": "float16",
624
+ "format": "f32-to-bf16",
625
+ "nbytes": 1536,
626
+ "byteOffset": 29531136
627
+ },
628
+ {
629
+ "name": "transformer.h.4.ln_2.weight",
630
+ "shape": [
631
+ 768
632
+ ],
633
+ "dtype": "float16",
634
+ "format": "f32-to-bf16",
635
+ "nbytes": 1536,
636
+ "byteOffset": 29532672
637
+ },
638
+ {
639
+ "name": "transformer.h.4.ln_2.bias",
640
+ "shape": [
641
+ 768
642
+ ],
643
+ "dtype": "float16",
644
+ "format": "f32-to-bf16",
645
+ "nbytes": 1536,
646
+ "byteOffset": 29534208
647
+ }
648
+ ],
649
+ "md5sum": "94d1cea00ea6a1492cb7bb23eb5d64b5"
650
+ },
651
+ {
652
+ "dataPath": "params_shard_4.bin",
653
+ "format": "raw-shard",
654
+ "nbytes": 33076224,
655
+ "records": [
656
+ {
657
+ "name": "transformer.h.4.mlp.c_fc.weight",
658
+ "shape": [
659
+ 3072,
660
+ 768
661
+ ],
662
+ "dtype": "float16",
663
+ "format": "f32-to-bf16",
664
+ "nbytes": 4718592,
665
+ "byteOffset": 0
666
+ },
667
+ {
668
+ "name": "transformer.h.4.mlp.c_fc.bias",
669
+ "shape": [
670
+ 3072
671
+ ],
672
+ "dtype": "float16",
673
+ "format": "f32-to-bf16",
674
+ "nbytes": 6144,
675
+ "byteOffset": 4718592
676
+ },
677
+ {
678
+ "name": "transformer.h.4.mlp.c_proj.weight",
679
+ "shape": [
680
+ 768,
681
+ 3072
682
+ ],
683
+ "dtype": "float16",
684
+ "format": "f32-to-bf16",
685
+ "nbytes": 4718592,
686
+ "byteOffset": 4724736
687
+ },
688
+ {
689
+ "name": "transformer.h.4.mlp.c_proj.bias",
690
+ "shape": [
691
+ 768
692
+ ],
693
+ "dtype": "float16",
694
+ "format": "f32-to-bf16",
695
+ "nbytes": 1536,
696
+ "byteOffset": 9443328
697
+ },
698
+ {
699
+ "name": "transformer.h.5.ln_1.weight",
700
+ "shape": [
701
+ 768
702
+ ],
703
+ "dtype": "float16",
704
+ "format": "f32-to-bf16",
705
+ "nbytes": 1536,
706
+ "byteOffset": 9444864
707
+ },
708
+ {
709
+ "name": "transformer.h.5.ln_1.bias",
710
+ "shape": [
711
+ 768
712
+ ],
713
+ "dtype": "float16",
714
+ "format": "f32-to-bf16",
715
+ "nbytes": 1536,
716
+ "byteOffset": 9446400
717
+ },
718
+ {
719
+ "name": "transformer.h.5.attn.c_attn.weight",
720
+ "shape": [
721
+ 2304,
722
+ 768
723
+ ],
724
+ "dtype": "float16",
725
+ "format": "f32-to-bf16",
726
+ "nbytes": 3538944,
727
+ "byteOffset": 9447936
728
+ },
729
+ {
730
+ "name": "transformer.h.5.attn.c_attn.bias",
731
+ "shape": [
732
+ 2304
733
+ ],
734
+ "dtype": "float16",
735
+ "format": "f32-to-bf16",
736
+ "nbytes": 4608,
737
+ "byteOffset": 12986880
738
+ },
739
+ {
740
+ "name": "transformer.h.5.attn.c_proj.weight",
741
+ "shape": [
742
+ 768,
743
+ 768
744
+ ],
745
+ "dtype": "float16",
746
+ "format": "f32-to-bf16",
747
+ "nbytes": 1179648,
748
+ "byteOffset": 12991488
749
+ },
750
+ {
751
+ "name": "transformer.h.5.attn.c_proj.bias",
752
+ "shape": [
753
+ 768
754
+ ],
755
+ "dtype": "float16",
756
+ "format": "f32-to-bf16",
757
+ "nbytes": 1536,
758
+ "byteOffset": 14171136
759
+ },
760
+ {
761
+ "name": "transformer.h.5.ln_2.weight",
762
+ "shape": [
763
+ 768
764
+ ],
765
+ "dtype": "float16",
766
+ "format": "f32-to-bf16",
767
+ "nbytes": 1536,
768
+ "byteOffset": 14172672
769
+ },
770
+ {
771
+ "name": "transformer.h.5.ln_2.bias",
772
+ "shape": [
773
+ 768
774
+ ],
775
+ "dtype": "float16",
776
+ "format": "f32-to-bf16",
777
+ "nbytes": 1536,
778
+ "byteOffset": 14174208
779
+ },
780
+ {
781
+ "name": "transformer.h.5.mlp.c_fc.weight",
782
+ "shape": [
783
+ 3072,
784
+ 768
785
+ ],
786
+ "dtype": "float16",
787
+ "format": "f32-to-bf16",
788
+ "nbytes": 4718592,
789
+ "byteOffset": 14175744
790
+ },
791
+ {
792
+ "name": "transformer.h.5.mlp.c_fc.bias",
793
+ "shape": [
794
+ 3072
795
+ ],
796
+ "dtype": "float16",
797
+ "format": "f32-to-bf16",
798
+ "nbytes": 6144,
799
+ "byteOffset": 18894336
800
+ },
801
+ {
802
+ "name": "transformer.h.5.mlp.c_proj.weight",
803
+ "shape": [
804
+ 768,
805
+ 3072
806
+ ],
807
+ "dtype": "float16",
808
+ "format": "f32-to-bf16",
809
+ "nbytes": 4718592,
810
+ "byteOffset": 18900480
811
+ },
812
+ {
813
+ "name": "transformer.h.5.mlp.c_proj.bias",
814
+ "shape": [
815
+ 768
816
+ ],
817
+ "dtype": "float16",
818
+ "format": "f32-to-bf16",
819
+ "nbytes": 1536,
820
+ "byteOffset": 23619072
821
+ },
822
+ {
823
+ "name": "transformer.h.6.ln_1.weight",
824
+ "shape": [
825
+ 768
826
+ ],
827
+ "dtype": "float16",
828
+ "format": "f32-to-bf16",
829
+ "nbytes": 1536,
830
+ "byteOffset": 23620608
831
+ },
832
+ {
833
+ "name": "transformer.h.6.ln_1.bias",
834
+ "shape": [
835
+ 768
836
+ ],
837
+ "dtype": "float16",
838
+ "format": "f32-to-bf16",
839
+ "nbytes": 1536,
840
+ "byteOffset": 23622144
841
+ },
842
+ {
843
+ "name": "transformer.h.6.attn.c_attn.weight",
844
+ "shape": [
845
+ 2304,
846
+ 768
847
+ ],
848
+ "dtype": "float16",
849
+ "format": "f32-to-bf16",
850
+ "nbytes": 3538944,
851
+ "byteOffset": 23623680
852
+ },
853
+ {
854
+ "name": "transformer.h.6.attn.c_attn.bias",
855
+ "shape": [
856
+ 2304
857
+ ],
858
+ "dtype": "float16",
859
+ "format": "f32-to-bf16",
860
+ "nbytes": 4608,
861
+ "byteOffset": 27162624
862
+ },
863
+ {
864
+ "name": "transformer.h.6.attn.c_proj.weight",
865
+ "shape": [
866
+ 768,
867
+ 768
868
+ ],
869
+ "dtype": "float16",
870
+ "format": "f32-to-bf16",
871
+ "nbytes": 1179648,
872
+ "byteOffset": 27167232
873
+ },
874
+ {
875
+ "name": "transformer.h.6.attn.c_proj.bias",
876
+ "shape": [
877
+ 768
878
+ ],
879
+ "dtype": "float16",
880
+ "format": "f32-to-bf16",
881
+ "nbytes": 1536,
882
+ "byteOffset": 28346880
883
+ },
884
+ {
885
+ "name": "transformer.h.6.ln_2.weight",
886
+ "shape": [
887
+ 768
888
+ ],
889
+ "dtype": "float16",
890
+ "format": "f32-to-bf16",
891
+ "nbytes": 1536,
892
+ "byteOffset": 28348416
893
+ },
894
+ {
895
+ "name": "transformer.h.6.ln_2.bias",
896
+ "shape": [
897
+ 768
898
+ ],
899
+ "dtype": "float16",
900
+ "format": "f32-to-bf16",
901
+ "nbytes": 1536,
902
+ "byteOffset": 28349952
903
+ },
904
+ {
905
+ "name": "transformer.h.6.mlp.c_fc.weight",
906
+ "shape": [
907
+ 3072,
908
+ 768
909
+ ],
910
+ "dtype": "float16",
911
+ "format": "f32-to-bf16",
912
+ "nbytes": 4718592,
913
+ "byteOffset": 28351488
914
+ },
915
+ {
916
+ "name": "transformer.h.6.mlp.c_fc.bias",
917
+ "shape": [
918
+ 3072
919
+ ],
920
+ "dtype": "float16",
921
+ "format": "f32-to-bf16",
922
+ "nbytes": 6144,
923
+ "byteOffset": 33070080
924
+ }
925
+ ],
926
+ "md5sum": "43957cf0e0da3bc016aa35099e8a4e53"
927
+ },
928
+ {
929
+ "dataPath": "params_shard_5.bin",
930
+ "format": "raw-shard",
931
+ "nbytes": 33074688,
932
+ "records": [
933
+ {
934
+ "name": "transformer.h.6.mlp.c_proj.weight",
935
+ "shape": [
936
+ 768,
937
+ 3072
938
+ ],
939
+ "dtype": "float16",
940
+ "format": "f32-to-bf16",
941
+ "nbytes": 4718592,
942
+ "byteOffset": 0
943
+ },
944
+ {
945
+ "name": "transformer.h.6.mlp.c_proj.bias",
946
+ "shape": [
947
+ 768
948
+ ],
949
+ "dtype": "float16",
950
+ "format": "f32-to-bf16",
951
+ "nbytes": 1536,
952
+ "byteOffset": 4718592
953
+ },
954
+ {
955
+ "name": "transformer.h.7.ln_1.weight",
956
+ "shape": [
957
+ 768
958
+ ],
959
+ "dtype": "float16",
960
+ "format": "f32-to-bf16",
961
+ "nbytes": 1536,
962
+ "byteOffset": 4720128
963
+ },
964
+ {
965
+ "name": "transformer.h.7.ln_1.bias",
966
+ "shape": [
967
+ 768
968
+ ],
969
+ "dtype": "float16",
970
+ "format": "f32-to-bf16",
971
+ "nbytes": 1536,
972
+ "byteOffset": 4721664
973
+ },
974
+ {
975
+ "name": "transformer.h.7.attn.c_attn.weight",
976
+ "shape": [
977
+ 2304,
978
+ 768
979
+ ],
980
+ "dtype": "float16",
981
+ "format": "f32-to-bf16",
982
+ "nbytes": 3538944,
983
+ "byteOffset": 4723200
984
+ },
985
+ {
986
+ "name": "transformer.h.7.attn.c_attn.bias",
987
+ "shape": [
988
+ 2304
989
+ ],
990
+ "dtype": "float16",
991
+ "format": "f32-to-bf16",
992
+ "nbytes": 4608,
993
+ "byteOffset": 8262144
994
+ },
995
+ {
996
+ "name": "transformer.h.7.attn.c_proj.weight",
997
+ "shape": [
998
+ 768,
999
+ 768
1000
+ ],
1001
+ "dtype": "float16",
1002
+ "format": "f32-to-bf16",
1003
+ "nbytes": 1179648,
1004
+ "byteOffset": 8266752
1005
+ },
1006
+ {
1007
+ "name": "transformer.h.7.attn.c_proj.bias",
1008
+ "shape": [
1009
+ 768
1010
+ ],
1011
+ "dtype": "float16",
1012
+ "format": "f32-to-bf16",
1013
+ "nbytes": 1536,
1014
+ "byteOffset": 9446400
1015
+ },
1016
+ {
1017
+ "name": "transformer.h.7.ln_2.weight",
1018
+ "shape": [
1019
+ 768
1020
+ ],
1021
+ "dtype": "float16",
1022
+ "format": "f32-to-bf16",
1023
+ "nbytes": 1536,
1024
+ "byteOffset": 9447936
1025
+ },
1026
+ {
1027
+ "name": "transformer.h.7.ln_2.bias",
1028
+ "shape": [
1029
+ 768
1030
+ ],
1031
+ "dtype": "float16",
1032
+ "format": "f32-to-bf16",
1033
+ "nbytes": 1536,
1034
+ "byteOffset": 9449472
1035
+ },
1036
+ {
1037
+ "name": "transformer.h.7.mlp.c_fc.weight",
1038
+ "shape": [
1039
+ 3072,
1040
+ 768
1041
+ ],
1042
+ "dtype": "float16",
1043
+ "format": "f32-to-bf16",
1044
+ "nbytes": 4718592,
1045
+ "byteOffset": 9451008
1046
+ },
1047
+ {
1048
+ "name": "transformer.h.7.mlp.c_fc.bias",
1049
+ "shape": [
1050
+ 3072
1051
+ ],
1052
+ "dtype": "float16",
1053
+ "format": "f32-to-bf16",
1054
+ "nbytes": 6144,
1055
+ "byteOffset": 14169600
1056
+ },
1057
+ {
1058
+ "name": "transformer.h.7.mlp.c_proj.weight",
1059
+ "shape": [
1060
+ 768,
1061
+ 3072
1062
+ ],
1063
+ "dtype": "float16",
1064
+ "format": "f32-to-bf16",
1065
+ "nbytes": 4718592,
1066
+ "byteOffset": 14175744
1067
+ },
1068
+ {
1069
+ "name": "transformer.h.7.mlp.c_proj.bias",
1070
+ "shape": [
1071
+ 768
1072
+ ],
1073
+ "dtype": "float16",
1074
+ "format": "f32-to-bf16",
1075
+ "nbytes": 1536,
1076
+ "byteOffset": 18894336
1077
+ },
1078
+ {
1079
+ "name": "transformer.h.8.ln_1.weight",
1080
+ "shape": [
1081
+ 768
1082
+ ],
1083
+ "dtype": "float16",
1084
+ "format": "f32-to-bf16",
1085
+ "nbytes": 1536,
1086
+ "byteOffset": 18895872
1087
+ },
1088
+ {
1089
+ "name": "transformer.h.8.ln_1.bias",
1090
+ "shape": [
1091
+ 768
1092
+ ],
1093
+ "dtype": "float16",
1094
+ "format": "f32-to-bf16",
1095
+ "nbytes": 1536,
1096
+ "byteOffset": 18897408
1097
+ },
1098
+ {
1099
+ "name": "transformer.h.8.attn.c_attn.weight",
1100
+ "shape": [
1101
+ 2304,
1102
+ 768
1103
+ ],
1104
+ "dtype": "float16",
1105
+ "format": "f32-to-bf16",
1106
+ "nbytes": 3538944,
1107
+ "byteOffset": 18898944
1108
+ },
1109
+ {
1110
+ "name": "transformer.h.8.attn.c_attn.bias",
1111
+ "shape": [
1112
+ 2304
1113
+ ],
1114
+ "dtype": "float16",
1115
+ "format": "f32-to-bf16",
1116
+ "nbytes": 4608,
1117
+ "byteOffset": 22437888
1118
+ },
1119
+ {
1120
+ "name": "transformer.h.8.attn.c_proj.weight",
1121
+ "shape": [
1122
+ 768,
1123
+ 768
1124
+ ],
1125
+ "dtype": "float16",
1126
+ "format": "f32-to-bf16",
1127
+ "nbytes": 1179648,
1128
+ "byteOffset": 22442496
1129
+ },
1130
+ {
1131
+ "name": "transformer.h.8.attn.c_proj.bias",
1132
+ "shape": [
1133
+ 768
1134
+ ],
1135
+ "dtype": "float16",
1136
+ "format": "f32-to-bf16",
1137
+ "nbytes": 1536,
1138
+ "byteOffset": 23622144
1139
+ },
1140
+ {
1141
+ "name": "transformer.h.8.ln_2.weight",
1142
+ "shape": [
1143
+ 768
1144
+ ],
1145
+ "dtype": "float16",
1146
+ "format": "f32-to-bf16",
1147
+ "nbytes": 1536,
1148
+ "byteOffset": 23623680
1149
+ },
1150
+ {
1151
+ "name": "transformer.h.8.ln_2.bias",
1152
+ "shape": [
1153
+ 768
1154
+ ],
1155
+ "dtype": "float16",
1156
+ "format": "f32-to-bf16",
1157
+ "nbytes": 1536,
1158
+ "byteOffset": 23625216
1159
+ },
1160
+ {
1161
+ "name": "transformer.h.8.mlp.c_fc.weight",
1162
+ "shape": [
1163
+ 3072,
1164
+ 768
1165
+ ],
1166
+ "dtype": "float16",
1167
+ "format": "f32-to-bf16",
1168
+ "nbytes": 4718592,
1169
+ "byteOffset": 23626752
1170
+ },
1171
+ {
1172
+ "name": "transformer.h.8.mlp.c_fc.bias",
1173
+ "shape": [
1174
+ 3072
1175
+ ],
1176
+ "dtype": "float16",
1177
+ "format": "f32-to-bf16",
1178
+ "nbytes": 6144,
1179
+ "byteOffset": 28345344
1180
+ },
1181
+ {
1182
+ "name": "transformer.h.8.mlp.c_proj.weight",
1183
+ "shape": [
1184
+ 768,
1185
+ 3072
1186
+ ],
1187
+ "dtype": "float16",
1188
+ "format": "f32-to-bf16",
1189
+ "nbytes": 4718592,
1190
+ "byteOffset": 28351488
1191
+ },
1192
+ {
1193
+ "name": "transformer.h.8.mlp.c_proj.bias",
1194
+ "shape": [
1195
+ 768
1196
+ ],
1197
+ "dtype": "float16",
1198
+ "format": "f32-to-bf16",
1199
+ "nbytes": 1536,
1200
+ "byteOffset": 33070080
1201
+ },
1202
+ {
1203
+ "name": "transformer.h.9.ln_1.weight",
1204
+ "shape": [
1205
+ 768
1206
+ ],
1207
+ "dtype": "float16",
1208
+ "format": "f32-to-bf16",
1209
+ "nbytes": 1536,
1210
+ "byteOffset": 33071616
1211
+ },
1212
+ {
1213
+ "name": "transformer.h.9.ln_1.bias",
1214
+ "shape": [
1215
+ 768
1216
+ ],
1217
+ "dtype": "float16",
1218
+ "format": "f32-to-bf16",
1219
+ "nbytes": 1536,
1220
+ "byteOffset": 33073152
1221
+ }
1222
+ ],
1223
+ "md5sum": "4cbbe2cce00b37c72cd95d22b3aa03d3"
1224
+ },
1225
+ {
1226
+ "dataPath": "params_shard_6.bin",
1227
+ "format": "raw-shard",
1228
+ "nbytes": 33079296,
1229
+ "records": [
1230
+ {
1231
+ "name": "transformer.h.9.attn.c_attn.weight",
1232
+ "shape": [
1233
+ 2304,
1234
+ 768
1235
+ ],
1236
+ "dtype": "float16",
1237
+ "format": "f32-to-bf16",
1238
+ "nbytes": 3538944,
1239
+ "byteOffset": 0
1240
+ },
1241
+ {
1242
+ "name": "transformer.h.9.attn.c_attn.bias",
1243
+ "shape": [
1244
+ 2304
1245
+ ],
1246
+ "dtype": "float16",
1247
+ "format": "f32-to-bf16",
1248
+ "nbytes": 4608,
1249
+ "byteOffset": 3538944
1250
+ },
1251
+ {
1252
+ "name": "transformer.h.9.attn.c_proj.weight",
1253
+ "shape": [
1254
+ 768,
1255
+ 768
1256
+ ],
1257
+ "dtype": "float16",
1258
+ "format": "f32-to-bf16",
1259
+ "nbytes": 1179648,
1260
+ "byteOffset": 3543552
1261
+ },
1262
+ {
1263
+ "name": "transformer.h.9.attn.c_proj.bias",
1264
+ "shape": [
1265
+ 768
1266
+ ],
1267
+ "dtype": "float16",
1268
+ "format": "f32-to-bf16",
1269
+ "nbytes": 1536,
1270
+ "byteOffset": 4723200
1271
+ },
1272
+ {
1273
+ "name": "transformer.h.9.ln_2.weight",
1274
+ "shape": [
1275
+ 768
1276
+ ],
1277
+ "dtype": "float16",
1278
+ "format": "f32-to-bf16",
1279
+ "nbytes": 1536,
1280
+ "byteOffset": 4724736
1281
+ },
1282
+ {
1283
+ "name": "transformer.h.9.ln_2.bias",
1284
+ "shape": [
1285
+ 768
1286
+ ],
1287
+ "dtype": "float16",
1288
+ "format": "f32-to-bf16",
1289
+ "nbytes": 1536,
1290
+ "byteOffset": 4726272
1291
+ },
1292
+ {
1293
+ "name": "transformer.h.9.mlp.c_fc.weight",
1294
+ "shape": [
1295
+ 3072,
1296
+ 768
1297
+ ],
1298
+ "dtype": "float16",
1299
+ "format": "f32-to-bf16",
1300
+ "nbytes": 4718592,
1301
+ "byteOffset": 4727808
1302
+ },
1303
+ {
1304
+ "name": "transformer.h.9.mlp.c_fc.bias",
1305
+ "shape": [
1306
+ 3072
1307
+ ],
1308
+ "dtype": "float16",
1309
+ "format": "f32-to-bf16",
1310
+ "nbytes": 6144,
1311
+ "byteOffset": 9446400
1312
+ },
1313
+ {
1314
+ "name": "transformer.h.9.mlp.c_proj.weight",
1315
+ "shape": [
1316
+ 768,
1317
+ 3072
1318
+ ],
1319
+ "dtype": "float16",
1320
+ "format": "f32-to-bf16",
1321
+ "nbytes": 4718592,
1322
+ "byteOffset": 9452544
1323
+ },
1324
+ {
1325
+ "name": "transformer.h.9.mlp.c_proj.bias",
1326
+ "shape": [
1327
+ 768
1328
+ ],
1329
+ "dtype": "float16",
1330
+ "format": "f32-to-bf16",
1331
+ "nbytes": 1536,
1332
+ "byteOffset": 14171136
1333
+ },
1334
+ {
1335
+ "name": "transformer.h.10.ln_1.weight",
1336
+ "shape": [
1337
+ 768
1338
+ ],
1339
+ "dtype": "float16",
1340
+ "format": "f32-to-bf16",
1341
+ "nbytes": 1536,
1342
+ "byteOffset": 14172672
1343
+ },
1344
+ {
1345
+ "name": "transformer.h.10.ln_1.bias",
1346
+ "shape": [
1347
+ 768
1348
+ ],
1349
+ "dtype": "float16",
1350
+ "format": "f32-to-bf16",
1351
+ "nbytes": 1536,
1352
+ "byteOffset": 14174208
1353
+ },
1354
+ {
1355
+ "name": "transformer.h.10.attn.c_attn.weight",
1356
+ "shape": [
1357
+ 2304,
1358
+ 768
1359
+ ],
1360
+ "dtype": "float16",
1361
+ "format": "f32-to-bf16",
1362
+ "nbytes": 3538944,
1363
+ "byteOffset": 14175744
1364
+ },
1365
+ {
1366
+ "name": "transformer.h.10.attn.c_attn.bias",
1367
+ "shape": [
1368
+ 2304
1369
+ ],
1370
+ "dtype": "float16",
1371
+ "format": "f32-to-bf16",
1372
+ "nbytes": 4608,
1373
+ "byteOffset": 17714688
1374
+ },
1375
+ {
1376
+ "name": "transformer.h.10.attn.c_proj.weight",
1377
+ "shape": [
1378
+ 768,
1379
+ 768
1380
+ ],
1381
+ "dtype": "float16",
1382
+ "format": "f32-to-bf16",
1383
+ "nbytes": 1179648,
1384
+ "byteOffset": 17719296
1385
+ },
1386
+ {
1387
+ "name": "transformer.h.10.attn.c_proj.bias",
1388
+ "shape": [
1389
+ 768
1390
+ ],
1391
+ "dtype": "float16",
1392
+ "format": "f32-to-bf16",
1393
+ "nbytes": 1536,
1394
+ "byteOffset": 18898944
1395
+ },
1396
+ {
1397
+ "name": "transformer.h.10.ln_2.weight",
1398
+ "shape": [
1399
+ 768
1400
+ ],
1401
+ "dtype": "float16",
1402
+ "format": "f32-to-bf16",
1403
+ "nbytes": 1536,
1404
+ "byteOffset": 18900480
1405
+ },
1406
+ {
1407
+ "name": "transformer.h.10.ln_2.bias",
1408
+ "shape": [
1409
+ 768
1410
+ ],
1411
+ "dtype": "float16",
1412
+ "format": "f32-to-bf16",
1413
+ "nbytes": 1536,
1414
+ "byteOffset": 18902016
1415
+ },
1416
+ {
1417
+ "name": "transformer.h.10.mlp.c_fc.weight",
1418
+ "shape": [
1419
+ 3072,
1420
+ 768
1421
+ ],
1422
+ "dtype": "float16",
1423
+ "format": "f32-to-bf16",
1424
+ "nbytes": 4718592,
1425
+ "byteOffset": 18903552
1426
+ },
1427
+ {
1428
+ "name": "transformer.h.10.mlp.c_fc.bias",
1429
+ "shape": [
1430
+ 3072
1431
+ ],
1432
+ "dtype": "float16",
1433
+ "format": "f32-to-bf16",
1434
+ "nbytes": 6144,
1435
+ "byteOffset": 23622144
1436
+ },
1437
+ {
1438
+ "name": "transformer.h.10.mlp.c_proj.weight",
1439
+ "shape": [
1440
+ 768,
1441
+ 3072
1442
+ ],
1443
+ "dtype": "float16",
1444
+ "format": "f32-to-bf16",
1445
+ "nbytes": 4718592,
1446
+ "byteOffset": 23628288
1447
+ },
1448
+ {
1449
+ "name": "transformer.h.10.mlp.c_proj.bias",
1450
+ "shape": [
1451
+ 768
1452
+ ],
1453
+ "dtype": "float16",
1454
+ "format": "f32-to-bf16",
1455
+ "nbytes": 1536,
1456
+ "byteOffset": 28346880
1457
+ },
1458
+ {
1459
+ "name": "transformer.h.11.ln_1.weight",
1460
+ "shape": [
1461
+ 768
1462
+ ],
1463
+ "dtype": "float16",
1464
+ "format": "f32-to-bf16",
1465
+ "nbytes": 1536,
1466
+ "byteOffset": 28348416
1467
+ },
1468
+ {
1469
+ "name": "transformer.h.11.ln_1.bias",
1470
+ "shape": [
1471
+ 768
1472
+ ],
1473
+ "dtype": "float16",
1474
+ "format": "f32-to-bf16",
1475
+ "nbytes": 1536,
1476
+ "byteOffset": 28349952
1477
+ },
1478
+ {
1479
+ "name": "transformer.h.11.attn.c_attn.weight",
1480
+ "shape": [
1481
+ 2304,
1482
+ 768
1483
+ ],
1484
+ "dtype": "float16",
1485
+ "format": "f32-to-bf16",
1486
+ "nbytes": 3538944,
1487
+ "byteOffset": 28351488
1488
+ },
1489
+ {
1490
+ "name": "transformer.h.11.attn.c_attn.bias",
1491
+ "shape": [
1492
+ 2304
1493
+ ],
1494
+ "dtype": "float16",
1495
+ "format": "f32-to-bf16",
1496
+ "nbytes": 4608,
1497
+ "byteOffset": 31890432
1498
+ },
1499
+ {
1500
+ "name": "transformer.h.11.attn.c_proj.weight",
1501
+ "shape": [
1502
+ 768,
1503
+ 768
1504
+ ],
1505
+ "dtype": "float16",
1506
+ "format": "f32-to-bf16",
1507
+ "nbytes": 1179648,
1508
+ "byteOffset": 31895040
1509
+ },
1510
+ {
1511
+ "name": "transformer.h.11.attn.c_proj.bias",
1512
+ "shape": [
1513
+ 768
1514
+ ],
1515
+ "dtype": "float16",
1516
+ "format": "f32-to-bf16",
1517
+ "nbytes": 1536,
1518
+ "byteOffset": 33074688
1519
+ },
1520
+ {
1521
+ "name": "transformer.h.11.ln_2.weight",
1522
+ "shape": [
1523
+ 768
1524
+ ],
1525
+ "dtype": "float16",
1526
+ "format": "f32-to-bf16",
1527
+ "nbytes": 1536,
1528
+ "byteOffset": 33076224
1529
+ },
1530
+ {
1531
+ "name": "transformer.h.11.ln_2.bias",
1532
+ "shape": [
1533
+ 768
1534
+ ],
1535
+ "dtype": "float16",
1536
+ "format": "f32-to-bf16",
1537
+ "nbytes": 1536,
1538
+ "byteOffset": 33077760
1539
+ }
1540
+ ],
1541
+ "md5sum": "4f1aa298f8b0af477effc227d22f5ef1"
1542
+ },
1543
+ {
1544
+ "dataPath": "params_shard_7.bin",
1545
+ "format": "raw-shard",
1546
+ "nbytes": 9447936,
1547
+ "records": [
1548
+ {
1549
+ "name": "transformer.h.11.mlp.c_fc.weight",
1550
+ "shape": [
1551
+ 3072,
1552
+ 768
1553
+ ],
1554
+ "dtype": "float16",
1555
+ "format": "f32-to-bf16",
1556
+ "nbytes": 4718592,
1557
+ "byteOffset": 0
1558
+ },
1559
+ {
1560
+ "name": "transformer.h.11.mlp.c_fc.bias",
1561
+ "shape": [
1562
+ 3072
1563
+ ],
1564
+ "dtype": "float16",
1565
+ "format": "f32-to-bf16",
1566
+ "nbytes": 6144,
1567
+ "byteOffset": 4718592
1568
+ },
1569
+ {
1570
+ "name": "transformer.h.11.mlp.c_proj.weight",
1571
+ "shape": [
1572
+ 768,
1573
+ 3072
1574
+ ],
1575
+ "dtype": "float16",
1576
+ "format": "f32-to-bf16",
1577
+ "nbytes": 4718592,
1578
+ "byteOffset": 4724736
1579
+ },
1580
+ {
1581
+ "name": "transformer.h.11.mlp.c_proj.bias",
1582
+ "shape": [
1583
+ 768
1584
+ ],
1585
+ "dtype": "float16",
1586
+ "format": "f32-to-bf16",
1587
+ "nbytes": 1536,
1588
+ "byteOffset": 9443328
1589
+ },
1590
+ {
1591
+ "name": "transformer.ln_f.weight",
1592
+ "shape": [
1593
+ 768
1594
+ ],
1595
+ "dtype": "float16",
1596
+ "format": "f32-to-bf16",
1597
+ "nbytes": 1536,
1598
+ "byteOffset": 9444864
1599
+ },
1600
+ {
1601
+ "name": "transformer.ln_f.bias",
1602
+ "shape": [
1603
+ 768
1604
+ ],
1605
+ "dtype": "float16",
1606
+ "format": "f32-to-bf16",
1607
+ "nbytes": 1536,
1608
+ "byteOffset": 9446400
1609
+ }
1610
+ ],
1611
+ "md5sum": "1532478e2b064b31091ff718a8c67188"
1612
+ }
1613
+ ]
1614
+ }