FallnAI/flux-quantized
110
1{
2 "time_text_embed.timestep_embedder.linear_1": {
3 "weights": "qint8",
4 "activations": "none"
5 },
6 "time_text_embed.timestep_embedder.linear_2": {
7 "weights": "qint8",
8 "activations": "none"
9 },
10 "time_text_embed.text_embedder.linear_1": {
11 "weights": "qint8",
12 "activations": "none"
13 },
14 "time_text_embed.text_embedder.linear_2": {
15 "weights": "qint8",
16 "activations": "none"
17 },
18 "context_embedder": {
19 "weights": "qint8",
20 "activations": "none"
21 },
22 "x_embedder": {
23 "weights": "qint8",
24 "activations": "none"
25 },
26 "transformer_blocks.0.norm1.linear": {
27 "weights": "qint8",
28 "activations": "none"
29 },
30 "transformer_blocks.0.norm1_context.linear": {
31 "weights": "qint8",
32 "activations": "none"
33 },
34 "transformer_blocks.0.attn.to_q": {
35 "weights": "qint8",
36 "activations": "none"
37 },
38 "transformer_blocks.0.attn.to_k": {
39 "weights": "qint8",
40 "activations": "none"
41 },
42 "transformer_blocks.0.attn.to_v": {
43 "weights": "qint8",
44 "activations": "none"
45 },
46 "transformer_blocks.0.attn.add_k_proj": {
47 "weights": "qint8",
48 "activations": "none"
49 },
50 "transformer_blocks.0.attn.add_v_proj": {
51 "weights": "qint8",
52 "activations": "none"
53 },
54 "transformer_blocks.0.attn.add_q_proj": {
55 "weights": "qint8",
56 "activations": "none"
57 },
58 "transformer_blocks.0.attn.to_out.0": {
59 "weights": "qint8",
60 "activations": "none"
61 },
62 "transformer_blocks.0.attn.to_add_out": {
63 "weights": "qint8",
64 "activations": "none"
65 },
66 "transformer_blocks.0.ff.net.0.proj": {
67 "weights": "qint8",
68 "activations": "none"
69 },
70 "transformer_blocks.0.ff.net.2": {
71 "weights": "qint8",
72 "activations": "none"
73 },
74 "transformer_blocks.0.ff_context.net.0.proj": {
75 "weights": "qint8",
76 "activations": "none"
77 },
78 "transformer_blocks.0.ff_context.net.2": {
79 "weights": "qint8",
80 "activations": "none"
81 },
82 "transformer_blocks.1.norm1.linear": {
83 "weights": "qint8",
84 "activations": "none"
85 },
86 "transformer_blocks.1.norm1_context.linear": {
87 "weights": "qint8",
88 "activations": "none"
89 },
90 "transformer_blocks.1.attn.to_q": {
91 "weights": "qint8",
92 "activations": "none"
93 },
94 "transformer_blocks.1.attn.to_k": {
95 "weights": "qint8",
96 "activations": "none"
97 },
98 "transformer_blocks.1.attn.to_v": {
99 "weights": "qint8",
100 "activations": "none"
101 },
102 "transformer_blocks.1.attn.add_k_proj": {
103 "weights": "qint8",
104 "activations": "none"
105 },
106 "transformer_blocks.1.attn.add_v_proj": {
107 "weights": "qint8",
108 "activations": "none"
109 },
110 "transformer_blocks.1.attn.add_q_proj": {
111 "weights": "qint8",
112 "activations": "none"
113 },
114 "transformer_blocks.1.attn.to_out.0": {
115 "weights": "qint8",
116 "activations": "none"
117 },
118 "transformer_blocks.1.attn.to_add_out": {
119 "weights": "qint8",
120 "activations": "none"
121 },
122 "transformer_blocks.1.ff.net.0.proj": {
123 "weights": "qint8",
124 "activations": "none"
125 },
126 "transformer_blocks.1.ff.net.2": {
127 "weights": "qint8",
128 "activations": "none"
129 },
130 "transformer_blocks.1.ff_context.net.0.proj": {
131 "weights": "qint8",
132 "activations": "none"
133 },
134 "transformer_blocks.1.ff_context.net.2": {
135 "weights": "qint8",
136 "activations": "none"
137 },
138 "transformer_blocks.2.norm1.linear": {
139 "weights": "qint8",
140 "activations": "none"
141 },
142 "transformer_blocks.2.norm1_context.linear": {
143 "weights": "qint8",
144 "activations": "none"
145 },
146 "transformer_blocks.2.attn.to_q": {
147 "weights": "qint8",
148 "activations": "none"
149 },
150 "transformer_blocks.2.attn.to_k": {
151 "weights": "qint8",
152 "activations": "none"
153 },
154 "transformer_blocks.2.attn.to_v": {
155 "weights": "qint8",
156 "activations": "none"
157 },
158 "transformer_blocks.2.attn.add_k_proj": {
159 "weights": "qint8",
160 "activations": "none"
161 },
162 "transformer_blocks.2.attn.add_v_proj": {
163 "weights": "qint8",
164 "activations": "none"
165 },
166 "transformer_blocks.2.attn.add_q_proj": {
167 "weights": "qint8",
168 "activations": "none"
169 },
170 "transformer_blocks.2.attn.to_out.0": {
171 "weights": "qint8",
172 "activations": "none"
173 },
174 "transformer_blocks.2.attn.to_add_out": {
175 "weights": "qint8",
176 "activations": "none"
177 },
178 "transformer_blocks.2.ff.net.0.proj": {
179 "weights": "qint8",
180 "activations": "none"
181 },
182 "transformer_blocks.2.ff.net.2": {
183 "weights": "qint8",
184 "activations": "none"
185 },
186 "transformer_blocks.2.ff_context.net.0.proj": {
187 "weights": "qint8",
188 "activations": "none"
189 },
190 "transformer_blocks.2.ff_context.net.2": {
191 "weights": "qint8",
192 "activations": "none"
193 },
194 "transformer_blocks.3.norm1.linear": {
195 "weights": "qint8",
196 "activations": "none"
197 },
198 "transformer_blocks.3.norm1_context.linear": {
199 "weights": "qint8",
200 "activations": "none"
201 },
202 "transformer_blocks.3.attn.to_q": {
203 "weights": "qint8",
204 "activations": "none"
205 },
206 "transformer_blocks.3.attn.to_k": {
207 "weights": "qint8",
208 "activations": "none"
209 },
210 "transformer_blocks.3.attn.to_v": {
211 "weights": "qint8",
212 "activations": "none"
213 },
214 "transformer_blocks.3.attn.add_k_proj": {
215 "weights": "qint8",
216 "activations": "none"
217 },
218 "transformer_blocks.3.attn.add_v_proj": {
219 "weights": "qint8",
220 "activations": "none"
221 },
222 "transformer_blocks.3.attn.add_q_proj": {
223 "weights": "qint8",
224 "activations": "none"
225 },
226 "transformer_blocks.3.attn.to_out.0": {
227 "weights": "qint8",
228 "activations": "none"
229 },
230 "transformer_blocks.3.attn.to_add_out": {
231 "weights": "qint8",
232 "activations": "none"
233 },
234 "transformer_blocks.3.ff.net.0.proj": {
235 "weights": "qint8",
236 "activations": "none"
237 },
238 "transformer_blocks.3.ff.net.2": {
239 "weights": "qint8",
240 "activations": "none"
241 },
242 "transformer_blocks.3.ff_context.net.0.proj": {
243 "weights": "qint8",
244 "activations": "none"
245 },
246 "transformer_blocks.3.ff_context.net.2": {
247 "weights": "qint8",
248 "activations": "none"
249 },
250 "transformer_blocks.4.norm1.linear": {
251 "weights": "qint8",
252 "activations": "none"
253 },
254 "transformer_blocks.4.norm1_context.linear": {
255 "weights": "qint8",
256 "activations": "none"
257 },
258 "transformer_blocks.4.attn.to_q": {
259 "weights": "qint8",
260 "activations": "none"
261 },
262 "transformer_blocks.4.attn.to_k": {
263 "weights": "qint8",
264 "activations": "none"
265 },
266 "transformer_blocks.4.attn.to_v": {
267 "weights": "qint8",
268 "activations": "none"
269 },
270 "transformer_blocks.4.attn.add_k_proj": {
271 "weights": "qint8",
272 "activations": "none"
273 },
274 "transformer_blocks.4.attn.add_v_proj": {
275 "weights": "qint8",
276 "activations": "none"
277 },
278 "transformer_blocks.4.attn.add_q_proj": {
279 "weights": "qint8",
280 "activations": "none"
281 },
282 "transformer_blocks.4.attn.to_out.0": {
283 "weights": "qint8",
284 "activations": "none"
285 },
286 "transformer_blocks.4.attn.to_add_out": {
287 "weights": "qint8",
288 "activations": "none"
289 },
290 "transformer_blocks.4.ff.net.0.proj": {
291 "weights": "qint8",
292 "activations": "none"
293 },
294 "transformer_blocks.4.ff.net.2": {
295 "weights": "qint8",
296 "activations": "none"
297 },
298 "transformer_blocks.4.ff_context.net.0.proj": {
299 "weights": "qint8",
300 "activations": "none"
301 },
302 "transformer_blocks.4.ff_context.net.2": {
303 "weights": "qint8",
304 "activations": "none"
305 },
306 "transformer_blocks.5.norm1.linear": {
307 "weights": "qint8",
308 "activations": "none"
309 },
310 "transformer_blocks.5.norm1_context.linear": {
311 "weights": "qint8",
312 "activations": "none"
313 },
314 "transformer_blocks.5.attn.to_q": {
315 "weights": "qint8",
316 "activations": "none"
317 },
318 "transformer_blocks.5.attn.to_k": {
319 "weights": "qint8",
320 "activations": "none"
321 },
322 "transformer_blocks.5.attn.to_v": {
323 "weights": "qint8",
324 "activations": "none"
325 },
326 "transformer_blocks.5.attn.add_k_proj": {
327 "weights": "qint8",
328 "activations": "none"
329 },
330 "transformer_blocks.5.attn.add_v_proj": {
331 "weights": "qint8",
332 "activations": "none"
333 },
334 "transformer_blocks.5.attn.add_q_proj": {
335 "weights": "qint8",
336 "activations": "none"
337 },
338 "transformer_blocks.5.attn.to_out.0": {
339 "weights": "qint8",
340 "activations": "none"
341 },
342 "transformer_blocks.5.attn.to_add_out": {
343 "weights": "qint8",
344 "activations": "none"
345 },
346 "transformer_blocks.5.ff.net.0.proj": {
347 "weights": "qint8",
348 "activations": "none"
349 },
350 "transformer_blocks.5.ff.net.2": {
351 "weights": "qint8",
352 "activations": "none"
353 },
354 "transformer_blocks.5.ff_context.net.0.proj": {
355 "weights": "qint8",
356 "activations": "none"
357 },
358 "transformer_blocks.5.ff_context.net.2": {
359 "weights": "qint8",
360 "activations": "none"
361 },
362 "transformer_blocks.6.norm1.linear": {
363 "weights": "qint8",
364 "activations": "none"
365 },
366 "transformer_blocks.6.norm1_context.linear": {
367 "weights": "qint8",
368 "activations": "none"
369 },
370 "transformer_blocks.6.attn.to_q": {
371 "weights": "qint8",
372 "activations": "none"
373 },
374 "transformer_blocks.6.attn.to_k": {
375 "weights": "qint8",
376 "activations": "none"
377 },
378 "transformer_blocks.6.attn.to_v": {
379 "weights": "qint8",
380 "activations": "none"
381 },
382 "transformer_blocks.6.attn.add_k_proj": {
383 "weights": "qint8",
384 "activations": "none"
385 },
386 "transformer_blocks.6.attn.add_v_proj": {
387 "weights": "qint8",
388 "activations": "none"
389 },
390 "transformer_blocks.6.attn.add_q_proj": {
391 "weights": "qint8",
392 "activations": "none"
393 },
394 "transformer_blocks.6.attn.to_out.0": {
395 "weights": "qint8",
396 "activations": "none"
397 },
398 "transformer_blocks.6.attn.to_add_out": {
399 "weights": "qint8",
400 "activations": "none"
401 },
402 "transformer_blocks.6.ff.net.0.proj": {
403 "weights": "qint8",
404 "activations": "none"
405 },
406 "transformer_blocks.6.ff.net.2": {
407 "weights": "qint8",
408 "activations": "none"
409 },
410 "transformer_blocks.6.ff_context.net.0.proj": {
411 "weights": "qint8",
412 "activations": "none"
413 },
414 "transformer_blocks.6.ff_context.net.2": {
415 "weights": "qint8",
416 "activations": "none"
417 },
418 "transformer_blocks.7.norm1.linear": {
419 "weights": "qint8",
420 "activations": "none"
421 },
422 "transformer_blocks.7.norm1_context.linear": {
423 "weights": "qint8",
424 "activations": "none"
425 },
426 "transformer_blocks.7.attn.to_q": {
427 "weights": "qint8",
428 "activations": "none"
429 },
430 "transformer_blocks.7.attn.to_k": {
431 "weights": "qint8",
432 "activations": "none"
433 },
434 "transformer_blocks.7.attn.to_v": {
435 "weights": "qint8",
436 "activations": "none"
437 },
438 "transformer_blocks.7.attn.add_k_proj": {
439 "weights": "qint8",
440 "activations": "none"
441 },
442 "transformer_blocks.7.attn.add_v_proj": {
443 "weights": "qint8",
444 "activations": "none"
445 },
446 "transformer_blocks.7.attn.add_q_proj": {
447 "weights": "qint8",
448 "activations": "none"
449 },
450 "transformer_blocks.7.attn.to_out.0": {
451 "weights": "qint8",
452 "activations": "none"
453 },
454 "transformer_blocks.7.attn.to_add_out": {
455 "weights": "qint8",
456 "activations": "none"
457 },
458 "transformer_blocks.7.ff.net.0.proj": {
459 "weights": "qint8",
460 "activations": "none"
461 },
462 "transformer_blocks.7.ff.net.2": {
463 "weights": "qint8",
464 "activations": "none"
465 },
466 "transformer_blocks.7.ff_context.net.0.proj": {
467 "weights": "qint8",
468 "activations": "none"
469 },
470 "transformer_blocks.7.ff_context.net.2": {
471 "weights": "qint8",
472 "activations": "none"
473 },
474 "transformer_blocks.8.norm1.linear": {
475 "weights": "qint8",
476 "activations": "none"
477 },
478 "transformer_blocks.8.norm1_context.linear": {
479 "weights": "qint8",
480 "activations": "none"
481 },
482 "transformer_blocks.8.attn.to_q": {
483 "weights": "qint8",
484 "activations": "none"
485 },
486 "transformer_blocks.8.attn.to_k": {
487 "weights": "qint8",
488 "activations": "none"
489 },
490 "transformer_blocks.8.attn.to_v": {
491 "weights": "qint8",
492 "activations": "none"
493 },
494 "transformer_blocks.8.attn.add_k_proj": {
495 "weights": "qint8",
496 "activations": "none"
497 },
498 "transformer_blocks.8.attn.add_v_proj": {
499 "weights": "qint8",
500 "activations": "none"
501 },
502 "transformer_blocks.8.attn.add_q_proj": {
503 "weights": "qint8",
504 "activations": "none"
505 },
506 "transformer_blocks.8.attn.to_out.0": {
507 "weights": "qint8",
508 "activations": "none"
509 },
510 "transformer_blocks.8.attn.to_add_out": {
511 "weights": "qint8",
512 "activations": "none"
513 },
514 "transformer_blocks.8.ff.net.0.proj": {
515 "weights": "qint8",
516 "activations": "none"
517 },
518 "transformer_blocks.8.ff.net.2": {
519 "weights": "qint8",
520 "activations": "none"
521 },
522 "transformer_blocks.8.ff_context.net.0.proj": {
523 "weights": "qint8",
524 "activations": "none"
525 },
526 "transformer_blocks.8.ff_context.net.2": {
527 "weights": "qint8",
528 "activations": "none"
529 },
530 "transformer_blocks.9.norm1.linear": {
531 "weights": "qint8",
532 "activations": "none"
533 },
534 "transformer_blocks.9.norm1_context.linear": {
535 "weights": "qint8",
536 "activations": "none"
537 },
538 "transformer_blocks.9.attn.to_q": {
539 "weights": "qint8",
540 "activations": "none"
541 },
542 "transformer_blocks.9.attn.to_k": {
543 "weights": "qint8",
544 "activations": "none"
545 },
546 "transformer_blocks.9.attn.to_v": {
547 "weights": "qint8",
548 "activations": "none"
549 },
550 "transformer_blocks.9.attn.add_k_proj": {
551 "weights": "qint8",
552 "activations": "none"
553 },
554 "transformer_blocks.9.attn.add_v_proj": {
555 "weights": "qint8",
556 "activations": "none"
557 },
558 "transformer_blocks.9.attn.add_q_proj": {
559 "weights": "qint8",
560 "activations": "none"
561 },
562 "transformer_blocks.9.attn.to_out.0": {
563 "weights": "qint8",
564 "activations": "none"
565 },
566 "transformer_blocks.9.attn.to_add_out": {
567 "weights": "qint8",
568 "activations": "none"
569 },
570 "transformer_blocks.9.ff.net.0.proj": {
571 "weights": "qint8",
572 "activations": "none"
573 },
574 "transformer_blocks.9.ff.net.2": {
575 "weights": "qint8",
576 "activations": "none"
577 },
578 "transformer_blocks.9.ff_context.net.0.proj": {
579 "weights": "qint8",
580 "activations": "none"
581 },
582 "transformer_blocks.9.ff_context.net.2": {
583 "weights": "qint8",
584 "activations": "none"
585 },
586 "transformer_blocks.10.norm1.linear": {
587 "weights": "qint8",
588 "activations": "none"
589 },
590 "transformer_blocks.10.norm1_context.linear": {
591 "weights": "qint8",
592 "activations": "none"
593 },
594 "transformer_blocks.10.attn.to_q": {
595 "weights": "qint8",
596 "activations": "none"
597 },
598 "transformer_blocks.10.attn.to_k": {
599 "weights": "qint8",
600 "activations": "none"
601 },
602 "transformer_blocks.10.attn.to_v": {
603 "weights": "qint8",
604 "activations": "none"
605 },
606 "transformer_blocks.10.attn.add_k_proj": {
607 "weights": "qint8",
608 "activations": "none"
609 },
610 "transformer_blocks.10.attn.add_v_proj": {
611 "weights": "qint8",
612 "activations": "none"
613 },
614 "transformer_blocks.10.attn.add_q_proj": {
615 "weights": "qint8",
616 "activations": "none"
617 },
618 "transformer_blocks.10.attn.to_out.0": {
619 "weights": "qint8",
620 "activations": "none"
621 },
622 "transformer_blocks.10.attn.to_add_out": {
623 "weights": "qint8",
624 "activations": "none"
625 },
626 "transformer_blocks.10.ff.net.0.proj": {
627 "weights": "qint8",
628 "activations": "none"
629 },
630 "transformer_blocks.10.ff.net.2": {
631 "weights": "qint8",
632 "activations": "none"
633 },
634 "transformer_blocks.10.ff_context.net.0.proj": {
635 "weights": "qint8",
636 "activations": "none"
637 },
638 "transformer_blocks.10.ff_context.net.2": {
639 "weights": "qint8",
640 "activations": "none"
641 },
642 "transformer_blocks.11.norm1.linear": {
643 "weights": "qint8",
644 "activations": "none"
645 },
646 "transformer_blocks.11.norm1_context.linear": {
647 "weights": "qint8",
648 "activations": "none"
649 },
650 "transformer_blocks.11.attn.to_q": {
651 "weights": "qint8",
652 "activations": "none"
653 },
654 "transformer_blocks.11.attn.to_k": {
655 "weights": "qint8",
656 "activations": "none"
657 },
658 "transformer_blocks.11.attn.to_v": {
659 "weights": "qint8",
660 "activations": "none"
661 },
662 "transformer_blocks.11.attn.add_k_proj": {
663 "weights": "qint8",
664 "activations": "none"
665 },
666 "transformer_blocks.11.attn.add_v_proj": {
667 "weights": "qint8",
668 "activations": "none"
669 },
670 "transformer_blocks.11.attn.add_q_proj": {
671 "weights": "qint8",
672 "activations": "none"
673 },
674 "transformer_blocks.11.attn.to_out.0": {
675 "weights": "qint8",
676 "activations": "none"
677 },
678 "transformer_blocks.11.attn.to_add_out": {
679 "weights": "qint8",
680 "activations": "none"
681 },
682 "transformer_blocks.11.ff.net.0.proj": {
683 "weights": "qint8",
684 "activations": "none"
685 },
686 "transformer_blocks.11.ff.net.2": {
687 "weights": "qint8",
688 "activations": "none"
689 },
690 "transformer_blocks.11.ff_context.net.0.proj": {
691 "weights": "qint8",
692 "activations": "none"
693 },
694 "transformer_blocks.11.ff_context.net.2": {
695 "weights": "qint8",
696 "activations": "none"
697 },
698 "transformer_blocks.12.norm1.linear": {
699 "weights": "qint8",
700 "activations": "none"
701 },
702 "transformer_blocks.12.norm1_context.linear": {
703 "weights": "qint8",
704 "activations": "none"
705 },
706 "transformer_blocks.12.attn.to_q": {
707 "weights": "qint8",
708 "activations": "none"
709 },
710 "transformer_blocks.12.attn.to_k": {
711 "weights": "qint8",
712 "activations": "none"
713 },
714 "transformer_blocks.12.attn.to_v": {
715 "weights": "qint8",
716 "activations": "none"
717 },
718 "transformer_blocks.12.attn.add_k_proj": {
719 "weights": "qint8",
720 "activations": "none"
721 },
722 "transformer_blocks.12.attn.add_v_proj": {
723 "weights": "qint8",
724 "activations": "none"
725 },
726 "transformer_blocks.12.attn.add_q_proj": {
727 "weights": "qint8",
728 "activations": "none"
729 },
730 "transformer_blocks.12.attn.to_out.0": {
731 "weights": "qint8",
732 "activations": "none"
733 },
734 "transformer_blocks.12.attn.to_add_out": {
735 "weights": "qint8",
736 "activations": "none"
737 },
738 "transformer_blocks.12.ff.net.0.proj": {
739 "weights": "qint8",
740 "activations": "none"
741 },
742 "transformer_blocks.12.ff.net.2": {
743 "weights": "qint8",
744 "activations": "none"
745 },
746 "transformer_blocks.12.ff_context.net.0.proj": {
747 "weights": "qint8",
748 "activations": "none"
749 },
750 "transformer_blocks.12.ff_context.net.2": {
751 "weights": "qint8",
752 "activations": "none"
753 },
754 "transformer_blocks.13.norm1.linear": {
755 "weights": "qint8",
756 "activations": "none"
757 },
758 "transformer_blocks.13.norm1_context.linear": {
759 "weights": "qint8",
760 "activations": "none"
761 },
762 "transformer_blocks.13.attn.to_q": {
763 "weights": "qint8",
764 "activations": "none"
765 },
766 "transformer_blocks.13.attn.to_k": {
767 "weights": "qint8",
768 "activations": "none"
769 },
770 "transformer_blocks.13.attn.to_v": {
771 "weights": "qint8",
772 "activations": "none"
773 },
774 "transformer_blocks.13.attn.add_k_proj": {
775 "weights": "qint8",
776 "activations": "none"
777 },
778 "transformer_blocks.13.attn.add_v_proj": {
779 "weights": "qint8",
780 "activations": "none"
781 },
782 "transformer_blocks.13.attn.add_q_proj": {
783 "weights": "qint8",
784 "activations": "none"
785 },
786 "transformer_blocks.13.attn.to_out.0": {
787 "weights": "qint8",
788 "activations": "none"
789 },
790 "transformer_blocks.13.attn.to_add_out": {
791 "weights": "qint8",
792 "activations": "none"
793 },
794 "transformer_blocks.13.ff.net.0.proj": {
795 "weights": "qint8",
796 "activations": "none"
797 },
798 "transformer_blocks.13.ff.net.2": {
799 "weights": "qint8",
800 "activations": "none"
801 },
802 "transformer_blocks.13.ff_context.net.0.proj": {
803 "weights": "qint8",
804 "activations": "none"
805 },
806 "transformer_blocks.13.ff_context.net.2": {
807 "weights": "qint8",
808 "activations": "none"
809 },
810 "transformer_blocks.14.norm1.linear": {
811 "weights": "qint8",
812 "activations": "none"
813 },
814 "transformer_blocks.14.norm1_context.linear": {
815 "weights": "qint8",
816 "activations": "none"
817 },
818 "transformer_blocks.14.attn.to_q": {
819 "weights": "qint8",
820 "activations": "none"
821 },
822 "transformer_blocks.14.attn.to_k": {
823 "weights": "qint8",
824 "activations": "none"
825 },
826 "transformer_blocks.14.attn.to_v": {
827 "weights": "qint8",
828 "activations": "none"
829 },
830 "transformer_blocks.14.attn.add_k_proj": {
831 "weights": "qint8",
832 "activations": "none"
833 },
834 "transformer_blocks.14.attn.add_v_proj": {
835 "weights": "qint8",
836 "activations": "none"
837 },
838 "transformer_blocks.14.attn.add_q_proj": {
839 "weights": "qint8",
840 "activations": "none"
841 },
842 "transformer_blocks.14.attn.to_out.0": {
843 "weights": "qint8",
844 "activations": "none"
845 },
846 "transformer_blocks.14.attn.to_add_out": {
847 "weights": "qint8",
848 "activations": "none"
849 },
850 "transformer_blocks.14.ff.net.0.proj": {
851 "weights": "qint8",
852 "activations": "none"
853 },
854 "transformer_blocks.14.ff.net.2": {
855 "weights": "qint8",
856 "activations": "none"
857 },
858 "transformer_blocks.14.ff_context.net.0.proj": {
859 "weights": "qint8",
860 "activations": "none"
861 },
862 "transformer_blocks.14.ff_context.net.2": {
863 "weights": "qint8",
864 "activations": "none"
865 },
866 "transformer_blocks.15.norm1.linear": {
867 "weights": "qint8",
868 "activations": "none"
869 },
870 "transformer_blocks.15.norm1_context.linear": {
871 "weights": "qint8",
872 "activations": "none"
873 },
874 "transformer_blocks.15.attn.to_q": {
875 "weights": "qint8",
876 "activations": "none"
877 },
878 "transformer_blocks.15.attn.to_k": {
879 "weights": "qint8",
880 "activations": "none"
881 },
882 "transformer_blocks.15.attn.to_v": {
883 "weights": "qint8",
884 "activations": "none"
885 },
886 "transformer_blocks.15.attn.add_k_proj": {
887 "weights": "qint8",
888 "activations": "none"
889 },
890 "transformer_blocks.15.attn.add_v_proj": {
891 "weights": "qint8",
892 "activations": "none"
893 },
894 "transformer_blocks.15.attn.add_q_proj": {
895 "weights": "qint8",
896 "activations": "none"
897 },
898 "transformer_blocks.15.attn.to_out.0": {
899 "weights": "qint8",
900 "activations": "none"
901 },
902 "transformer_blocks.15.attn.to_add_out": {
903 "weights": "qint8",
904 "activations": "none"
905 },
906 "transformer_blocks.15.ff.net.0.proj": {
907 "weights": "qint8",
908 "activations": "none"
909 },
910 "transformer_blocks.15.ff.net.2": {
911 "weights": "qint8",
912 "activations": "none"
913 },
914 "transformer_blocks.15.ff_context.net.0.proj": {
915 "weights": "qint8",
916 "activations": "none"
917 },
918 "transformer_blocks.15.ff_context.net.2": {
919 "weights": "qint8",
920 "activations": "none"
921 },
922 "transformer_blocks.16.norm1.linear": {
923 "weights": "qint8",
924 "activations": "none"
925 },
926 "transformer_blocks.16.norm1_context.linear": {
927 "weights": "qint8",
928 "activations": "none"
929 },
930 "transformer_blocks.16.attn.to_q": {
931 "weights": "qint8",
932 "activations": "none"
933 },
934 "transformer_blocks.16.attn.to_k": {
935 "weights": "qint8",
936 "activations": "none"
937 },
938 "transformer_blocks.16.attn.to_v": {
939 "weights": "qint8",
940 "activations": "none"
941 },
942 "transformer_blocks.16.attn.add_k_proj": {
943 "weights": "qint8",
944 "activations": "none"
945 },
946 "transformer_blocks.16.attn.add_v_proj": {
947 "weights": "qint8",
948 "activations": "none"
949 },
950 "transformer_blocks.16.attn.add_q_proj": {
951 "weights": "qint8",
952 "activations": "none"
953 },
954 "transformer_blocks.16.attn.to_out.0": {
955 "weights": "qint8",
956 "activations": "none"
957 },
958 "transformer_blocks.16.attn.to_add_out": {
959 "weights": "qint8",
960 "activations": "none"
961 },
962 "transformer_blocks.16.ff.net.0.proj": {
963 "weights": "qint8",
964 "activations": "none"
965 },
966 "transformer_blocks.16.ff.net.2": {
967 "weights": "qint8",
968 "activations": "none"
969 },
970 "transformer_blocks.16.ff_context.net.0.proj": {
971 "weights": "qint8",
972 "activations": "none"
973 },
974 "transformer_blocks.16.ff_context.net.2": {
975 "weights": "qint8",
976 "activations": "none"
977 },
978 "transformer_blocks.17.norm1.linear": {
979 "weights": "qint8",
980 "activations": "none"
981 },
982 "transformer_blocks.17.norm1_context.linear": {
983 "weights": "qint8",
984 "activations": "none"
985 },
986 "transformer_blocks.17.attn.to_q": {
987 "weights": "qint8",
988 "activations": "none"
989 },
990 "transformer_blocks.17.attn.to_k": {
991 "weights": "qint8",
992 "activations": "none"
993 },
994 "transformer_blocks.17.attn.to_v": {
995 "weights": "qint8",
996 "activations": "none"
997 },
998 "transformer_blocks.17.attn.add_k_proj": {
999 "weights": "qint8",
1000 "activations": "none"
1001 },
1002 "transformer_blocks.17.attn.add_v_proj": {
1003 "weights": "qint8",
1004 "activations": "none"
1005 },
1006 "transformer_blocks.17.attn.add_q_proj": {
1007 "weights": "qint8",
1008 "activations": "none"
1009 },
1010 "transformer_blocks.17.attn.to_out.0": {
1011 "weights": "qint8",
1012 "activations": "none"
1013 },
1014 "transformer_blocks.17.attn.to_add_out": {
1015 "weights": "qint8",
1016 "activations": "none"
1017 },
1018 "transformer_blocks.17.ff.net.0.proj": {
1019 "weights": "qint8",
1020 "activations": "none"
1021 },
1022 "transformer_blocks.17.ff.net.2": {
1023 "weights": "qint8",
1024 "activations": "none"
1025 },
1026 "transformer_blocks.17.ff_context.net.0.proj": {
1027 "weights": "qint8",
1028 "activations": "none"
1029 },
1030 "transformer_blocks.17.ff_context.net.2": {
1031 "weights": "qint8",
1032 "activations": "none"
1033 },
1034 "transformer_blocks.18.norm1.linear": {
1035 "weights": "qint8",
1036 "activations": "none"
1037 },
1038 "transformer_blocks.18.norm1_context.linear": {
1039 "weights": "qint8",
1040 "activations": "none"
1041 },
1042 "transformer_blocks.18.attn.to_q": {
1043 "weights": "qint8",
1044 "activations": "none"
1045 },
1046 "transformer_blocks.18.attn.to_k": {
1047 "weights": "qint8",
1048 "activations": "none"
1049 },
1050 "transformer_blocks.18.attn.to_v": {
1051 "weights": "qint8",
1052 "activations": "none"
1053 },
1054 "transformer_blocks.18.attn.add_k_proj": {
1055 "weights": "qint8",
1056 "activations": "none"
1057 },
1058 "transformer_blocks.18.attn.add_v_proj": {
1059 "weights": "qint8",
1060 "activations": "none"
1061 },
1062 "transformer_blocks.18.attn.add_q_proj": {
1063 "weights": "qint8",
1064 "activations": "none"
1065 },
1066 "transformer_blocks.18.attn.to_out.0": {
1067 "weights": "qint8",
1068 "activations": "none"
1069 },
1070 "transformer_blocks.18.attn.to_add_out": {
1071 "weights": "qint8",
1072 "activations": "none"
1073 },
1074 "transformer_blocks.18.ff.net.0.proj": {
1075 "weights": "qint8",
1076 "activations": "none"
1077 },
1078 "transformer_blocks.18.ff.net.2": {
1079 "weights": "qint8",
1080 "activations": "none"
1081 },
1082 "transformer_blocks.18.ff_context.net.0.proj": {
1083 "weights": "qint8",
1084 "activations": "none"
1085 },
1086 "transformer_blocks.18.ff_context.net.2": {
1087 "weights": "qint8",
1088 "activations": "none"
1089 },
1090 "single_transformer_blocks.0.norm.linear": {
1091 "weights": "qint8",
1092 "activations": "none"
1093 },
1094 "single_transformer_blocks.0.proj_mlp": {
1095 "weights": "qint8",
1096 "activations": "none"
1097 },
1098 "single_transformer_blocks.0.proj_out": {
1099 "weights": "qint8",
1100 "activations": "none"
1101 },
1102 "single_transformer_blocks.0.attn.to_q": {
1103 "weights": "qint8",
1104 "activations": "none"
1105 },
1106 "single_transformer_blocks.0.attn.to_k": {
1107 "weights": "qint8",
1108 "activations": "none"
1109 },
1110 "single_transformer_blocks.0.attn.to_v": {
1111 "weights": "qint8",
1112 "activations": "none"
1113 },
1114 "single_transformer_blocks.1.norm.linear": {
1115 "weights": "qint8",
1116 "activations": "none"
1117 },
1118 "single_transformer_blocks.1.proj_mlp": {
1119 "weights": "qint8",
1120 "activations": "none"
1121 },
1122 "single_transformer_blocks.1.proj_out": {
1123 "weights": "qint8",
1124 "activations": "none"
1125 },
1126 "single_transformer_blocks.1.attn.to_q": {
1127 "weights": "qint8",
1128 "activations": "none"
1129 },
1130 "single_transformer_blocks.1.attn.to_k": {
1131 "weights": "qint8",
1132 "activations": "none"
1133 },
1134 "single_transformer_blocks.1.attn.to_v": {
1135 "weights": "qint8",
1136 "activations": "none"
1137 },
1138 "single_transformer_blocks.2.norm.linear": {
1139 "weights": "qint8",
1140 "activations": "none"
1141 },
1142 "single_transformer_blocks.2.proj_mlp": {
1143 "weights": "qint8",
1144 "activations": "none"
1145 },
1146 "single_transformer_blocks.2.proj_out": {
1147 "weights": "qint8",
1148 "activations": "none"
1149 },
1150 "single_transformer_blocks.2.attn.to_q": {
1151 "weights": "qint8",
1152 "activations": "none"
1153 },
1154 "single_transformer_blocks.2.attn.to_k": {
1155 "weights": "qint8",
1156 "activations": "none"
1157 },
1158 "single_transformer_blocks.2.attn.to_v": {
1159 "weights": "qint8",
1160 "activations": "none"
1161 },
1162 "single_transformer_blocks.3.norm.linear": {
1163 "weights": "qint8",
1164 "activations": "none"
1165 },
1166 "single_transformer_blocks.3.proj_mlp": {
1167 "weights": "qint8",
1168 "activations": "none"
1169 },
1170 "single_transformer_blocks.3.proj_out": {
1171 "weights": "qint8",
1172 "activations": "none"
1173 },
1174 "single_transformer_blocks.3.attn.to_q": {
1175 "weights": "qint8",
1176 "activations": "none"
1177 },
1178 "single_transformer_blocks.3.attn.to_k": {
1179 "weights": "qint8",
1180 "activations": "none"
1181 },
1182 "single_transformer_blocks.3.attn.to_v": {
1183 "weights": "qint8",
1184 "activations": "none"
1185 },
1186 "single_transformer_blocks.4.norm.linear": {
1187 "weights": "qint8",
1188 "activations": "none"
1189 },
1190 "single_transformer_blocks.4.proj_mlp": {
1191 "weights": "qint8",
1192 "activations": "none"
1193 },
1194 "single_transformer_blocks.4.proj_out": {
1195 "weights": "qint8",
1196 "activations": "none"
1197 },
1198 "single_transformer_blocks.4.attn.to_q": {
1199 "weights": "qint8",
1200 "activations": "none"
