Jhinslain/jhinslain-1b-webgpu
0
1{
2 "metadata": {
3 "ParamSize": 283,
4 "ParamBytes": 1807423488.0,
5 "BitsPerParam": 4.500626782697164
6 },
7 "records": [
8 {
9 "dataPath": "params_shard_0.bin",
10 "format": "raw-shard",
11 "nbytes": 197001216,
12 "records": [
13 {
14 "name": "model.embed_tokens.q_weight",
15 "shape": [
16 128256,
17 384
18 ],
19 "dtype": "uint32",
20 "format": "f32-to-bf16",
21 "nbytes": 197001216,
22 "byteOffset": 0
23 }
24 ],
25 "md5sum": "1d23d906fea7be657da47864b6dbe0fb"
26 },
27 {
28 "dataPath": "params_shard_1.bin",
29 "format": "raw-shard",
30 "nbytes": 24631296,
31 "records": [
32 {
33 "name": "model.embed_tokens.q_scale",
34 "shape": [
35 128256,
36 96
37 ],
38 "dtype": "float16",
39 "format": "f32-to-bf16",
40 "nbytes": 24625152,
41 "byteOffset": 0
42 },
43 {
44 "name": "model.layers.0.input_layernorm.weight",
45 "shape": [
46 3072
47 ],
48 "dtype": "float16",
49 "format": "f32-to-bf16",
50 "nbytes": 6144,
51 "byteOffset": 24625152
52 }
53 ],
54 "md5sum": "5ac94bd3abfff44fc23c41e36d66afaa"
55 },
56 {
57 "dataPath": "params_shard_2.bin",
58 "format": "raw-shard",
59 "nbytes": 25165824,
60 "records": [
61 {
62 "name": "model.layers.0.mlp.gate_up_proj.q_weight",
63 "shape": [
64 16384,
65 384
66 ],
67 "dtype": "uint32",
68 "format": "f32-to-bf16",
69 "nbytes": 25165824,
70 "byteOffset": 0
71 }
72 ],
73 "md5sum": "ce4a4953a1f0a4bb8d451aa94f0e64fb"
74 },
75 {
76 "dataPath": "params_shard_3.bin",
77 "format": "raw-shard",
78 "nbytes": 31469568,
79 "records": [
80 {
81 "name": "model.layers.0.mlp.down_proj.q_weight",
82 "shape": [
83 3072,
84 1024
85 ],
86 "dtype": "uint32",
87 "format": "f32-to-bf16",
88 "nbytes": 12582912,
89 "byteOffset": 0
90 },
91 {
92 "name": "model.layers.0.mlp.down_proj.q_scale",
93 "shape": [
94 3072,
95 256
96 ],
97 "dtype": "float16",
98 "format": "f32-to-bf16",
99 "nbytes": 1572864,
100 "byteOffset": 12582912
101 },
102 {
103 "name": "model.layers.0.mlp.gate_up_proj.q_scale",
104 "shape": [
105 16384,
106 96
107 ],
108 "dtype": "float16",
109 "format": "f32-to-bf16",
110 "nbytes": 3145728,
111 "byteOffset": 14155776
112 },
113 {
114 "name": "model.layers.0.post_attention_layernorm.weight",
115 "shape": [
116 3072
117 ],
118 "dtype": "float16",
119 "format": "f32-to-bf16",
120 "nbytes": 6144,
121 "byteOffset": 17301504
122 },
123 {
124 "name": "model.layers.0.self_attn.qkv_proj.q_weight",
125 "shape": [
126 5120,
127 384
128 ],
129 "dtype": "uint32",
130 "format": "f32-to-bf16",
131 "nbytes": 7864320,
132 "byteOffset": 17307648
133 },
134 {
135 "name": "model.layers.0.self_attn.qkv_proj.q_scale",
136 "shape": [
137 5120,
138 96
139 ],
140 "dtype": "float16",
141 "format": "f32-to-bf16",
142 "nbytes": 983040,
143 "byteOffset": 25171968
144 },
145 {
146 "name": "model.layers.0.self_attn.o_proj.q_weight",
147 "shape": [
148 3072,
149 384
150 ],
151 "dtype": "uint32",
152 "format": "f32-to-bf16",
153 "nbytes": 4718592,
154 "byteOffset": 26155008
155 },
156 {
157 "name": "model.layers.0.self_attn.o_proj.q_scale",
158 "shape": [
159 3072,
160 96
161 ],
162 "dtype": "float16",
163 "format": "f32-to-bf16",
164 "nbytes": 589824,
165 "byteOffset": 30873600
166 },
167 {
168 "name": "model.layers.1.input_layernorm.weight",
169 "shape": [
170 3072
171 ],
172 "dtype": "float16",
173 "format": "f32-to-bf16",
174 "nbytes": 6144,
175 "byteOffset": 31463424
176 }
177 ],
178 "md5sum": "5f147be52f0ed6e2845853b0cd46024e"
179 },
180 {
181 "dataPath": "params_shard_4.bin",
182 "format": "raw-shard",
183 "nbytes": 25165824,
184 "records": [
185 {
186 "name": "model.layers.1.mlp.gate_up_proj.q_weight",
187 "shape": [
188 16384,
189 384
190 ],
191 "dtype": "uint32",
192 "format": "f32-to-bf16",
193 "nbytes": 25165824,
194 "byteOffset": 0
195 }
196 ],
197 "md5sum": "99a3682f5c8efd7387b8c72c7b35ea3e"
198 },
199 {
200 "dataPath": "params_shard_5.bin",
201 "format": "raw-shard",
202 "nbytes": 31469568,
203 "records": [
204 {
205 "name": "model.layers.1.mlp.down_proj.q_weight",
206 "shape": [
207 3072,
208 1024
209 ],
210 "dtype": "uint32",
211 "format": "f32-to-bf16",
212 "nbytes": 12582912,
213 "byteOffset": 0
214 },
215 {
216 "name": "model.layers.1.mlp.down_proj.q_scale",
217 "shape": [
218 3072,
219 256
220 ],
221 "dtype": "float16",
222 "format": "f32-to-bf16",
223 "nbytes": 1572864,
224 "byteOffset": 12582912
225 },
226 {
227 "name": "model.layers.1.mlp.gate_up_proj.q_scale",
228 "shape": [
229 16384,
230 96
231 ],
232 "dtype": "float16",
233 "format": "f32-to-bf16",
234 "nbytes": 3145728,
235 "byteOffset": 14155776
236 },
237 {
238 "name": "model.layers.1.post_attention_layernorm.weight",
239 "shape": [
240 3072
241 ],
242 "dtype": "float16",
243 "format": "f32-to-bf16",
244 "nbytes": 6144,
245 "byteOffset": 17301504
246 },
247 {
248 "name": "model.layers.1.self_attn.qkv_proj.q_weight",
249 "shape": [
250 5120,
251 384
252 ],
253 "dtype": "uint32",
254 "format": "f32-to-bf16",
255 "nbytes": 7864320,
256 "byteOffset": 17307648
257 },
258 {
259 "name": "model.layers.1.self_attn.qkv_proj.q_scale",
260 "shape": [
261 5120,
262 96
263 ],
264 "dtype": "float16",
265 "format": "f32-to-bf16",
266 "nbytes": 983040,
267 "byteOffset": 25171968
268 },
269 {
270 "name": "model.layers.1.self_attn.o_proj.q_weight",
271 "shape": [
272 3072,
273 384
274 ],
275 "dtype": "uint32",
276 "format": "f32-to-bf16",
277 "nbytes": 4718592,
278 "byteOffset": 26155008
279 },
280 {
281 "name": "model.layers.1.self_attn.o_proj.q_scale",
282 "shape": [
283 3072,
284 96
285 ],
286 "dtype": "float16",
287 "format": "f32-to-bf16",
288 "nbytes": 589824,
289 "byteOffset": 30873600
290 },
291 {
292 "name": "model.layers.10.input_layernorm.weight",
293 "shape": [
294 3072
295 ],
296 "dtype": "float16",
297 "format": "f32-to-bf16",
298 "nbytes": 6144,
299 "byteOffset": 31463424
300 }
301 ],
302 "md5sum": "665d7503a547d2ea988dac1e9fd1712b"
303 },
304 {
305 "dataPath": "params_shard_6.bin",
306 "format": "raw-shard",
307 "nbytes": 25165824,
308 "records": [
309 {
310 "name": "model.layers.10.mlp.gate_up_proj.q_weight",
311 "shape": [
312 16384,
313 384
314 ],
315 "dtype": "uint32",
316 "format": "f32-to-bf16",
317 "nbytes": 25165824,
318 "byteOffset": 0
319 }
320 ],
321 "md5sum": "d44a0ea9a4480b44cea97205a38660fc"
322 },
323 {
324 "dataPath": "params_shard_7.bin",
325 "format": "raw-shard",
326 "nbytes": 31469568,
327 "records": [
328 {
329 "name": "model.layers.10.mlp.down_proj.q_weight",
330 "shape": [
331 3072,
332 1024
333 ],
334 "dtype": "uint32",
335 "format": "f32-to-bf16",
336 "nbytes": 12582912,
337 "byteOffset": 0
338 },
339 {
340 "name": "model.layers.10.mlp.down_proj.q_scale",
341 "shape": [
342 3072,
343 256
344 ],
345 "dtype": "float16",
346 "format": "f32-to-bf16",
347 "nbytes": 1572864,
348 "byteOffset": 12582912
349 },
350 {
351 "name": "model.layers.10.mlp.gate_up_proj.q_scale",
352 "shape": [
353 16384,
354 96
355 ],
356 "dtype": "float16",
357 "format": "f32-to-bf16",
358 "nbytes": 3145728,
359 "byteOffset": 14155776
360 },
361 {
362 "name": "model.layers.10.post_attention_layernorm.weight",
363 "shape": [
364 3072
365 ],
366 "dtype": "float16",
367 "format": "f32-to-bf16",
368 "nbytes": 6144,
369 "byteOffset": 17301504
370 },
371 {
372 "name": "model.layers.10.self_attn.qkv_proj.q_weight",
373 "shape": [
374 5120,
375 384
376 ],
377 "dtype": "uint32",
378 "format": "f32-to-bf16",
379 "nbytes": 7864320,
380 "byteOffset": 17307648
381 },
382 {
383 "name": "model.layers.10.self_attn.qkv_proj.q_scale",
384 "shape": [
385 5120,
386 96
387 ],
388 "dtype": "float16",
389 "format": "f32-to-bf16",
390 "nbytes": 983040,
391 "byteOffset": 25171968
392 },
393 {
394 "name": "model.layers.10.self_attn.o_proj.q_weight",
395 "shape": [
396 3072,
397 384
398 ],
399 "dtype": "uint32",
400 "format": "f32-to-bf16",
401 "nbytes": 4718592,
402 "byteOffset": 26155008
403 },
404 {
405 "name": "model.layers.10.self_attn.o_proj.q_scale",
406 "shape": [
407 3072,
408 96
409 ],
410 "dtype": "float16",
411 "format": "f32-to-bf16",
412 "nbytes": 589824,
413 "byteOffset": 30873600
414 },
415 {
416 "name": "model.layers.11.input_layernorm.weight",
417 "shape": [
418 3072
419 ],
420 "dtype": "float16",
421 "format": "f32-to-bf16",
422 "nbytes": 6144,
423 "byteOffset": 31463424
424 }
425 ],
426 "md5sum": "91c5d212e13abd4259e11bff37a383da"
427 },
428 {
429 "dataPath": "params_shard_8.bin",
430 "format": "raw-shard",
431 "nbytes": 25165824,
432 "records": [
433 {
434 "name": "model.layers.11.mlp.gate_up_proj.q_weight",
435 "shape": [
436 16384,
437 384
438 ],
439 "dtype": "uint32",
440 "format": "f32-to-bf16",
441 "nbytes": 25165824,
442 "byteOffset": 0
443 }
444 ],
445 "md5sum": "058e0056c8e2e0636e57e2a317fc2d06"
446 },
447 {
448 "dataPath": "params_shard_9.bin",
449 "format": "raw-shard",
450 "nbytes": 31469568,
451 "records": [
452 {
453 "name": "model.layers.11.mlp.down_proj.q_weight",
454 "shape": [
455 3072,
456 1024
457 ],
458 "dtype": "uint32",
459 "format": "f32-to-bf16",
460 "nbytes": 12582912,
461 "byteOffset": 0
462 },
463 {
464 "name": "model.layers.11.mlp.down_proj.q_scale",
465 "shape": [
466 3072,
467 256
468 ],
469 "dtype": "float16",
470 "format": "f32-to-bf16",
471 "nbytes": 1572864,
472 "byteOffset": 12582912
473 },
474 {
475 "name": "model.layers.11.mlp.gate_up_proj.q_scale",
476 "shape": [
477 16384,
478 96
479 ],
480 "dtype": "float16",
481 "format": "f32-to-bf16",
482 "nbytes": 3145728,
483 "byteOffset": 14155776
484 },
485 {
486 "name": "model.layers.11.post_attention_layernorm.weight",
487 "shape": [
488 3072
489 ],
490 "dtype": "float16",
491 "format": "f32-to-bf16",
492 "nbytes": 6144,
493 "byteOffset": 17301504
494 },
495 {
496 "name": "model.layers.11.self_attn.qkv_proj.q_weight",
497 "shape": [
498 5120,
499 384
500 ],
501 "dtype": "uint32",
502 "format": "f32-to-bf16",
503 "nbytes": 7864320,
504 "byteOffset": 17307648
505 },
506 {
507 "name": "model.layers.11.self_attn.qkv_proj.q_scale",
508 "shape": [
509 5120,
510 96
511 ],
512 "dtype": "float16",
513 "format": "f32-to-bf16",
514 "nbytes": 983040,
515 "byteOffset": 25171968
516 },
517 {
518 "name": "model.layers.11.self_attn.o_proj.q_weight",
519 "shape": [
520 3072,
521 384
522 ],
523 "dtype": "uint32",
524 "format": "f32-to-bf16",
525 "nbytes": 4718592,
526 "byteOffset": 26155008
527 },
528 {
529 "name": "model.layers.11.self_attn.o_proj.q_scale",
530 "shape": [
531 3072,
532 96
533 ],
534 "dtype": "float16",
535 "format": "f32-to-bf16",
536 "nbytes": 589824,
537 "byteOffset": 30873600
538 },
539 {
540 "name": "model.layers.12.input_layernorm.weight",
541 "shape": [
542 3072
543 ],
544 "dtype": "float16",
545 "format": "f32-to-bf16",
546 "nbytes": 6144,
547 "byteOffset": 31463424
548 }
549 ],
550 "md5sum": "d7b05671ee9cf1c884397e257b4a9631"
551 },
552 {
553 "dataPath": "params_shard_10.bin",
554 "format": "raw-shard",
555 "nbytes": 25165824,
556 "records": [
557 {
558 "name": "model.layers.12.mlp.gate_up_proj.q_weight",
559 "shape": [
560 16384,
561 384
562 ],
563 "dtype": "uint32",
564 "format": "f32-to-bf16",
565 "nbytes": 25165824,
566 "byteOffset": 0
567 }
568 ],
569 "md5sum": "ac5e029ffbeae72822aa260076fc542b"
570 },
571 {
572 "dataPath": "params_shard_11.bin",
573 "format": "raw-shard",
574 "nbytes": 31469568,
575 "records": [
576 {
577 "name": "model.layers.12.mlp.down_proj.q_weight",
578 "shape": [
579 3072,
580 1024
581 ],
582 "dtype": "uint32",
583 "format": "f32-to-bf16",
584 "nbytes": 12582912,
585 "byteOffset": 0
586 },
587 {
588 "name": "model.layers.12.mlp.down_proj.q_scale",
589 "shape": [
590 3072,
591 256
592 ],
593 "dtype": "float16",
594 "format": "f32-to-bf16",
595 "nbytes": 1572864,
596 "byteOffset": 12582912
597 },
598 {
599 "name": "model.layers.12.mlp.gate_up_proj.q_scale",
600 "shape": [
601 16384,
602 96
603 ],
604 "dtype": "float16",
605 "format": "f32-to-bf16",
606 "nbytes": 3145728,
607 "byteOffset": 14155776
608 },
609 {
610 "name": "model.layers.12.post_attention_layernorm.weight",
611 "shape": [
612 3072
613 ],
614 "dtype": "float16",
615 "format": "f32-to-bf16",
616 "nbytes": 6144,
617 "byteOffset": 17301504
618 },
619 {
620 "name": "model.layers.12.self_attn.qkv_proj.q_weight",
621 "shape": [
622 5120,
623 384
624 ],
625 "dtype": "uint32",
626 "format": "f32-to-bf16",
627 "nbytes": 7864320,
628 "byteOffset": 17307648
629 },
630 {
631 "name": "model.layers.12.self_attn.qkv_proj.q_scale",
632 "shape": [
633 5120,
634 96
635 ],
636 "dtype": "float16",
637 "format": "f32-to-bf16",
638 "nbytes": 983040,
639 "byteOffset": 25171968
640 },
641 {
642 "name": "model.layers.12.self_attn.o_proj.q_weight",
643 "shape": [
644 3072,
645 384
646 ],
647 "dtype": "uint32",
648 "format": "f32-to-bf16",
649 "nbytes": 4718592,
650 "byteOffset": 26155008
651 },
652 {
653 "name": "model.layers.12.self_attn.o_proj.q_scale",
654 "shape": [
655 3072,
656 96
657 ],
658 "dtype": "float16",
659 "format": "f32-to-bf16",
660 "nbytes": 589824,
661 "byteOffset": 30873600
662 },
663 {
664 "name": "model.layers.13.input_layernorm.weight",
665 "shape": [
666 3072
667 ],
668 "dtype": "float16",
669 "format": "f32-to-bf16",
670 "nbytes": 6144,
671 "byteOffset": 31463424
672 }
673 ],
674 "md5sum": "13e57265a45505946a6ba86234551443"
675 },
676 {
677 "dataPath": "params_shard_12.bin",
678 "format": "raw-shard",
679 "nbytes": 25165824,
680 "records": [
681 {
682 "name": "model.layers.13.mlp.gate_up_proj.q_weight",
683 "shape": [
684 16384,
685 384
686 ],
687 "dtype": "uint32",
688 "format": "f32-to-bf16",
689 "nbytes": 25165824,
690 "byteOffset": 0
691 }
692 ],
693 "md5sum": "1153c8f795d9e1cc1cb630511506dc51"
694 },
695 {
696 "dataPath": "params_shard_13.bin",
697 "format": "raw-shard",
698 "nbytes": 31469568,
699 "records": [
700 {
701 "name": "model.layers.13.mlp.down_proj.q_weight",
702 "shape": [
703 3072,
704 1024
705 ],
706 "dtype": "uint32",
707 "format": "f32-to-bf16",
708 "nbytes": 12582912,
709 "byteOffset": 0
710 },
711 {
712 "name": "model.layers.13.mlp.down_proj.q_scale",
713 "shape": [
714 3072,
715 256
716 ],
717 "dtype": "float16",
718 "format": "f32-to-bf16",
719 "nbytes": 1572864,
720 "byteOffset": 12582912
721 },
722 {
723 "name": "model.layers.13.mlp.gate_up_proj.q_scale",
724 "shape": [
725 16384,
726 96
727 ],
728 "dtype": "float16",
729 "format": "f32-to-bf16",
730 "nbytes": 3145728,
731 "byteOffset": 14155776
732 },
733 {
734 "name": "model.layers.13.post_attention_layernorm.weight",
735 "shape": [
736 3072
737 ],
738 "dtype": "float16",
739 "format": "f32-to-bf16",
740 "nbytes": 6144,
741 "byteOffset": 17301504
742 },
743 {
744 "name": "model.layers.13.self_attn.qkv_proj.q_weight",
745 "shape": [
746 5120,
747 384
748 ],
749 "dtype": "uint32",
750 "format": "f32-to-bf16",
751 "nbytes": 7864320,
752 "byteOffset": 17307648
753 },
754 {
755 "name": "model.layers.13.self_attn.qkv_proj.q_scale",
756 "shape": [
757 5120,
758 96
759 ],
760 "dtype": "float16",
761 "format": "f32-to-bf16",
762 "nbytes": 983040,
763 "byteOffset": 25171968
764 },
765 {
766 "name": "model.layers.13.self_attn.o_proj.q_weight",
767 "shape": [
768 3072,
769 384
770 ],
771 "dtype": "uint32",
772 "format": "f32-to-bf16",
773 "nbytes": 4718592,
774 "byteOffset": 26155008
775 },
776 {
777 "name": "model.layers.13.self_attn.o_proj.q_scale",
778 "shape": [
779 3072,
780 96
781 ],
782 "dtype": "float16",
783 "format": "f32-to-bf16",
784 "nbytes": 589824,
785 "byteOffset": 30873600
786 },
787 {
788 "name": "model.layers.14.input_layernorm.weight",
789 "shape": [
790 3072
791 ],
792 "dtype": "float16",
793 "format": "f32-to-bf16",
794 "nbytes": 6144,
795 "byteOffset": 31463424
796 }
797 ],
798 "md5sum": "2f35d98cdb6fd4a17ab699dca6a2cb81"
799 },
800 {
801 "dataPath": "params_shard_14.bin",
802 "format": "raw-shard",
803 "nbytes": 25165824,
804 "records": [
805 {
806 "name": "model.layers.14.mlp.gate_up_proj.q_weight",
807 "shape": [
808 16384,
809 384
810 ],
811 "dtype": "uint32",
812 "format": "f32-to-bf16",
813 "nbytes": 25165824,
814 "byteOffset": 0
815 }
816 ],
817 "md5sum": "07a00ce02924f1f366b12b3c938beb49"
818 },
819 {
820 "dataPath": "params_shard_15.bin",
821 "format": "raw-shard",
822 "nbytes": 31469568,
823 "records": [
824 {
825 "name": "model.layers.14.mlp.down_proj.q_weight",
826 "shape": [
827 3072,
828 1024
829 ],
830 "dtype": "uint32",
831 "format": "f32-to-bf16",
832 "nbytes": 12582912,
833 "byteOffset": 0
834 },
835 {
836 "name": "model.layers.14.mlp.down_proj.q_scale",
837 "shape": [
838 3072,
839 256
840 ],
841 "dtype": "float16",
842 "format": "f32-to-bf16",
843 "nbytes": 1572864,
844 "byteOffset": 12582912
845 },
846 {
847 "name": "model.layers.14.mlp.gate_up_proj.q_scale",
848 "shape": [
849 16384,
850 96
851 ],
852 "dtype": "float16",
853 "format": "f32-to-bf16",
854 "nbytes": 3145728,
855 "byteOffset": 14155776
856 },
857 {
858 "name": "model.layers.14.post_attention_layernorm.weight",
859 "shape": [
860 3072
861 ],
862 "dtype": "float16",
863 "format": "f32-to-bf16",
864 "nbytes": 6144,
865 "byteOffset": 17301504
866 },
867 {
868 "name": "model.layers.14.self_attn.qkv_proj.q_weight",
869 "shape": [
870 5120,
871 384
872 ],
873 "dtype": "uint32",
874 "format": "f32-to-bf16",
875 "nbytes": 7864320,
876 "byteOffset": 17307648
877 },
878 {
879 "name": "model.layers.14.self_attn.qkv_proj.q_scale",
880 "shape": [
881 5120,
882 96
883 ],
884 "dtype": "float16",
885 "format": "f32-to-bf16",
886 "nbytes": 983040,
887 "byteOffset": 25171968
888 },
889 {
890 "name": "model.layers.14.self_attn.o_proj.q_weight",
891 "shape": [
892 3072,
893 384
894 ],
895 "dtype": "uint32",
896 "format": "f32-to-bf16",
897 "nbytes": 4718592,
898 "byteOffset": 26155008
899 },
900 {
901 "name": "model.layers.14.self_attn.o_proj.q_scale",
902 "shape": [
903 3072,
904 96
905 ],
906 "dtype": "float16",
907 "format": "f32-to-bf16",
908 "nbytes": 589824,
909 "byteOffset": 30873600
910 },
911 {
912 "name": "model.layers.15.input_layernorm.weight",
913 "shape": [
914 3072
915 ],
916 "dtype": "float16",
917 "format": "f32-to-bf16",
918 "nbytes": 6144,
919 "byteOffset": 31463424
920 }
921 ],
922 "md5sum": "6d757b7e75dfe652c54ae211cc3d017f"
923 },
924 {
925 "dataPath": "params_shard_16.bin",
926 "format": "raw-shard",
927 "nbytes": 25165824,
928 "records": [
929 {
930 "name": "model.layers.15.mlp.gate_up_proj.q_weight",
931 "shape": [
932 16384,
933 384
934 ],
935 "dtype": "uint32",
936 "format": "f32-to-bf16",
937 "nbytes": 25165824,
938 "byteOffset": 0
939 }
940 ],
941 "md5sum": "0357997408bd3a66424218689aed2fd8"
942 },
943 {
944 "dataPath": "params_shard_17.bin",
945 "format": "raw-shard",
946 "nbytes": 31469568,
947 "records": [
948 {
949 "name": "model.layers.15.mlp.down_proj.q_weight",
950 "shape": [
951 3072,
952 1024
953 ],
954 "dtype": "uint32",
955 "format": "f32-to-bf16",
956 "nbytes": 12582912,
957 "byteOffset": 0
958 },
959 {
960 "name": "model.layers.15.mlp.down_proj.q_scale",
961 "shape": [
962 3072,
963 256
964 ],
965 "dtype": "float16",
966 "format": "f32-to-bf16",
967 "nbytes": 1572864,
968 "byteOffset": 12582912
969 },
970 {
971 "name": "model.layers.15.mlp.gate_up_proj.q_scale",
972 "shape": [
973 16384,
974 96
975 ],
976 "dtype": "float16",
977 "format": "f32-to-bf16",
978 "nbytes": 3145728,
979 "byteOffset": 14155776
980 },
981 {
982 "name": "model.layers.15.post_attention_layernorm.weight",
983 "shape": [
984 3072
985 ],
986 "dtype": "float16",
987 "format": "f32-to-bf16",
988 "nbytes": 6144,
989 "byteOffset": 17301504
990 },
991 {
992 "name": "model.layers.15.self_attn.qkv_proj.q_weight",
993 "shape": [
994 5120,
995 384
996 ],
997 "dtype": "uint32",
998 "format": "f32-to-bf16",
999 "nbytes": 7864320,
1000 "byteOffset": 17307648
1001 },
1002 {
1003 "name": "model.layers.15.self_attn.qkv_proj.q_scale",
1004 "shape": [
1005 5120,
1006 96
1007 ],
1008 "dtype": "float16",
1009 "format": "f32-to-bf16",
1010 "nbytes": 983040,
1011 "byteOffset": 25171968
1012 },
1013 {
1014 "name": "model.layers.15.self_attn.o_proj.q_weight",
1015 "shape": [
1016 3072,
1017 384
1018 ],
1019 "dtype": "uint32",
1020 "format": "f32-to-bf16",
1021 "nbytes": 4718592,
1022 "byteOffset": 26155008
1023 },
1024 {
1025 "name": "model.layers.15.self_attn.o_proj.q_scale",
1026 "shape": [
1027 3072,
1028 96
1029 ],
1030 "dtype": "float16",
1031 "format": "f32-to-bf16",
1032 "nbytes": 589824,
1033 "byteOffset": 30873600
1034 },
1035 {
1036 "name": "model.layers.16.input_layernorm.weight",
1037 "shape": [
1038 3072
1039 ],
1040 "dtype": "float16",
1041 "format": "f32-to-bf16",
1042 "nbytes": 6144,
1043 "byteOffset": 31463424
1044 }
1045 ],
1046 "md5sum": "1322c5b8234d7495b1086b6ac7a49a1e"
1047 },
1048 {
1049 "dataPath": "params_shard_18.bin",
1050 "format": "raw-shard",
1051 "nbytes": 25165824,
1052 "records": [
1053 {
1054 "name": "model.layers.16.mlp.gate_up_proj.q_weight",
1055 "shape": [
1056 16384,
1057 384
1058 ],
1059 "dtype": "uint32",
1060 "format": "f32-to-bf16",
1061 "nbytes": 25165824,
1062 "byteOffset": 0
1063 }
1064 ],
1065 "md5sum": "ccffe55782edde78791221919d4d711a"
1066 },
1067 {
1068 "dataPath": "params_shard_19.bin",
1069 "format": "raw-shard",
1070 "nbytes": 31469568,
1071 "records": [
1072 {
1073 "name": "model.layers.16.mlp.down_proj.q_weight",
1074 "shape": [
1075 3072,
1076 1024
1077 ],
1078 "dtype": "uint32",
1079 "format": "f32-to-bf16",
1080 "nbytes": 12582912,
1081 "byteOffset": 0
1082 },
1083 {
1084 "name": "model.layers.16.mlp.down_proj.q_scale",
1085 "shape": [
1086 3072,
1087 256
1088 ],
1089 "dtype": "float16",
1090 "format": "f32-to-bf16",
1091 "nbytes": 1572864,
1092 "byteOffset": 12582912
1093 },
1094 {
1095 "name": "model.layers.16.mlp.gate_up_proj.q_scale",
1096 "shape": [
1097 16384,
1098 96
1099 ],
1100 "dtype": "float16",
1101 "format": "f32-to-bf16",
1102 "nbytes": 3145728,
1103 "byteOffset": 14155776
1104 },
1105 {
1106 "name": "model.layers.16.post_attention_layernorm.weight",
1107 "shape": [
1108 3072
1109 ],
1110 "dtype": "float16",
1111 "format": "f32-to-bf16",
1112 "nbytes": 6144,
1113 "byteOffset": 17301504
1114 },
1115 {
1116 "name": "model.layers.16.self_attn.qkv_proj.q_weight",
1117 "shape": [
1118 5120,
1119 384
1120 ],
1121 "dtype": "uint32",
1122 "format": "f32-to-bf16",
1123 "nbytes": 7864320,
1124 "byteOffset": 17307648
1125 },
1126 {
1127 "name": "model.layers.16.self_attn.qkv_proj.q_scale",
1128 "shape": [
1129 5120,
1130 96
1131 ],
1132 "dtype": "float16",
1133 "format": "f32-to-bf16",
1134 "nbytes": 983040,
1135 "byteOffset": 25171968
1136 },
1137 {
1138 "name": "model.layers.16.self_attn.o_proj.q_weight",
1139 "shape": [
1140 3072,
1141 384
1142 ],
1143 "dtype": "uint32",
1144 "format": "f32-to-bf16",
1145 "nbytes": 4718592,
1146 "byteOffset": 26155008
1147 },
1148 {
1149 "name": "model.layers.16.self_attn.o_proj.q_scale",
1150 "shape": [
1151 3072,
1152 96
1153 ],
1154 "dtype": "float16",
1155 "format": "f32-to-bf16",
1156 "nbytes": 589824,
1157 "byteOffset": 30873600
1158 },
1159 {
1160 "name": "model.layers.17.input_layernorm.weight",
1161 "shape": [
1162 3072
1163 ],
1164 "dtype": "float16",
1165 "format": "f32-to-bf16",
1166 "nbytes": 6144,
1167 "byteOffset": 31463424
1168 }
1169 ],
1170 "md5sum": "d9c1a26269b15404d6e64d2ff7af221f"
1171 },
1172 {
1173 "dataPath": "params_shard_20.bin",
1174 "format": "raw-shard",
1175 "nbytes": 25165824,
1176 "records": [
1177 {
1178 "name": "model.layers.17.mlp.gate_up_proj.q_weight",
1179 "shape": [
1180 16384,
1181 384
1182 ],
1183 "dtype": "uint32",
1184 "format": "f32-to-bf16",
1185 "nbytes": 25165824,
1186 "byteOffset": 0
1187 }
1188 ],
1189 "md5sum": "c17d4178da4421580698cf1059a408f7"
1190 },
1191 {
1192 "dataPath": "params_shard_21.bin",
1193 "format": "raw-shard",
1194 "nbytes": 31469568,
1195 "records": [
1196 {
1197 "name": "model.layers.17.mlp.down_proj.q_weight",
1198 "shape": [
1199 3072,
1200 1024