ComfyPilotComfyPilot
工作流 / 视频 / Wan VACE控制视频
Wan VACE控制视频

Wan VACE控制视频

Wan2.1 VACE Control Video

使用Wan VACE通过控制输入视频和参考图像生成视频。

视频 28 个节点 2 个输入 148 次使用 ComfyUI ≥ 0.3.31 ComfyUI 31 KB

下载工作流 JSON 查看官方来源

下载后把文件拖进 ComfyUI 窗口即可打开。

需要的权重

这个工作流引用了 7 个权重文件，其中 7 个能在本站模型库里找到。

Wan21_CausVid_14B_T2V_lora_rank32.safetensors 在模型库查看
Wan21_CausVid_bidirect2_T2V_1_3B_lora_rank32.safetensors 在模型库查看
umt5_xxl_fp16.safetensors 在模型库查看
umt5_xxl_fp8_e4m3fn_scaled.safetensors 在模型库查看
wan2.1_vace_1.3B_fp16.safetensors 在模型库查看
wan2.1_vace_14B_fp16.safetensors 在模型库查看
wan_2.1_vae.safetensors 在模型库查看

节点构成

MarkdownNote × 6
UNETLoader × 2
LoraLoader × 2
CLIPLoader × 2
CLIPTextEncode × 2
CreateVideo × 1
KSampler × 1
ModelSamplingSD3 × 1
VAELoader × 1
TrimVideoLatent × 1
VAEDecode × 1
SaveAnimatedWEBP × 1
WanVaceToVideo × 1
SaveVideo × 1
PreviewImage × 1
GetVideoComponents × 1

官方标注的模型

Wan2.1Wan

工作流 JSON

和上面「下载工作流 JSON」是同一份文件。

1{
2 "id": "96995b8f-85c5-47af-b3cf-7b6a24675694",
3 "revision": 0,
4 "last_node_id": 150,
5 "last_link_id": 226,
6 "nodes": [
7 {
8 "id": 68,
9 "type": "CreateVideo",
10 "pos": [
11 1920,
12 50
13 ],
14 "size": [
15 270,
16 78
17 ],
18 "flags": {
19 "collapsed": false
20 },
21 "order": 26,
22 "mode": 0,
23 "inputs": [
24 {
25 "name": "images",
26 "type": "IMAGE",
27 "link": 139
28 },
29 {
30 "name": "audio",
31 "shape": 7,
32 "type": "AUDIO",
33 "link": null
34 }
35 ],
36 "outputs": [
37 {
38 "name": "VIDEO",
39 "type": "VIDEO",
40 "links": [
41 185
42 ]
43 }
44 ],
45 "properties": {
46 "cnr_id": "comfy-core",
47 "ver": "0.3.34",
48 "Node name for S&R": "CreateVideo"
49 },
50 "widgets_values": [
51 16
52 ]
53 },
54 {
55 "id": 3,
56 "type": "KSampler",
57 "pos": [
58 980,
59 50
60 ],
61 "size": [
62 315,
63 262
64 ],
65 "flags": [],
66 "order": 22,
67 "mode": 0,
68 "inputs": [
69 {
70 "name": "model",
71 "type": "MODEL",
72 "link": 95
73 },
74 {
75 "name": "positive",
76 "type": "CONDITIONING",
77 "link": 98
78 },
79 {
80 "name": "negative",
81 "type": "CONDITIONING",
82 "link": 99
83 },
84 {
85 "name": "latent_image",
86 "type": "LATENT",
87 "link": 160
88 }
89 ],
90 "outputs": [
91 {
92 "name": "LATENT",
93 "type": "LATENT",
94 "slot_index": 0,
95 "links": [
96 116
97 ]
98 }
99 ],
100 "properties": {
101 "cnr_id": "comfy-core",
102 "ver": "0.3.34",
103 "Node name for S&R": "KSampler"
104 },
105 "widgets_values": [
106 654654950714624,
107 "randomize",
108 4,
109 1,
110 "uni_pc",
111 "simple",
112 1
113 ]
114 },
115 {
116 "id": 48,
117 "type": "ModelSamplingSD3",
118 "pos": [
119 730,
120 10
121 ],
122 "size": [
123 210,
124 58
125 ],
126 "flags": {
127 "collapsed": true
128 },
129 "order": 16,
130 "mode": 0,
131 "inputs": [
132 {
133 "name": "model",
134 "type": "MODEL",
135 "link": 192
136 }
137 ],
138 "outputs": [
139 {
140 "name": "MODEL",
141 "type": "MODEL",
142 "slot_index": 0,
143 "links": [
144 95
145 ]
146 }
147 ],
148 "properties": {
149 "cnr_id": "comfy-core",
150 "ver": "0.3.34",
151 "Node name for S&R": "ModelSamplingSD3"
152 },
153 "widgets_values": [
154 8.000000000000002
155 ]
156 },
157 {
158 "id": 106,
159 "type": "UNETLoader",
160 "pos": [
161 -260,
162 90
163 ],
164 "size": [
165 350,
166 82
167 ],
168 "flags": [],
169 "order": 0,
170 "mode": 0,
171 "inputs": [],
172 "outputs": [
173 {
174 "name": "MODEL",
175 "type": "MODEL",
176 "slot_index": 0,
177 "links": [
178 188
179 ]
180 }
181 ],
182 "properties": {
183 "cnr_id": "comfy-core",
184 "ver": "0.3.34",
185 "Node name for S&R": "UNETLoader",
186 "models": [
187 {
188 "name": "wan2.1_vace_14B_fp16.safetensors",
189 "url": "https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/diffusion_models/wan2.1_vace_14B_fp16.safetensors",
190 "directory": "diffusion_models"
191 }
192 ]
193 },
194 "widgets_values": [
195 "wan2.1_vace_14B_fp16.safetensors",
196 "default"
197 ]
198 },
199 {
200 "id": 105,
201 "type": "VAELoader",
202 "pos": [
203 -260,
204 530
205 ],
206 "size": [
207 350,
208 58
209 ],
210 "flags": [],
211 "order": 1,
212 "mode": 0,
213 "inputs": [],
214 "outputs": [
215 {
216 "name": "VAE",
217 "type": "VAE",
218 "slot_index": 0,
219 "links": [
220 183,
221 184
222 ]
223 }
224 ],
225 "properties": {
226 "cnr_id": "comfy-core",
227 "ver": "0.3.34",
228 "Node name for S&R": "VAELoader",
229 "models": [
230 {
231 "name": "wan_2.1_vae.safetensors",
232 "url": "https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/vae/wan_2.1_vae.safetensors",
233 "directory": "vae"
234 }
235 ]
236 },
237 "widgets_values": [
238 "wan_2.1_vae.safetensors"
239 ]
240 },
241 {
242 "id": 116,
243 "type": "MarkdownNote",
244 "pos": [
245 640,
246 370
247 ],
248 "size": [
249 310,
250 110
251 ],
252 "flags": [],
253 "order": 2,
254 "mode": 0,
255 "inputs": [],
256 "outputs": [],
257 "title": "About Video Size",
258 "properties": [],
259 "widgets_values": [
260 "| Model | 480P | 720P |\n| ------------------------------------------------------------ | ---- | ---- |\n| [VACE-1.3B](https://huggingface.co/Wan-AI/Wan2.1-VACE-1.3B) | ✅ | ❌ |\n| [VACE-14B](https://huggingface.co/Wan-AI/Wan2.1-VACE-14B) | ✅ | ✅ |"
261 ],
262 "color": "#222",
263 "bgcolor": "#000"
264 },
265 {
266 "id": 58,
267 "type": "TrimVideoLatent",
268 "pos": [
269 980,
270 570
271 ],
272 "size": [
273 276.5860290527344,
274 58
275 ],
276 "flags": {
277 "collapsed": true
278 },
279 "order": 23,
280 "mode": 0,
281 "inputs": [
282 {
283 "name": "samples",
284 "type": "LATENT",
285 "link": 116
286 },
287 {
288 "name": "trim_amount",
289 "type": "INT",
290 "widget": {
291 "name": "trim_amount"
292 },
293 "link": 115
294 }
295 ],
296 "outputs": [
297 {
298 "name": "LATENT",
299 "type": "LATENT",
300 "links": [
301 117
302 ]
303 }
304 ],
305 "properties": {
306 "cnr_id": "comfy-core",
307 "ver": "0.3.34",
308 "Node name for S&R": "TrimVideoLatent"
309 },
310 "widgets_values": [
311 0
312 ]
313 },
314 {
315 "id": 8,
316 "type": "VAEDecode",
317 "pos": [
318 1170,
319 570
320 ],
321 "size": [
322 210,
323 46
324 ],
325 "flags": {
326 "collapsed": true
327 },
328 "order": 24,
329 "mode": 0,
330 "inputs": [
331 {
332 "name": "samples",
333 "type": "LATENT",
334 "link": 117
335 },
336 {
337 "name": "vae",
338 "type": "VAE",
339 "link": 184
340 }
341 ],
342 "outputs": [
343 {
344 "name": "IMAGE",
345 "type": "IMAGE",
346 "slot_index": 0,
347 "links": [
348 130,
349 139
350 ]
351 }
352 ],
353 "properties": {
354 "cnr_id": "comfy-core",
355 "ver": "0.3.34",
356 "Node name for S&R": "VAEDecode"
357 },
358 "widgets_values": []
359 },
360 {
361 "id": 70,
362 "type": "SaveAnimatedWEBP",
363 "pos": [
364 2260,
365 70
366 ],
367 "size": [
368 670,
369 800
370 ],
371 "flags": [],
372 "order": 25,
373 "mode": 4,
374 "inputs": [
375 {
376 "name": "images",
377 "type": "IMAGE",
378 "link": 130
379 }
380 ],
381 "outputs": [],
382 "properties": {
383 "cnr_id": "comfy-core",
384 "ver": "0.3.34",
385 "Node name for S&R": "SaveAnimatedWEBP"
386 },
387 "widgets_values": [
388 "ComfyUI",
389 6,
390 true,
391 80,
392 "default"
393 ]
394 },
395 {
396 "id": 108,
397 "type": "UNETLoader",
398 "pos": [
399 -670,
400 40
401 ],
402 "size": [
403 350,
404 82
405 ],
406 "flags": [],
407 "order": 3,
408 "mode": 4,
409 "inputs": [],
410 "outputs": [
411 {
412 "name": "MODEL",
413 "type": "MODEL",
414 "slot_index": 0,
415 "links": [
416 178
417 ]
418 }
419 ],
420 "properties": {
421 "cnr_id": "comfy-core",
422 "ver": "0.3.34",
423 "Node name for S&R": "UNETLoader",
424 "models": [
425 {
426 "name": "wan2.1_vace_1.3B_fp16.safetensors",
427 "url": "https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/diffusion_models/wan2.1_vace_1.3B_fp16.safetensors",
428 "directory": "diffusion_models"
429 }
430 ]
431 },
432 "widgets_values": [
433 "wan2.1_vace_1.3B_fp16.safetensors",
434 "default"
435 ]
436 },
437 {
438 "id": 109,
439 "type": "LoraLoader",
440 "pos": [
441 -670,
442 160
443 ],
444 "size": [
445 350,
446 126
447 ],
448 "flags": [],
449 "order": 13,
450 "mode": 4,
451 "inputs": [
452 {
453 "name": "model",
454 "type": "MODEL",
455 "link": 178
456 },
457 {
458 "name": "clip",
459 "type": "CLIP",
460 "link": 179
461 }
462 ],
463 "outputs": [
464 {
465 "name": "MODEL",
466 "type": "MODEL",
467 "links": []
468 },
469 {
470 "name": "CLIP",
471 "type": "CLIP",
472 "links": []
473 }
474 ],
475 "properties": {
476 "cnr_id": "comfy-core",
477 "ver": "0.3.38",
478 "Node name for S&R": "LoraLoader",
479 "models": [
480 {
481 "name": "Wan21_CausVid_bidirect2_T2V_1_3B_lora_rank32.safetensors",
482 "url": "https://huggingface.co/Kijai/WanVideo_comfy/resolve/main/Wan21_CausVid_bidirect2_T2V_1_3B_lora_rank32.safetensors",
483 "directory": "loras"
484 }
485 ]
486 },
487 "widgets_values": [
488 "Wan21_CausVid_bidirect2_T2V_1_3B_lora_rank32.safetensors",
489 0.7000000000000002,
490 1
491 ]
492 },
493 {
494 "id": 111,
495 "type": "CLIPLoader",
496 "pos": [
497 -680,
498 350
499 ],
500 "size": [
501 370,
502 106
503 ],
504 "flags": [],
505 "order": 4,
506 "mode": 4,
507 "inputs": [],
508 "outputs": [
509 {
510 "name": "CLIP",
511 "type": "CLIP",
512 "slot_index": 0,
513 "links": []
514 }
515 ],
516 "properties": {
517 "cnr_id": "comfy-core",
518 "ver": "0.3.34",
519 "Node name for S&R": "CLIPLoader",
520 "models": [
521 {
522 "name": "umt5_xxl_fp8_e4m3fn_scaled.safetensors",
523 "url": "https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/text_encoders/umt5_xxl_fp8_e4m3fn_scaled.safetensors?download=true",
524 "directory": "text_encoders"
525 }
526 ]
527 },
528 "widgets_values": [
529 "umt5_xxl_fp8_e4m3fn_scaled.safetensors",
530 "wan",
531 "default"
532 ]
533 },
534 {
535 "id": 110,
536 "type": "CLIPLoader",
537 "pos": [
538 -260,
539 380
540 ],
541 "size": [
542 350,
543 106
544 ],
545 "flags": [],
546 "order": 5,
547 "mode": 0,
548 "inputs": [],
549 "outputs": [
550 {
551 "name": "CLIP",
552 "type": "CLIP",
553 "slot_index": 0,
554 "links": [
555 179,
556 189
557 ]
558 }
559 ],
560 "properties": {
561 "cnr_id": "comfy-core",
562 "ver": "0.3.34",
563 "Node name for S&R": "CLIPLoader",
564 "models": [
565 {
566 "name": "umt5_xxl_fp16.safetensors",
567 "url": "https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/text_encoders/umt5_xxl_fp16.safetensors",
568 "directory": "text_encoders"
569 }
570 ]
571 },
572 "widgets_values": [
573 "umt5_xxl_fp16.safetensors",
574 "wan",
575 "default"
576 ]
577 },
578 {
579 "id": 49,
580 "type": "WanVaceToVideo",
581 "pos": [
582 630,
583 60
584 ],
585 "size": [
586 315,
587 254
588 ],
589 "flags": [],
590 "order": 21,
591 "mode": 0,
592 "inputs": [
593 {
594 "name": "positive",
595 "type": "CONDITIONING",
596 "link": 96
597 },
598 {
599 "name": "negative",
600 "type": "CONDITIONING",
601 "link": 97
602 },
603 {
604 "name": "vae",
605 "type": "VAE",
606 "link": 183
607 },
608 {
609 "name": "control_video",
610 "shape": 7,
611 "type": "IMAGE",
612 "link": 226
613 },
614 {
615 "name": "control_masks",
616 "shape": 7,
617 "type": "MASK",
618 "link": null
619 },
620 {
621 "name": "reference_image",
622 "shape": 7,
623 "type": "IMAGE",
624 "link": 216
625 }
626 ],
627 "outputs": [
628 {
629 "name": "positive",
630 "type": "CONDITIONING",
631 "links": [
632 98
633 ]
634 },
635 {
636 "name": "negative",
637 "type": "CONDITIONING",
638 "links": [
639 99
640 ]
641 },
642 {
643 "name": "latent",
644 "type": "LATENT",
645 "links": [
646 160
647 ]
648 },
649 {
650 "name": "trim_latent",
651 "type": "INT",
652 "links": [
653 115
654 ]
655 }
656 ],
657 "properties": {
658 "cnr_id": "comfy-core",
659 "ver": "0.3.34",
660 "Node name for S&R": "WanVaceToVideo"
661 },
662 "widgets_values": [
663 720,
664 720,
665 81,
666 1,
667 1
668 ]
669 },
670 {
671 "id": 114,
672 "type": "SaveVideo",
673 "pos": [
674 1360,
675 170
676 ],
677 "size": [
678 830,
679 928
680 ],
681 "flags": [],
682 "order": 27,
683 "mode": 0,
684 "inputs": [
685 {
686 "name": "video",
687 "type": "VIDEO",
688 "link": 185
689 }
690 ],
691 "outputs": [],
692 "properties": {
693 "cnr_id": "comfy-core",
694 "ver": "0.3.38",
695 "Node name for S&R": "SaveVideo"
696 },
697 "widgets_values": [
698 "video/ComfyUI",
699 "auto",
700 "auto"
701 ]
702 },
703 {
704 "id": 146,
705 "type": "PreviewImage",
706 "pos": [
707 920,
708 670
709 ],
710 "size": [
711 409.20550537109375,
712 246
713 ],
714 "flags": [],
715 "order": 20,
716 "mode": 4,
717 "inputs": [
718 {
719 "name": "images",
720 "type": "IMAGE",
721 "link": 224
722 }
723 ],
724 "outputs": [],
725 "properties": {
726 "cnr_id": "comfy-core",
727 "ver": "0.3.34",
728 "Node name for S&R": "PreviewImage"
729 },
730 "widgets_values": []
731 },
732 {
733 "id": 144,
734 "type": "GetVideoComponents",
735 "pos": [
736 380,
737 680
738 ],
739 "size": [
740 185.17733764648438,
741 66
742 ],
743 "flags": [],
744 "order": 15,
745 "mode": 0,
746 "inputs": [
747 {
748 "name": "video",
749 "type": "VIDEO",
750 "link": 223
751 }
752 ],
753 "outputs": [
754 {
755 "name": "images",
756 "type": "IMAGE",
757 "links": [
758 225
759 ]
760 },
761 {
762 "name": "audio",
763 "type": "AUDIO",
764 "links": null
765 },
766 {
767 "name": "fps",
768 "type": "FLOAT",
769 "links": null
770 }
771 ],
772 "properties": {
773 "cnr_id": "comfy-core",
774 "ver": "0.3.38",
775 "Node name for S&R": "GetVideoComponents"
776 },
777 "widgets_values": []
778 },
779 {
780 "id": 147,
781 "type": "Canny",
782 "pos": [
783 630,
784 680
785 ],
786 "size": [
787 270,
788 82
789 ],
790 "flags": [],
791 "order": 19,
792 "mode": 4,
793 "inputs": [
794 {
795 "name": "image",
796 "type": "IMAGE",
797 "link": 225
798 }
799 ],
800 "outputs": [
801 {
802 "name": "IMAGE",
803 "type": "IMAGE",
804 "links": [
805 224,
806 226
807 ]
808 }
809 ],
810 "properties": {
811 "cnr_id": "comfy-core",
812 "ver": "0.3.34",
813 "Node name for S&R": "Canny"
814 },
815 "widgets_values": [
816 0.4,
817 0.8
818 ]
819 },
820 {
821 "id": 112,
822 "type": "MarkdownNote",
823 "pos": [
824 980,
825 350
826 ],
827 "size": [
828 320,
829 170
830 ],
831 "flags": [],
832 "order": 6,
833 "mode": 0,
834 "inputs": [],
835 "outputs": [],
836 "title": "KSampler Setting",
837 "properties": [],
838 "widgets_values": [
839 "## Default\n\n- steps:20\n- cfg:6.0\n\n## For CausVid LoRA\n\n- steps: 2-4\n- cfg: 1.0\n\n"
840 ],
841 "color": "#222",
842 "bgcolor": "#000"
843 },
844 {
845 "id": 107,
846 "type": "LoraLoader",
847 "pos": [
848 -260,
849 210
850 ],
851 "size": [
852 350,
853 126
854 ],
855 "flags": [],
856 "order": 14,
857 "mode": 0,
858 "inputs": [
859 {
860 "name": "model",
861 "type": "MODEL",
862 "link": 188
863 },
864 {
865 "name": "clip",
866 "type": "CLIP",
867 "link": 189
868 }
869 ],
870 "outputs": [
871 {
872 "name": "MODEL",
873 "type": "MODEL",
874 "links": [
875 192
876 ]
877 },
878 {
879 "name": "CLIP",
880 "type": "CLIP",
881 "links": [
882 190,
883 191
884 ]
885 }
886 ],
887 "properties": {
888 "cnr_id": "comfy-core",
889 "ver": "0.3.38",
890 "Node name for S&R": "LoraLoader",
891 "models": [
892 {
893 "name": "Wan21_CausVid_14B_T2V_lora_rank32.safetensors",
894 "url": "https://huggingface.co/Kijai/WanVideo_comfy/resolve/main/Wan21_CausVid_14B_T2V_lora_rank32.safetensors",
895 "directory": "loras"
896 }
897 ]
898 },
899 "widgets_values": [
900 "Wan21_CausVid_14B_T2V_lora_rank32.safetensors",
901 0.30000000000000004,
902 1
903 ]
904 },
905 {
906 "id": 6,
907 "type": "CLIPTextEncode",
908 "pos": [
909 150,
910 50
911 ],
912 "size": [
913 420,
914 290
915 ],
916 "flags": [],
917 "order": 17,
918 "mode": 0,
919 "inputs": [
920 {
921 "name": "clip",
922 "type": "CLIP",
923 "link": 190
924 }
925 ],
926 "outputs": [
927 {
928 "name": "CONDITIONING",
929 "type": "CONDITIONING",
930 "slot_index": 0,
931 "links": [
932 96
933 ]
934 }
935 ],
936 "title": "CLIP Text Encode (Positive Prompt)",
937 "properties": {
938 "cnr_id": "comfy-core",
939 "ver": "0.3.34",
940 "Node name for S&R": "CLIPTextEncode"
941 },
942 "widgets_values": [
943 "The girl is dancing in a sea of flowers, slowly moving her hands. There is a close - up shot of her upper body. The character is surrounded by other transparent glass flowers in the style of Nicoletta Ceccoli, creating a beautiful, surreal, and emotionally expressive movie scene with a white, transparent feel and a dreamy atmosphere. "
944 ],
945 "color": "#232",
946 "bgcolor": "#353"
947 },
948 {
949 "id": 7,
950 "type": "CLIPTextEncode",
951 "pos": [
952 150,
953 390
954 ],
955 "size": [
956 425.27801513671875,
957 180.6060791015625
958 ],
959 "flags": [],
960 "order": 18,
961 "mode": 0,
962 "inputs": [
963 {
964 "name": "clip",
965 "type": "CLIP",
966 "link": 191
967 }
968 ],
969 "outputs": [
970 {
971 "name": "CONDITIONING",
972 "type": "CONDITIONING",
973 "slot_index": 0,
974 "links": [
975 97
976 ]
977 }
978 ],
979 "title": "CLIP Text Encode (Negative Prompt)",
980 "properties": {
981 "cnr_id": "comfy-core",
982 "ver": "0.3.34",
983 "Node name for S&R": "CLIPTextEncode"
984 },
985 "widgets_values": [
986 "过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，三条腿，背景人很多，倒着走,"
987 ],
988 "color": "#322",
989 "bgcolor": "#533"
990 },
991 {
992 "id": 148,
993 "type": "MarkdownNote",
994 "pos": [
995 630,
996 810
997 ],
998 "size": [
999 270,
1000 110
1001 ],
1002 "flags": [],
1003 "order": 7,
1004 "mode": 0,
1005 "inputs": [],
1006 "outputs": [],
1007 "properties": [],
1008 "widgets_values": [
1009 "You can replace this part with another image preprocessor. \n\nRelated custom node: \n[comfyui_controlnet_aux](https://github.com/Fannovel16/comfyui_controlnet_aux)\n"
1010 ],
1011 "color": "#222",
1012 "bgcolor": "#000"
1013 },
1014 {
1015 "id": 113,
1016 "type": "MarkdownNote",
1017 "pos": [
1018 -680,
1019 500
1020 ],
1021 "size": [
1022 370,
1023 140
1024 ],
1025 "flags": [],
1026 "order": 8,
1027 "mode": 0,
1028 "inputs": [],
1029 "outputs": [],
1030 "title": "Note",
1031 "properties": [],
1032 "widgets_values": [
1033 "The generation quality of the 14B model is better, but it takes longer to generate. If you have already downloaded the model, you can choose to directly use the nodes above, or just modify the model loaded by the loader node.\n\nThe corresponding LoRA should match the Diffusion Model. For example, the LoRA corresponding to the 14B diffusion model is the 14B LoRA. \n"
1034 ],
1035 "color": "#222",
1036 "bgcolor": "#000"
1037 },
1038 {
1039 "id": 149,
1040 "type": "MarkdownNote",
1041 "pos": [
1042 -290,
1043 -150
1044 ],
1045 "size": [
1046 410,
1047 100
1048 ],
1049 "flags": [],
1050 "order": 9,
1051 "mode": 0,
1052 "inputs": [],
1053 "outputs": [],
1054 "title": "About CausVid LoRA",
1055 "properties": [],
1056 "widgets_values": [
1057 "We have added CausVid LoRA by default to achieve acceleration. However, in some cases, the video may shake and become blurry. You might need to test different LoRA intensities to get the best results, which should be between 0.3 and 0.7. If you don't need it, you can use the bypass mode to disable it, and then restore the settings of `KSampler` to the default ones.\n"
1058 ],
1059 "color": "#222",
1060 "bgcolor": "#000"
1061 },
1062 {
1063 "id": 145,
1064 "type": "LoadVideo",
1065 "pos": [
1066 60,
1067 680
1068 ],
1069 "size": [
1070 274.080078125,
1071 348.080078125
1072 ],
1073 "flags": [],
1074 "order": 10,
1075 "mode": 0,
1076 "inputs": [],
1077 "outputs": [
1078 {
1079 "name": "VIDEO",
1080 "type": "VIDEO",
1081 "links": [
1082 223
1083 ]
1084 }
1085 ],
1086 "properties": {
1087 "cnr_id": "comfy-core",
1088 "ver": "0.3.38",
1089 "Node name for S&R": "LoadVideo"
1090 },
1091 "widgets_values": [
1092 "video_wan_vace_14B_v2v_reference_image_control_video.mp4",
1093 "image"
1094 ]
1095 },
1096 {
1097 "id": 134,
1098 "type": "LoadImage",
1099 "pos": [
1100 -280,
1101 680
1102 ],
1103 "size": [
1104 280,
1105 314
1106 ],
1107 "flags": [],
1108 "order": 11,
1109 "mode": 0,
1110 "inputs": [],
1111 "outputs": [
1112 {
1113 "name": "IMAGE",
1114 "type": "IMAGE",
1115 "links": [
1116 216
1117 ]
1118 },
1119 {
1120 "name": "MASK",
1121 "type": "MASK",
1122 "links": null
1123 }
1124 ],
1125 "properties": {
1126 "cnr_id": "comfy-core",
1127 "ver": "0.3.34",
1128 "Node name for S&R": "LoadImage"
1129 },
1130 "widgets_values": [
1131 "video_wan_vace_14B_v2v_reference_image.jpg",
1132 "image"
1133 ]
1134 },
1135 {
1136 "id": 150,
1137 "type": "MarkdownNote",
1138 "pos": [
1139 -1160,
1140 0
1141 ],
1142 "size": [
1143 470,
1144 730
1145 ],
1146 "flags": [],
1147 "order": 12,
1148 "mode": 0,
1149 "inputs": [],
1150 "outputs": [],
1151 "properties": [],
1152 "widgets_values": [
1153 "[Tutorial](https://docs.comfy.org/tutorials/video/wan/vace)\n\n## 14B Support 480P 720P\n\n- [Hugging Face:Comfy-Org/Wan_2.1_ComfyUI_repackaged](https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged)\n\n**ModelScope mirrors** (alternative repo location)\n\n- [ModelScope:Comfy-Org/Wan_2.1_ComfyUI_repackaged](https://modelscope.cn/models/Comfy-Org/Wan_2.1_ComfyUI_repackaged)\n\n**Diffusion Model**\n- [wan2.1_vace_14B_fp16.safetensors](https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/diffusion_models/wan2.1_vace_14B_fp16.safetensors) (32.29 GB)\n\n**LoRA**\n- [Wan21_CausVid_14B_T2V_lora_rank32.safetensors](https://huggingface.co/Kijai/WanVideo_comfy/blob/main/Wan21_CausVid_14B_T2V_lora_rank32.safetensors)\n\nIt takes about 40 minutes to complete at 81 frames 720P resolution with the RTX 4090 . \nAfter using Wan21_CausVid_14B_T2V_lora_rank32.safetensors, it only takes about 4 minutes.\n\n## 1.3B Support 480P only\n\n**Diffusion Model**\n- [wan2.1_vace_1.3B_fp16.safetensors](https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/diffusion_models/wan2.1_vace_1.3B_fp16.safetensors) (4.01 GB)\n\n**LoRA**\n- [Wan21_CausVid_bidirect2_T2V_1_3B_lora_rank32.safetensors](https://huggingface.co/Kijai/WanVideo_comfy/blob/main/Wan21_CausVid_bidirect2_T2V_1_3B_lora_rank32.safetensors)\n \n\n## Other Models\n\n* You may already have these models if you use Wan workflow before.\n\n**VAE**\n- [wan_2.1_vae.safetensors](https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/vae/wan_2.1_vae.safetensors?download=true) (242.1 MB)\n\n**Text encoders** Chose one of following model\n- [umt5_xxl_fp16.safetensors](https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/text_encoders/umt5_xxl_fp16.safetensors?download=true) (10.59 GB)\n- [umt5_xxl_fp8_e4m3fn_scaled.safetensors](https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/text_encoders/umt5_xxl_fp8_e4m3fn_scaled.safetensors?download=true) (6.27 GB)\n\n> You can choose between fp16 of fp8; I used fp16 to match what kijai's wrapper is compatible with.\n\nFile save location\n\n```\nComfyUI/\n├── models/\n│ ├── diffusion_models/\n│ │ ├-── wan2.1_vace_14B_fp16.safetensors\n│ │ └─── wan2.1_vace_1.3B_fp16.safetensors \n│ ├── text_encoders/\n│ │ └─── umt5_xxl_fp8_e4m3fn_scaled.safetensors # or fp16\n│ ├── loras/\n│ │ ├── Wan21_CausVid_14B_T2V_lora_rank32.safetensors\n│ │ └── Wan21_CausVid_bidirect2_T2V_1_3B_lora_rank32.safetensors\n│ └── vae/\n│ └── wan_2.1_vae.safetensors\n```\n"
1154 ],
1155 "color": "#222",
1156 "bgcolor": "#000"
1157 }
1158 ],
1159 "links": [
1160 [
1161 95,
1162 48,
1163 0,
1164 3,
1165 0,
1166 "MODEL"
1167 ],
1168 [
1169 96,
1170 6,
1171 0,
1172 49,
1173 0,
1174 "CONDITIONING"
1175 ],
1176 [
1177 97,
1178 7,
1179 0,
1180 49,
1181 1,
1182 "CONDITIONING"
1183 ],
1184 [
1185 98,
1186 49,
1187 0,
1188 3,
1189 1,
1190 "CONDITIONING"
1191 ],
1192 [
1193 99,
1194 49,
1195 1,
1196 3,
1197 2,
1198 "CONDITIONING"
1199 ],
1200 [
1201 115,
1202 49,
1203 3,
1204 58,
1205 1,
1206 "INT"
1207 ],
1208 [
1209 116,
1210 3,
1211 0,
1212 58,
1213 0,
1214 "LATENT"
1215 ],
1216 [
1217 117,
1218 58,
1219 0,
1220 8,
1221 0,
1222 "LATENT"
1223 ],
1224 [
1225 130,
1226 8,
1227 0,
1228 70,
1229 0,
1230 "IMAGE"
1231 ],
1232 [
1233 139,
1234 8,
1235 0,
1236 68,
1237 0,
1238 "IMAGE"
1239 ],
1240 [
1241 160,
1242 49,
1243 2,
1244 3,
1245 3,
1246 "LATENT"
1247 ],
1248 [
1249 178,
1250 108,
1251 0,
1252 109,
1253 0,
1254 "MODEL"
1255 ],
1256 [
1257 179,
1258 110,
1259 0,
1260 109,
1261 1,
1262 "CLIP"
1263 ],
1264 [
1265 183,
1266 105,
1267 0,
1268 49,
1269 2,
1270 "VAE"
1271 ],
1272 [
1273 184,
1274 105,
1275 0,
1276 8,
1277 1,
1278 "VAE"
1279 ],
1280 [
1281 185,
1282 68,
1283 0,
1284 114,
1285 0,
1286 "VIDEO"
1287 ],
1288 [
1289 188,
1290 106,
1291 0,
1292 107,
1293 0,
1294 "MODEL"
1295 ],
1296 [
1297 189,
1298 110,
1299 0,
1300 107,
1301 1,
1302 "CLIP"
1303 ],
1304 [
1305 190,
1306 107,
1307 1,
1308 6,
1309 0,
1310 "CLIP"
1311 ],
1312 [
1313 191,
1314 107,
1315 1,
1316 7,
1317 0,
1318 "CLIP"
1319 ],
1320 [
1321 192,
1322 107,
1323 0,
1324 48,
1325 0,
1326 "MODEL"
1327 ],
1328 [
1329 216,
1330 134,
1331 0,
1332 49,
1333 5,
1334 "IMAGE"
1335 ],
1336 [
1337 223,
1338 145,
1339 0,
1340 144,
1341 0,
1342 "VIDEO"
1343 ],
1344 [
1345 224,
1346 147,
1347 0,
1348 146,
1349 0,
1350 "IMAGE"
1351 ],
1352 [
1353 225,
1354 144,
1355 0,
1356 147,
1357 0,
1358 "IMAGE"
1359 ],
1360 [
1361 226,
1362 147,
1363 0,
1364 49,
1365 3,
1366 "IMAGE"
1367 ]
1368 ],
1369 "groups": [
1370 {
1371 "id": 2,
1372 "title": "Prompt",
1373 "bounding": [
1374 140,
1375 -30,
1376 450,
1377 620
1378 ],
1379 "color": "#3f789e",
1380 "font_size": 24,
1381 "flags": []
1382 },
1383 {
1384 "id": 3,
1385 "title": "Sampling & Decoding",
1386 "bounding": [
1387 610,
1388 -30,
1389 720,
1390 620
1391 ],
1392 "color": "#3f789e",
1393 "font_size": 24,
1394 "flags": []
1395 },
1396 {
1397 "id": 4,
1398 "title": "Save Video(Mp4)",
1399 "bounding": [
1400 1340,
1401 -30,
1402 880,
1403 1140
1404 ],
1405 "color": "#3f789e",
1406 "font_size": 24,
1407 "flags": []
1408 },
1409 {
1410 "id": 5,
1411 "title": "Save Video(WebP)",
1412 "bounding": [
1413 2240,
1414 -20,
1415 720,
1416 1130
1417 ],
1418 "color": "#3f789e",
1419 "font_size": 24,
1420 "flags": []
1421 },
1422 {
1423 "id": 9,
1424 "title": "Load models here",
1425 "bounding": [
1426 -290,
1427 -30,
1428 410,
1429 620
1430 ],
1431 "color": "#3f789e",
1432 "font_size": 24,
1433 "flags": []
1434 },
1435 {
1436 "id": 10,
1437 "title": "14B",
1438 "bounding": [
1439 -270,
1440 10,
1441 370,
1442 340
1443 ],
1444 "color": "#3f789e",
1445 "font_size": 24,
1446 "flags": []
1447 },
1448 {
1449 "id": 11,
1450 "title": "1.3B",
1451 "bounding": [
1452 -680,
1453 -30,
1454 370,
1455 330
1456 ],
1457 "color": "#3f789e",
1458 "font_size": 24,
1459 "flags": []
1460 },
1461 {
1462 "id": 13,
1463 "title": "Load control video",
1464 "bounding": [
1465 30,
1466 610,
1467 560,
1468 410
1469 ],
1470 "color": "#3f789e",
1471 "font_size": 24,
1472 "flags": []
1473 },
1474 {
1475 "id": 14,
1476 "title": "Load reference image",
1477 "bounding": [
1478 -290,
1479 610,
1480 300,
1481 410
1482 ],
1483 "color": "#3f789e",
1484 "font_size": 24,
1485 "flags": []
1486 },
1487 {
1488 "id": 15,
1489 "title": "Image preprocessing",
1490 "bounding": [
1491 610,
1492 610,
1493 720,
1494 410
1495 ],
1496 "color": "#3f789e",
1497 "font_size": 24,
1498 "flags": []
1499 }
1500 ],
1501 "config": [],
1502 "extra": {
1503 "ds": {
1504 "scale": 0.4064180107866041,
1505 "offset": [
1506 1027.671793047929,
1507 640.0501811606016
1508 ]
1509 },
1510 "frontendVersion": "1.27.10",
1511 "node_versions": {
1512 "comfy-core": "0.3.34"
1513 },
1514 "VHS_latentpreview": false,
1515 "VHS_latentpreviewrate": 0,
1516 "VHS_MetadataImage": true,
1517 "VHS_KeepIntermediate": true
1518 },
1519 "version": 0.4
1520}