工作流 /
视频 /
Wan2.1 VACE 文生视频
Wan2.1 VACE 文生视频
Wan2.1 VACE Text to Video
将文本描述转换为高质量视频。VACE-14B模型支持480p和720p两种分辨率。
视频
22 个节点 10 次使用 ComfyUI ≥ 0.3.31 ComfyUI 26 KB
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
× 5
UNETLoader
× 2
LoraLoader
× 2
CLIPLoader
× 2
CLIPTextEncode
× 2
CreateVideo
× 1
KSampler
× 1
WanVaceToVideo
× 1
ModelSamplingSD3
× 1
VAELoader
× 1
TrimVideoLatent
× 1
VAEDecode
× 1
SaveAnimatedWEBP
× 1
SaveVideo
× 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": 117,
5 "last_link_id": 192,
6 "nodes": [
7 {
8 "id": 68,
9 "type": "CreateVideo",
10 "pos": [
11 290,
12 690
13 ],
14 "size": [
15 270,
16 78
17 ],
18 "flags": {
19 "collapsed": false
20 },
21 "order": 20,
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 "Node name for S&R": "CreateVideo",
47 "cnr_id": "comfy-core",
48 "ver": "0.3.34"
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": 16,
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 "Node name for S&R": "KSampler",
102 "cnr_id": "comfy-core",
103 "ver": "0.3.34"
104 },
105 "widgets_values": [
106 675909971186865,
107 "randomize",
108 4,
109 1,
110 "uni_pc",
111 "simple",
112 1
113 ]
114 },
115 {
116 "id": 49,
117 "type": "WanVaceToVideo",
118 "pos": [
119 630,
120 60
121 ],
122 "size": [
123 315,
124 254
125 ],
126 "flags": [],
127 "order": 15,
128 "mode": 0,
129 "inputs": [
130 {
131 "name": "positive",
132 "type": "CONDITIONING",
133 "link": 96
134 },
135 {
136 "name": "negative",
137 "type": "CONDITIONING",
138 "link": 97
139 },
140 {
141 "name": "vae",
142 "type": "VAE",
143 "link": 183
144 },
145 {
146 "name": "control_video",
147 "shape": 7,
148 "type": "IMAGE",
149 "link": null
150 },
151 {
152 "name": "control_masks",
153 "shape": 7,
154 "type": "MASK",
155 "link": null
156 },
157 {
158 "name": "reference_image",
159 "shape": 7,
160 "type": "IMAGE",
161 "link": null
162 }
163 ],
164 "outputs": [
165 {
166 "name": "positive",
167 "type": "CONDITIONING",
168 "links": [
169 98
170 ]
171 },
172 {
173 "name": "negative",
174 "type": "CONDITIONING",
175 "links": [
176 99
177 ]
178 },
179 {
180 "name": "latent",
181 "type": "LATENT",
182 "links": [
183 160
184 ]
185 },
186 {
187 "name": "trim_latent",
188 "type": "INT",
189 "links": [
190 115
191 ]
192 }
193 ],
194 "properties": {
195 "Node name for S&R": "WanVaceToVideo",
196 "cnr_id": "comfy-core",
197 "ver": "0.3.34"
198 },
199 "widgets_values": [
200 1280,
201 720,
202 81,
203 1,
204 1
205 ]
206 },
207 {
208 "id": 48,
209 "type": "ModelSamplingSD3",
210 "pos": [
211 730,
212 10
213 ],
214 "size": [
215 210,
216 58
217 ],
218 "flags": {
219 "collapsed": true
220 },
221 "order": 12,
222 "mode": 0,
223 "inputs": [
224 {
225 "name": "model",
226 "type": "MODEL",
227 "link": 192
228 }
229 ],
230 "outputs": [
231 {
232 "name": "MODEL",
233 "type": "MODEL",
234 "slot_index": 0,
235 "links": [
236 95
237 ]
238 }
239 ],
240 "properties": {
241 "Node name for S&R": "ModelSamplingSD3",
242 "cnr_id": "comfy-core",
243 "ver": "0.3.34"
244 },
245 "widgets_values": [
246 8.000000000000002
247 ]
248 },
249 {
250 "id": 105,
251 "type": "VAELoader",
252 "pos": [
253 -260,
254 530
255 ],
256 "size": [
257 350,
258 58
259 ],
260 "flags": [],
261 "order": 0,
262 "mode": 0,
263 "inputs": [],
264 "outputs": [
265 {
266 "name": "VAE",
267 "type": "VAE",
268 "slot_index": 0,
269 "links": [
270 183,
271 184
272 ]
273 }
274 ],
275 "properties": {
276 "Node name for S&R": "VAELoader",
277 "cnr_id": "comfy-core",
278 "ver": "0.3.34",
279 "models": [
280 {
281 "name": "wan_2.1_vae.safetensors",
282 "url": "https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/vae/wan_2.1_vae.safetensors",
283 "directory": "vae"
284 }
285 ]
286 },
287 "widgets_values": [
288 "wan_2.1_vae.safetensors"
289 ],
290 "color": "#322",
291 "bgcolor": "#533"
292 },
293 {
294 "id": 116,
295 "type": "MarkdownNote",
296 "pos": [
297 640,
298 370
299 ],
300 "size": [
301 310,
302 110
303 ],
304 "flags": [],
305 "order": 1,
306 "mode": 0,
307 "inputs": [],
308 "outputs": [],
309 "title": "About Video Size",
310 "properties": [],
311 "widgets_values": [
312 "| 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) | ✅ | ✅ |"
313 ],
314 "color": "#222",
315 "bgcolor": "#000"
316 },
317 {
318 "id": 58,
319 "type": "TrimVideoLatent",
320 "pos": [
321 980,
322 570
323 ],
324 "size": [
325 276.5860290527344,
326 58
327 ],
328 "flags": {
329 "collapsed": true
330 },
331 "order": 17,
332 "mode": 0,
333 "inputs": [
334 {
335 "name": "samples",
336 "type": "LATENT",
337 "link": 116
338 },
339 {
340 "name": "trim_amount",
341 "type": "INT",
342 "widget": {
343 "name": "trim_amount"
344 },
345 "link": 115
346 }
347 ],
348 "outputs": [
349 {
350 "name": "LATENT",
351 "type": "LATENT",
352 "links": [
353 117
354 ]
355 }
356 ],
357 "properties": {
358 "Node name for S&R": "TrimVideoLatent",
359 "cnr_id": "comfy-core",
360 "ver": "0.3.34"
361 },
362 "widgets_values": [
363 0
364 ]
365 },
366 {
367 "id": 8,
368 "type": "VAEDecode",
369 "pos": [
370 1170,
371 570
372 ],
373 "size": [
374 210,
375 46
376 ],
377 "flags": {
378 "collapsed": true
379 },
380 "order": 18,
381 "mode": 0,
382 "inputs": [
383 {
384 "name": "samples",
385 "type": "LATENT",
386 "link": 117
387 },
388 {
389 "name": "vae",
390 "type": "VAE",
391 "link": 184
392 }
393 ],
394 "outputs": [
395 {
396 "name": "IMAGE",
397 "type": "IMAGE",
398 "slot_index": 0,
399 "links": [
400 130,
401 139
402 ]
403 }
404 ],
405 "properties": {
406 "Node name for S&R": "VAEDecode",
407 "cnr_id": "comfy-core",
408 "ver": "0.3.34"
409 },
410 "widgets_values": []
411 },
412 {
413 "id": 70,
414 "type": "SaveAnimatedWEBP",
415 "pos": [
416 630,
417 700
418 ],
419 "size": [
420 670,
421 800
422 ],
423 "flags": [],
424 "order": 19,
425 "mode": 4,
426 "inputs": [
427 {
428 "name": "images",
429 "type": "IMAGE",
430 "link": 130
431 }
432 ],
433 "outputs": [],
434 "properties": {
435 "Node name for S&R": "SaveAnimatedWEBP",
436 "cnr_id": "comfy-core",
437 "ver": "0.3.34"
438 },
439 "widgets_values": [
440 "ComfyUI",
441 6,
442 true,
443 80,
444 "default"
445 ]
446 },
447 {
448 "id": 108,
449 "type": "UNETLoader",
450 "pos": [
451 -670,
452 40
453 ],
454 "size": [
455 350,
456 82
457 ],
458 "flags": [],
459 "order": 2,
460 "mode": 4,
461 "inputs": [],
462 "outputs": [
463 {
464 "name": "MODEL",
465 "type": "MODEL",
466 "slot_index": 0,
467 "links": [
468 178
469 ]
470 }
471 ],
472 "properties": {
473 "Node name for S&R": "UNETLoader",
474 "cnr_id": "comfy-core",
475 "ver": "0.3.34",
476 "models": [
477 {
478 "name": "wan2.1_vace_1.3B_fp16.safetensors",
479 "url": "https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/diffusion_models/wan2.1_vace_1.3B_fp16.safetensors",
480 "directory": "diffusion_models"
481 }
482 ]
483 },
484 "widgets_values": [
485 "wan2.1_vace_1.3B_fp16.safetensors",
486 "default"
487 ],
488 "color": "#322",
489 "bgcolor": "#533"
490 },
491 {
492 "id": 109,
493 "type": "LoraLoader",
494 "pos": [
495 -670,
496 160
497 ],
498 "size": [
499 350,
500 126
501 ],
502 "flags": [],
503 "order": 10,
504 "mode": 4,
505 "inputs": [
506 {
507 "name": "model",
508 "type": "MODEL",
509 "link": 178
510 },
511 {
512 "name": "clip",
513 "type": "CLIP",
514 "link": 179
515 }
516 ],
517 "outputs": [
518 {
519 "name": "MODEL",
520 "type": "MODEL",
521 "links": []
522 },
523 {
524 "name": "CLIP",
525 "type": "CLIP",
526 "links": []
527 }
528 ],
529 "properties": {
530 "Node name for S&R": "LoraLoader",
531 "cnr_id": "comfy-core",
532 "ver": "0.3.38",
533 "models": [
534 {
535 "name": "Wan21_CausVid_bidirect2_T2V_1_3B_lora_rank32.safetensors",
536 "url": "https://huggingface.co/Kijai/WanVideo_comfy/resolve/main/Wan21_CausVid_bidirect2_T2V_1_3B_lora_rank32.safetensors",
537 "directory": "loras"
538 }
539 ]
540 },
541 "widgets_values": [
542 "Wan21_CausVid_bidirect2_T2V_1_3B_lora_rank32.safetensors",
543 0.7000000000000002,
544 1
545 ],
546 "color": "#322",
547 "bgcolor": "#533"
548 },
549 {
550 "id": 111,
551 "type": "CLIPLoader",
552 "pos": [
553 -680,
554 350
555 ],
556 "size": [
557 370,
558 106
559 ],
560 "flags": [],
561 "order": 3,
562 "mode": 4,
563 "inputs": [],
564 "outputs": [
565 {
566 "name": "CLIP",
567 "type": "CLIP",
568 "slot_index": 0,
569 "links": []
570 }
571 ],
572 "properties": {
573 "Node name for S&R": "CLIPLoader",
574 "cnr_id": "comfy-core",
575 "ver": "0.3.34",
576 "models": [
577 {
578 "name": "umt5_xxl_fp8_e4m3fn_scaled.safetensors",
579 "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",
580 "directory": "text_encoders"
581 }
582 ]
583 },
584 "widgets_values": [
585 "umt5_xxl_fp8_e4m3fn_scaled.safetensors",
586 "wan",
587 "default"
588 ],
589 "color": "#322",
590 "bgcolor": "#533"
591 },
592 {
593 "id": 110,
594 "type": "CLIPLoader",
595 "pos": [
596 -260,
597 380
598 ],
599 "size": [
600 350,
601 106
602 ],
603 "flags": [],
604 "order": 4,
605 "mode": 0,
606 "inputs": [],
607 "outputs": [
608 {
609 "name": "CLIP",
610 "type": "CLIP",
611 "slot_index": 0,
612 "links": [
613 179,
614 189
615 ]
616 }
617 ],
618 "properties": {
619 "Node name for S&R": "CLIPLoader",
620 "cnr_id": "comfy-core",
621 "ver": "0.3.34",
622 "models": [
623 {
624 "name": "umt5_xxl_fp16.safetensors",
625 "url": "https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/text_encoders/umt5_xxl_fp16.safetensors",
626 "directory": "text_encoders"
627 }
628 ]
629 },
630 "widgets_values": [
631 "umt5_xxl_fp16.safetensors",
632 "wan",
633 "default"
634 ],
635 "color": "#322",
636 "bgcolor": "#533"
637 },
638 {
639 "id": 107,
640 "type": "LoraLoader",
641 "pos": [
642 -260,
643 210
644 ],
645 "size": [
646 350,
647 126
648 ],
649 "flags": [],
650 "order": 11,
651 "mode": 0,
652 "inputs": [
653 {
654 "name": "model",
655 "type": "MODEL",
656 "link": 188
657 },
658 {
659 "name": "clip",
660 "type": "CLIP",
661 "link": 189
662 }
663 ],
664 "outputs": [
665 {
666 "name": "MODEL",
667 "type": "MODEL",
668 "links": [
669 192
670 ]
671 },
672 {
673 "name": "CLIP",
674 "type": "CLIP",
675 "links": [
676 190,
677 191
678 ]
679 }
680 ],
681 "properties": {
682 "Node name for S&R": "LoraLoader",
683 "cnr_id": "comfy-core",
684 "ver": "0.3.38",
685 "models": [
686 {
687 "name": "Wan21_CausVid_14B_T2V_lora_rank32.safetensors",
688 "url": "https://huggingface.co/Kijai/WanVideo_comfy/resolve/main/Wan21_CausVid_14B_T2V_lora_rank32.safetensors",
689 "directory": "loras"
690 }
691 ]
692 },
693 "widgets_values": [
694 "Wan21_CausVid_14B_T2V_lora_rank32.safetensors",
695 0.7000000000000002,
696 1
697 ],
698 "color": "#322",
699 "bgcolor": "#533"
700 },
701 {
702 "id": 112,
703 "type": "MarkdownNote",
704 "pos": [
705 980,
706 350
707 ],
708 "size": [
709 320,
710 170
711 ],
712 "flags": [],
713 "order": 5,
714 "mode": 0,
715 "inputs": [],
716 "outputs": [],
717 "title": "KSampler Setting",
718 "properties": [],
719 "widgets_values": [
720 "## 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"
721 ],
722 "color": "#222",
723 "bgcolor": "#000"
724 },
725 {
726 "id": 6,
727 "type": "CLIPTextEncode",
728 "pos": [
729 150,
730 50
731 ],
732 "size": [
733 430,
734 290
735 ],
736 "flags": [],
737 "order": 13,
738 "mode": 0,
739 "inputs": [
740 {
741 "name": "clip",
742 "type": "CLIP",
743 "link": 190
744 }
745 ],
746 "outputs": [
747 {
748 "name": "CONDITIONING",
749 "type": "CONDITIONING",
750 "slot_index": 0,
751 "links": [
752 96
753 ]
754 }
755 ],
756 "title": "CLIP Text Encode (Positive Prompt)",
757 "properties": {
758 "Node name for S&R": "CLIPTextEncode",
759 "cnr_id": "comfy-core",
760 "ver": "0.3.34"
761 },
762 "widgets_values": [
763 "Fujifilm Portra 400H film still, slammed Nissan Skyline R33 GTR LM JGTC, in heavy motion blur, 7-11 Tokyo, Midnight"
764 ],
765 "color": "#232",
766 "bgcolor": "#353"
767 },
768 {
769 "id": 7,
770 "type": "CLIPTextEncode",
771 "pos": [
772 150,
773 400
774 ],
775 "size": [
776 425.27801513671875,
777 180.6060791015625
778 ],
779 "flags": [],
780 "order": 14,
781 "mode": 0,
782 "inputs": [
783 {
784 "name": "clip",
785 "type": "CLIP",
786 "link": 191
787 }
788 ],
789 "outputs": [
790 {
791 "name": "CONDITIONING",
792 "type": "CONDITIONING",
793 "slot_index": 0,
794 "links": [
795 97
796 ]
797 }
798 ],
799 "title": "CLIP Text Encode (Negative Prompt)",
800 "properties": {
801 "Node name for S&R": "CLIPTextEncode",
802 "cnr_id": "comfy-core",
803 "ver": "0.3.34"
804 },
805 "widgets_values": [
806 "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,"
807 ],
808 "color": "#322",
809 "bgcolor": "#533"
810 },
811 {
812 "id": 114,
813 "type": "SaveVideo",
814 "pos": [
815 -270,
816 810
817 ],
818 "size": [
819 830,
820 928
821 ],
822 "flags": [],
823 "order": 21,
824 "mode": 0,
825 "inputs": [
826 {
827 "name": "video",
828 "type": "VIDEO",
829 "link": 185
830 }
831 ],
832 "outputs": [],
833 "properties": {
834 "Node name for S&R": "SaveVideo",
835 "cnr_id": "comfy-core",
836 "ver": "0.3.38"
837 },
838 "widgets_values": [
839 "video/ComfyUI",
840 "auto",
841 "auto"
842 ]
843 },
844 {
845 "id": 106,
846 "type": "UNETLoader",
847 "pos": [
848 -260,
849 90
850 ],
851 "size": [
852 350,
853 82
854 ],
855 "flags": [],
856 "order": 6,
857 "mode": 0,
858 "inputs": [],
859 "outputs": [
860 {
861 "name": "MODEL",
862 "type": "MODEL",
863 "slot_index": 0,
864 "links": [
865 188
866 ]
867 }
868 ],
869 "properties": {
870 "Node name for S&R": "UNETLoader",
871 "cnr_id": "comfy-core",
872 "ver": "0.3.34",
873 "models": [
874 {
875 "name": "wan2.1_vace_14B_fp16.safetensors",
876 "url": "https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/diffusion_models/wan2.1_vace_14B_fp16.safetensors",
877 "directory": "diffusion_models"
878 }
879 ]
880 },
881 "widgets_values": [
882 "wan2.1_vace_14B_fp16.safetensors",
883 "default"
884 ],
885 "color": "#322",
886 "bgcolor": "#533"
887 },
888 {
889 "id": 117,
890 "type": "MarkdownNote",
891 "pos": [
892 -290,
893 -150
894 ],
895 "size": [
896 410,
897 100
898 ],
899 "flags": [],
900 "order": 7,
901 "mode": 0,
902 "inputs": [],
903 "outputs": [],
904 "title": "About CausVid LoRA",
905 "properties": [],
906 "widgets_values": [
907 "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."
908 ],
909 "color": "#222",
910 "bgcolor": "#000"
911 },
912 {
913 "id": 113,
914 "type": "MarkdownNote",
915 "pos": [
916 -680,
917 500
918 ],
919 "size": [
920 370,
921 130
922 ],
923 "flags": [],
924 "order": 8,
925 "mode": 0,
926 "inputs": [],
927 "outputs": [],
928 "title": "Note",
929 "properties": [],
930 "widgets_values": [
931 "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"
932 ],
933 "color": "#222",
934 "bgcolor": "#000"
935 },
936 {
937 "id": 77,
938 "type": "MarkdownNote",
939 "pos": [
940 -1170,
941 0
942 ],
943 "size": [
944 470,
945 730
946 ],
947 "flags": [],
948 "order": 9,
949 "mode": 0,
950 "inputs": [],
951 "outputs": [],
952 "properties": [],
953 "widgets_values": [
954 "[Tutorial](https://docs.comfy.org/tutorials/video/wan/vace) | [教程](https://docs.comfy.org/zh-CN/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"
955 ],
956 "color": "#222",
957 "bgcolor": "#000"
958 }
959 ],
960 "links": [
961 [
962 95,
963 48,
964 0,
965 3,
966 0,
967 "MODEL"
968 ],
969 [
970 96,
971 6,
972 0,
973 49,
974 0,
975 "CONDITIONING"
976 ],
977 [
978 97,
979 7,
980 0,
981 49,
982 1,
983 "CONDITIONING"
984 ],
985 [
986 98,
987 49,
988 0,
989 3,
990 1,
991 "CONDITIONING"
992 ],
993 [
994 99,
995 49,
996 1,
997 3,
998 2,
999 "CONDITIONING"
1000 ],
1001 [
1002 115,
1003 49,
1004 3,
1005 58,
1006 1,
1007 "INT"
1008 ],
1009 [
1010 116,
1011 3,
1012 0,
1013 58,
1014 0,
1015 "LATENT"
1016 ],
1017 [
1018 117,
1019 58,
1020 0,
1021 8,
1022 0,
1023 "LATENT"
1024 ],
1025 [
1026 130,
1027 8,
1028 0,
1029 70,
1030 0,
1031 "IMAGE"
1032 ],
1033 [
1034 139,
1035 8,
1036 0,
1037 68,
1038 0,
1039 "IMAGE"
1040 ],
1041 [
1042 160,
1043 49,
1044 2,
1045 3,
1046 3,
1047 "LATENT"
1048 ],
1049 [
1050 178,
1051 108,
1052 0,
1053 109,
1054 0,
1055 "MODEL"
1056 ],
1057 [
1058 179,
1059 110,
1060 0,
1061 109,
1062 1,
1063 "CLIP"
1064 ],
1065 [
1066 183,
1067 105,
1068 0,
1069 49,
1070 2,
1071 "VAE"
1072 ],
1073 [
1074 184,
1075 105,
1076 0,
1077 8,
1078 1,
1079 "VAE"
1080 ],
1081 [
1082 185,
1083 68,
1084 0,
1085 114,
1086 0,
1087 "VIDEO"
1088 ],
1089 [
1090 188,
1091 106,
1092 0,
1093 107,
1094 0,
1095 "MODEL"
1096 ],
1097 [
1098 189,
1099 110,
1100 0,
1101 107,
1102 1,
1103 "CLIP"
1104 ],
1105 [
1106 190,
1107 107,
1108 1,
1109 6,
1110 0,
1111 "CLIP"
1112 ],
1113 [
1114 191,
1115 107,
1116 1,
1117 7,
1118 0,
1119 "CLIP"
1120 ],
1121 [
1122 192,
1123 107,
1124 0,
1125 48,
1126 0,
1127 "MODEL"
1128 ]
1129 ],
1130 "groups": [
1131 {
1132 "id": 2,
1133 "title": "Prompt",
1134 "bounding": [
1135 140,
1136 -30,
1137 450,
1138 620
1139 ],
1140 "color": "#3f789e",
1141 "font_size": 24,
1142 "flags": []
1143 },
1144 {
1145 "id": 3,
1146 "title": "Sampling & Decoding",
1147 "bounding": [
1148 610,
1149 -30,
1150 720,
1151 620
1152 ],
1153 "color": "#3f789e",
1154 "font_size": 24,
1155 "flags": []
1156 },
1157 {
1158 "id": 4,
1159 "title": "Save Video(Mp4)",
1160 "bounding": [
1161 -290,
1162 610,
1163 880,
1164 1140
1165 ],
1166 "color": "#3f789e",
1167 "font_size": 24,
1168 "flags": []
1169 },
1170 {
1171 "id": 5,
1172 "title": "Save Video(WebP)",
1173 "bounding": [
1174 610,
1175 610,
1176 720,
1177 1130
1178 ],
1179 "color": "#3f789e",
1180 "font_size": 24,
1181 "flags": []
1182 },
1183 {
1184 "id": 9,
1185 "title": "Load models here",
1186 "bounding": [
1187 -290,
1188 -30,
1189 410,
1190 620
1191 ],
1192 "color": "#3f789e",
1193 "font_size": 24,
1194 "flags": []
1195 },
1196 {
1197 "id": 10,
1198 "title": "14B",
1199 "bounding": [
1200 -270,
1201 10,
1202 370,
1203 340
1204 ],
1205 "color": "#3f789e",
1206 "font_size": 24,
1207 "flags": []
1208 },
1209 {
1210 "id": 11,
1211 "title": "1.3B",
1212 "bounding": [
1213 -680,
1214 -30,
1215 370,
1216 330
1217 ],
1218 "color": "#3f789e",
1219 "font_size": 24,
1220 "flags": []
1221 }
1222 ],
1223 "config": [],
1224 "extra": {
1225 "ds": {
1226 "scale": 0.6727499949325629,
1227 "offset": [
1228 1597.6037842126482,
1229 318.3006235572395
1230 ]
1231 },
1232 "frontendVersion": "1.24.0-1",
1233 "node_versions": {
1234 "comfy-core": "0.3.34"
1235 },
1236 "VHS_latentpreview": false,
1237 "VHS_latentpreviewrate": 0,
1238 "VHS_MetadataImage": true,
1239 "VHS_KeepIntermediate": true
1240 },
1241 "version": 0.4
1242}