ComfyPilotComfyPilot
工作流 / 视频 / FastVideo FastH3:文生视频
FastVideo FastH3:文生视频

FastVideo FastH3:文生视频

FastVideo FastH3: Text to Video

使用蒸馏八步模型直接从文本生成同步的视频和音频,输出一个视频,无需任何文件输入。非常适合快速文生视频原型设计、音视频内容创作以及创意叙事,在这些场景中,快速生成比最大细节更重要。

视频 5 个节点 423 次使用 ComfyUI ≥ 0.36.0 ComfyUI 59 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 4 个权重文件,其中 3 个能在本站模型库里找到。

fastvideo_fasth3_8step_v2_pruned_int8_convrot.safetensors 模型库暂未收录
minimax_h3_audio_vae_fp32.safetensors 在模型库查看
minimax_h3_video_vae_int8_convrot.safetensors 在模型库查看
qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors 在模型库查看

节点构成

MarkdownNote × 2
SaveVideo × 1
4c314f31-ecda-4b08-ae98-faaba1bf613f × 1
ResolutionSelector × 1

官方标注的模型

MiniMax H3

工作流 JSON

和上面「下载工作流 JSON」是同一份文件。

1{
2 "id": "e3f2b845-8f2c-4b5a-9caf-eac1029d3e7e",
3 "revision": 0,
4 "last_node_id": 148,
5 "last_link_id": 288,
6 "nodes": [
7 {
8 "id": 92,
9 "type": "SaveVideo",
10 "pos": [
11 -370,
12 4880
13 ],
14 "size": [
15 900,
16 126
17 ],
18 "flags": [],
19 "order": 4,
20 "mode": 0,
21 "inputs": [
22 {
23 "name": "video",
24 "type": "VIDEO",
25 "link": 194
26 }
27 ],
28 "outputs": [
29 {
30 "name": "video",
31 "type": "VIDEO",
32 "links": null
33 }
34 ],
35 "properties": {
36 "cnr_id": "comfy-core",
37 "ver": "0.33.0"
38 },
39 "widgets_values": [
40 "video/MiniMax_H3",
41 "auto",
42 "auto",
43 "auto"
44 ],
45 "widgets_values_named": {
46 "filename_prefix": "video/MiniMax_H3",
47 "format": "auto",
48 "format.codec": "auto",
49 "codec": "auto"
50 }
51 },
52 {
53 "id": 105,
54 "type": "4c314f31-ecda-4b08-ae98-faaba1bf613f",
55 "pos": [
56 -900,
57 4880
58 ],
59 "size": [
60 480,
61 630
62 ],
63 "flags": [],
64 "order": 3,
65 "mode": 0,
66 "inputs": [
67 {
68 "name": "first_frame",
69 "shape": 7,
70 "type": "IMAGE",
71 "link": null
72 },
73 {
74 "name": "last_frame",
75 "shape": 7,
76 "type": "IMAGE",
77 "link": null
78 },
79 {
80 "name": "width",
81 "type": "INT",
82 "widget": {
83 "name": "width"
84 },
85 "link": 273
86 },
87 {
88 "name": "height",
89 "type": "INT",
90 "widget": {
91 "name": "height"
92 },
93 "link": 274
94 },
95 {
96 "label": "duration",
97 "name": "value_1",
98 "type": "FLOAT",
99 "widget": {
100 "name": "value_1"
101 },
102 "link": null
103 },
104 {
105 "label": "audio_vae",
106 "name": "vae_name_1",
107 "type": "COMBO",
108 "widget": {
109 "name": "vae_name_1"
110 },
111 "link": null
112 }
113 ],
114 "outputs": [
115 {
116 "name": "VIDEO",
117 "type": "VIDEO",
118 "links": [
119 194
120 ]
121 }
122 ],
123 "properties": {
124 "cnr_id": "comfy-core",
125 "ver": "0.33.0",
126 "previewExposures": []
127 },
128 "widgets_values": [
129 "integrated_multimodal_description: [Shot 1] Cinematic, a wide-angle shot filmed from an extreme low-angle, worm's-eye perspective. In the extreme foreground of the square frame, a massive tan suede work boot features a thick, chunky brown rubber sole and tightly laced brown shoelaces, rendered slightly out of focus. Above the boot, a young adult man wears loose-fitting, olive green cargo pants, an unbuttoned dark navy blue heavy-cotton overshirt, and a tan baseball cap. He has a serious expression, looking directly down into the lens. Behind him, a modern skyscraper made of reflective glass panels stretches into an overcast, pale grey sky. Large, bold, white typography reading \"STREETWEAR\" bursts into the frame, positioned seamlessly between the man's back and the skyscraper, creating a three-dimensional interlaced title sequence effect. The man lowers his raised foot, and the massive tan boot descends rapidly, striking the pavement. He then raises both hands to adjust his cap's brim. The camera pushes in slowly, intensifying the visual scale of the figure and the floating text. [Shot 2] At 00:02.100, the camera cuts to a medium shot from a slightly elevated low-angle vantage point. The man from Shot 1 now stands fully planted on the ground against the sheer glass facade of the skyscraper. Fast-paced, neon-yellow text reading \"VOL. 1\" flashes rapidly in the extreme foreground, momentarily overlapping his navy jacket. He drops his hands and executes a sharp, forceful turn of his shoulders to the right, striking a dynamic, angular pose. His jacket flares out, revealing a neutral-toned t-shirt underneath. His gaze shifts away from the camera, looking off into the distance. The camera arcs right at a medium speed, tracking his rotation and capturing the shifting reflections of the city on the glass building behind him. [Shot 3] At 00:03.600, the camera cuts to a wide tracking shot positioned just above the pavement level. The man from Shot 1 strides purposefully forward along the grey concrete sidewalk. Clean, white typography reading \"AVAILABLE NOW\" scrolls vertically down the right side of the frame, weaving behind his swinging right arm to maintain the interlaced text aesthetic. The focus is directed toward the lower half of his body, highlighting the movement of his olive green cargo pants as they wrinkle dynamically. His heavy tan boots stomp rhythmically on the concrete. The camera pulls out at a fast speed, maintaining a constant distance just ahead of his advancing boots while the vertical grid lines of the skyscraper blur in the background.\noverall_soundscape: A faint, continuous urban hum of distant city traffic provides a subtle ambient baseline. A loud, resonant thud is heard prominently in the foreground as the heavy tan boot strikes the concrete pavement, followed by the crisp, audible rustle of thick cotton and denim as the man adjusts his cap. When the shot changes, a sharp, distinct swish of heavy fabric accompanies his sudden shoulder rotation. Finally, a rhythmic series of heavy, pronounced footsteps—thick rubber soles slapping firmly against the concrete—dominates the foreground, accompanied by the continuous, rough rustling of cargo pants.\nnon_diegetic_music: Electronic bass music, fast tempo, featuring a heavy, rhythmic drum machine beat, deep synthesizer drones, and sudden aggressive bass drops with no melodic swell.",
130 1344,
131 768,
132 5,
133 631153434041465,
134 "fastvideo_fasth3_8step_v2_pruned_int8_convrot.safetensors",
135 "qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors",
136 "minimax_h3_video_vae_int8_convrot.safetensors",
137 "minimax_h3_audio_vae_fp32.safetensors",
138 "comfy kitchen attention"
139 ],
140 "widgets_values_named": {
141 "prompt": "integrated_multimodal_description: [Shot 1] Cinematic, a wide-angle shot filmed from an extreme low-angle, worm's-eye perspective. In the extreme foreground of the square frame, a massive tan suede work boot features a thick, chunky brown rubber sole and tightly laced brown shoelaces, rendered slightly out of focus. Above the boot, a young adult man wears loose-fitting, olive green cargo pants, an unbuttoned dark navy blue heavy-cotton overshirt, and a tan baseball cap. He has a serious expression, looking directly down into the lens. Behind him, a modern skyscraper made of reflective glass panels stretches into an overcast, pale grey sky. Large, bold, white typography reading \"STREETWEAR\" bursts into the frame, positioned seamlessly between the man's back and the skyscraper, creating a three-dimensional interlaced title sequence effect. The man lowers his raised foot, and the massive tan boot descends rapidly, striking the pavement. He then raises both hands to adjust his cap's brim. The camera pushes in slowly, intensifying the visual scale of the figure and the floating text. [Shot 2] At 00:02.100, the camera cuts to a medium shot from a slightly elevated low-angle vantage point. The man from Shot 1 now stands fully planted on the ground against the sheer glass facade of the skyscraper. Fast-paced, neon-yellow text reading \"VOL. 1\" flashes rapidly in the extreme foreground, momentarily overlapping his navy jacket. He drops his hands and executes a sharp, forceful turn of his shoulders to the right, striking a dynamic, angular pose. His jacket flares out, revealing a neutral-toned t-shirt underneath. His gaze shifts away from the camera, looking off into the distance. The camera arcs right at a medium speed, tracking his rotation and capturing the shifting reflections of the city on the glass building behind him. [Shot 3] At 00:03.600, the camera cuts to a wide tracking shot positioned just above the pavement level. The man from Shot 1 strides purposefully forward along the grey concrete sidewalk. Clean, white typography reading \"AVAILABLE NOW\" scrolls vertically down the right side of the frame, weaving behind his swinging right arm to maintain the interlaced text aesthetic. The focus is directed toward the lower half of his body, highlighting the movement of his olive green cargo pants as they wrinkle dynamically. His heavy tan boots stomp rhythmically on the concrete. The camera pulls out at a fast speed, maintaining a constant distance just ahead of his advancing boots while the vertical grid lines of the skyscraper blur in the background.\noverall_soundscape: A faint, continuous urban hum of distant city traffic provides a subtle ambient baseline. A loud, resonant thud is heard prominently in the foreground as the heavy tan boot strikes the concrete pavement, followed by the crisp, audible rustle of thick cotton and denim as the man adjusts his cap. When the shot changes, a sharp, distinct swish of heavy fabric accompanies his sudden shoulder rotation. Finally, a rhythmic series of heavy, pronounced footsteps—thick rubber soles slapping firmly against the concrete—dominates the foreground, accompanied by the continuous, rough rustling of cargo pants.\nnon_diegetic_music: Electronic bass music, fast tempo, featuring a heavy, rhythmic drum machine beat, deep synthesizer drones, and sudden aggressive bass drops with no melodic swell.",
142 "width": 1344,
143 "height": 768,
144 "value_1": 5,
145 "noise_seed": 631153434041465,
146 "unet_name": "fastvideo_fasth3_8step_v2_pruned_int8_convrot.safetensors",
147 "clip_name": "qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors",
148 "vae_name": "minimax_h3_video_vae_int8_convrot.safetensors",
149 "vae_name_1": "minimax_h3_audio_vae_fp32.safetensors",
150 "attention": "comfy kitchen attention"
151 }
152 },
153 {
154 "id": 139,
155 "type": "MarkdownNote",
156 "pos": [
157 -1730,
158 4880
159 ],
160 "size": [
161 450,
162 630
163 ],
164 "flags": [],
165 "order": 0,
166 "mode": 0,
167 "inputs": [],
168 "outputs": [],
169 "title": "Note: Fast Video FastH3",
170 "properties": [],
171 "widgets_values": [
172 "## About this workflow\n\nThis template runs the **Text to Video** task (`MiniMaxH3ImageToVideo` node with no images connected) using the [FastVideo FastH3 8-Step V2](https://huggingface.co/FastVideo/FastVideo-FastH3-8-Step-V2) checkpoint, a DMD2-distilled model that generates synchronized video and audio in **8 sampling steps** instead of the base model's full schedule.\n\n**Checkpoint repo**: [FastVideo/FastVideo-FastH3-Comfy](https://huggingface.co/FastVideo/FastVideo-FastH3-Comfy)\n\n**Scope**: this distilled checkpoint supports t2va only. FL2VA (first/last-frame) and Ref2VA (multi-reference conditioning) were not distilled; use the base MiniMax H3 model for image-conditioned tasks.\n\n**Key inputs**\n\n- **prompt**: describe the shots, motion, and the accompanying audio (dialogue, SFX, music) in one block\n- **width / height**: set via Resolution Selector. H3's native canvas is a 768px short edge, capped at 768x1344 pixels, rounded to a multiple of 32\n- **duration (seconds)**: converted to a valid frame `length` by the Math Expression node, snapping up to the model's 17-frame-per-block (17k+5) grid at 24fps"
173 ],
174 "widgets_values_named": {
175 "text": "## About this workflow\n\nThis template runs the **Text to Video** task (`MiniMaxH3ImageToVideo` node with no images connected) using the [FastVideo FastH3 8-Step V2](https://huggingface.co/FastVideo/FastVideo-FastH3-8-Step-V2) checkpoint, a DMD2-distilled model that generates synchronized video and audio in **8 sampling steps** instead of the base model's full schedule.\n\n**Checkpoint repo**: [FastVideo/FastVideo-FastH3-Comfy](https://huggingface.co/FastVideo/FastVideo-FastH3-Comfy)\n\n**Scope**: this distilled checkpoint supports t2va only. FL2VA (first/last-frame) and Ref2VA (multi-reference conditioning) were not distilled; use the base MiniMax H3 model for image-conditioned tasks.\n\n**Key inputs**\n\n- **prompt**: describe the shots, motion, and the accompanying audio (dialogue, SFX, music) in one block\n- **width / height**: set via Resolution Selector. H3's native canvas is a 768px short edge, capped at 768x1344 pixels, rounded to a multiple of 32\n- **duration (seconds)**: converted to a valid frame `length` by the Math Expression node, snapping up to the model's 17-frame-per-block (17k+5) grid at 24fps"
176 },
177 "color": "#222",
178 "bgcolor": "#000"
179 },
180 {
181 "id": 140,
182 "type": "MarkdownNote",
183 "pos": [
184 -2220,
185 4880
186 ],
187 "size": [
188 430,
189 650
190 ],
191 "flags": [],
192 "order": 1,
193 "mode": 0,
194 "inputs": [],
195 "outputs": [],
196 "title": "Note: Model Links",
197 "properties": [],
198 "widgets_values": [
199 "Guide: [Subgraph](https://docs.comfy.org/interface/features/subgraph)\n\n## Model Links\n\n- [Hugging Face:Comfy-Org/MiniMax-H3](https://huggingface.co/Comfy-Org/MiniMax-H3)\n\n**ModelScope mirrors** (alternative repo location)\n\n- [ModelScope:Comfy-Org/MiniMax-H3](https://modelscope.cn/models/Comfy-Org/MiniMax-H3)\n\n**vae**\n\n- [minimax_h3_video_vae_int8_convrot.safetensors](https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/vae/minimax_h3_video_vae_int8_convrot.safetensors) (2.62 GB)\n- [minimax_h3_audio_vae_fp32.safetensors](https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/vae/minimax_h3_audio_vae_fp32.safetensors) (577.2 MB)\n\n**diffusion_models**\n\n- [fastvideo_fasth3_8step_v2_pruned_int8_convrot.safetensors](https://huggingface.co/FastVideo/FastVideo-FastH3-Comfy/resolve/main/diffusion_models/fastvideo_fasth3_8step_v2_pruned_int8_convrot.safetensors) (20.61 GB)\n\n**text_encoders**\n\n- [qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors](https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/text_encoders/qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors) (14.61 GB)\n\n## Model Storage Location\n\n```\n📂 ComfyUI/\n├── 📂 models/\n│ ├── 📂 vae/\n│ │ ├── minimax_h3_video_vae_int8_convrot.safetensors\n│ │ └── minimax_h3_audio_vae_fp32.safetensors\n│ ├── 📂 diffusion_models/\n│ │ └── fastvideo_fasth3_8step_v2_pruned_int8_convrot.safetensors\n│ └── 📂 text_encoders/\n│ └── qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors\n```\n\n## Report Issue\n\nNote: Please update ComfyUI first ([guide](https://docs.comfy.org/installation/update_comfyui)) and prepare required models. Desktop/Cloud updates follow stable releases, so some nightly-supported models may not be available yet.\n\n- Cannot run / runtime errors: [ComfyUI/issues](https://github.com/comfyanonymous/ComfyUI/issues)\n- UI / frontend issues: [ComfyUI_frontend/issues](https://github.com/Comfy-Org/ComfyUI_frontend/issues)\n- Workflow issues: [workflow_templates/issues](https://github.com/Comfy-Org/workflow_templates/issues)\n"
200 ],
201 "widgets_values_named": {
202 "text": "Guide: [Subgraph](https://docs.comfy.org/interface/features/subgraph)\n\n## Model Links\n\n- [Hugging Face:Comfy-Org/MiniMax-H3](https://huggingface.co/Comfy-Org/MiniMax-H3)\n\n**ModelScope mirrors** (alternative repo location)\n\n- [ModelScope:Comfy-Org/MiniMax-H3](https://modelscope.cn/models/Comfy-Org/MiniMax-H3)\n\n**vae**\n\n- [minimax_h3_video_vae_int8_convrot.safetensors](https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/vae/minimax_h3_video_vae_int8_convrot.safetensors) (2.62 GB)\n- [minimax_h3_audio_vae_fp32.safetensors](https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/vae/minimax_h3_audio_vae_fp32.safetensors) (577.2 MB)\n\n**diffusion_models**\n\n- [fastvideo_fasth3_8step_v2_pruned_int8_convrot.safetensors](https://huggingface.co/FastVideo/FastVideo-FastH3-Comfy/resolve/main/diffusion_models/fastvideo_fasth3_8step_v2_pruned_int8_convrot.safetensors) (20.61 GB)\n\n**text_encoders**\n\n- [qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors](https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/text_encoders/qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors) (14.61 GB)\n\n## Model Storage Location\n\n```\n📂 ComfyUI/\n├── 📂 models/\n│ ├── 📂 vae/\n│ │ ├── minimax_h3_video_vae_int8_convrot.safetensors\n│ │ └── minimax_h3_audio_vae_fp32.safetensors\n│ ├── 📂 diffusion_models/\n│ │ └── fastvideo_fasth3_8step_v2_pruned_int8_convrot.safetensors\n│ └── 📂 text_encoders/\n│ └── qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors\n```\n\n## Report Issue\n\nNote: Please update ComfyUI first ([guide](https://docs.comfy.org/installation/update_comfyui)) and prepare required models. Desktop/Cloud updates follow stable releases, so some nightly-supported models may not be available yet.\n\n- Cannot run / runtime errors: [ComfyUI/issues](https://github.com/comfyanonymous/ComfyUI/issues)\n- UI / frontend issues: [ComfyUI_frontend/issues](https://github.com/Comfy-Org/ComfyUI_frontend/issues)\n- Workflow issues: [workflow_templates/issues](https://github.com/Comfy-Org/workflow_templates/issues)\n"
203 },
204 "color": "#222",
205 "bgcolor": "#000"
206 },
207 {
208 "id": 143,
209 "type": "ResolutionSelector",
210 "pos": [
211 -1230,
212 5230
213 ],
214 "size": [
215 270,
216 190
217 ],
218 "flags": [],
219 "order": 2,
220 "mode": 0,
221 "showAdvanced": true,
222 "inputs": [],
223 "outputs": [
224 {
225 "name": "width",
226 "type": "INT",
227 "links": [
228 273
229 ]
230 },
231 {
232 "name": "height",
233 "type": "INT",
234 "links": [
235 274
236 ]
237 }
238 ],
239 "properties": {
240 "Node name for S&R": "ResolutionSelector"
241 },
242 "widgets_values": [
243 "1:1 (Square)",
244 0.4,
245 32
246 ],
247 "widgets_values_named": {
248 "aspect_ratio": "1:1 (Square)",
249 "megapixels": 0.4,
250 "multiple": 32
251 }
252 }
253 ],
254 "links": [
255 [
256 194,
257 105,
258 0,
259 92,
260 0,
261 "VIDEO"
262 ],
263 [
264 273,
265 143,
266 0,
267 105,
268 2,
269 "INT"
270 ],
271 [
272 274,
273 143,
274 1,
275 105,
276 3,
277 "INT"
278 ]
279 ],
280 "groups": [],
281 "definitions": {
282 "subgraphs": [
283 {
284 "id": "4c314f31-ecda-4b08-ae98-faaba1bf613f",
285 "version": 1,
286 "state": {
287 "lastGroupId": 7,
288 "lastNodeId": 148,
289 "lastLinkId": 288,
290 "lastRerouteId": 0
291 },
292 "revision": 0,
293 "config": [],
294 "name": "First Last Frame to Video (FastVideo FastH3)",
295 "inputNode": {
296 "id": -10,
297 "bounding": [
298 -2640,
299 4810,
300 128,
301 288
302 ]
303 },
304 "outputNode": {
305 "id": -20,
306 "bounding": [
307 2130,
308 5070,
309 128,
310 68
311 ]
312 },
313 "inputs": [
314 {
315 "id": "d6eaa195-2266-4016-b7ed-b2d17a3e53c2",
316 "name": "first_frame",
317 "type": "IMAGE",
318 "linkIds": [
319 195
320 ],
321 "pos": [
322 -2536,
323 4834
324 ]
325 },
326 {
327 "id": "03b95567-f496-4279-9d38-989dd34fa882",
328 "name": "last_frame",
329 "type": "IMAGE",
330 "linkIds": [
331 196
332 ],
333 "pos": [
334 -2536,
335 4854
336 ]
337 },
338 {
339 "id": "d7302ca7-24ed-44c4-8ac9-736540dab7fb",
340 "name": "prompt",
341 "type": "STRING",
342 "linkIds": [
343 197
344 ],
345 "pos": [
346 -2536,
347 4874
348 ]
349 },
350 {
351 "id": "709e2d94-3172-496f-ad86-7d672f3df568",
352 "name": "width",
353 "type": "INT",
354 "linkIds": [
355 200
356 ],
357 "pos": [
358 -2536,
359 4894
360 ]
361 },
362 {
363 "id": "8205b85b-19bb-47f6-8418-89cd42940c1d",
364 "name": "height",
365 "type": "INT",
366 "linkIds": [
367 201
368 ],
369 "pos": [
370 -2536,
371 4914
372 ]
373 },
374 {
375 "id": "a40e5e96-4307-4a8e-a4f8-13a27d4bc9d3",
376 "name": "value_1",
377 "type": "FLOAT",
378 "linkIds": [
379 206
380 ],
381 "label": "duration",
382 "pos": [
383 -2536,
384 4934
385 ]
386 },
387 {
388 "id": "9f8734bc-c2b3-43b0-911e-e89560e8b777",
389 "name": "noise_seed",
390 "type": "INT",
391 "linkIds": [
392 207
393 ],
394 "pos": [
395 -2536,
396 4954
397 ]
398 },
399 {
400 "id": "9a6d2811-673c-47e0-aa94-dda7c83621e4",
401 "name": "unet_name",
402 "type": "COMBO",
403 "linkIds": [
404 221
405 ],
406 "pos": [
407 -2536,
408 4974
409 ]
410 },
411 {
412 "id": "7100a787-5536-4002-840b-1d2fd071010a",
413 "name": "clip_name",
414 "type": "COMBO",
415 "linkIds": [
416 227
417 ],
418 "pos": [
419 -2536,
420 4994
421 ]
422 },
423 {
424 "id": "d40632fd-e0ae-4319-b85d-5479db4d6e11",
425 "name": "vae_name",
426 "type": "COMBO",
427 "linkIds": [
428 223
429 ],
430 "pos": [
431 -2536,
432 5014
433 ]
434 },
435 {
436 "id": "2992852c-4b20-439a-8771-c866ec1996e6",
437 "name": "vae_name_1",
438 "type": "COMBO",
439 "linkIds": [
440 224
441 ],
442 "label": "audio_vae",
443 "pos": [
444 -2536,
445 5034
446 ]
447 },
448 {
449 "id": "42566ff0-1439-4215-9724-33d6c0282e21",
450 "name": "attention",
451 "type": "COMBO",
452 "linkIds": [
453 258
454 ],
455 "pos": [
456 -2536,
457 5054
458 ]
459 }
460 ],
461 "outputs": [
462 {
463 "id": "adb2611c-e490-4fca-8067-b718c992f8cc",
464 "name": "VIDEO",
465 "type": "VIDEO",
466 "linkIds": [
467 168
468 ],
469 "localized_name": "VIDEO",
470 "pos": [
471 2154,
472 5094
473 ]
474 }
475 ],
476 "widgets": [],
477 "nodes": [
478 {
479 "id": 11,
480 "type": "VAELoader",
481 "pos": [
482 -2030,
483 5110
484 ],
485 "size": [
486 640,
487 100
488 ],
489 "flags": [],
490 "order": 4,
491 "mode": 0,
492 "inputs": [
493 {
494 "localized_name": "vae_name",
495 "name": "vae_name",
496 "type": "COMBO",
497 "widget": {
498 "name": "vae_name"
499 },
500 "link": 223
501 }
502 ],
503 "outputs": [
504 {
505 "localized_name": "VAE",
506 "name": "VAE",
507 "type": "VAE",
508 "links": [
509 8,
510 190
511 ]
512 }
513 ],
514 "properties": {
515 "Node name for S&R": "VAELoader",
516 "cnr_id": "comfy-core",
517 "ver": "0.33.0",
518 "models": [
519 {
520 "name": "minimax_h3_video_vae_int8_convrot.safetensors",
521 "url": "https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/vae/minimax_h3_video_vae_int8_convrot.safetensors",
522 "directory": "vae"
523 }
524 ]
525 },
526 "widgets_values": [
527 "minimax_h3_video_vae_int8_convrot.safetensors"
528 ],
529 "widgets_values_named": {
530 "vae_name": "minimax_h3_video_vae_int8_convrot.safetensors"
531 }
532 },
533 {
534 "id": 24,
535 "type": "VAELoader",
536 "pos": [
537 -2030,
538 5290
539 ],
540 "size": [
541 650,
542 100
543 ],
544 "flags": [],
545 "order": 10,
546 "mode": 0,
547 "inputs": [
548 {
549 "localized_name": "vae_name",
550 "name": "vae_name",
551 "type": "COMBO",
552 "widget": {
553 "name": "vae_name"
554 },
555 "link": 224
556 }
557 ],
558 "outputs": [
559 {
560 "localized_name": "VAE",
561 "name": "VAE",
562 "type": "VAE",
563 "links": [
564 23
565 ]
566 }
567 ],
568 "properties": {
569 "Node name for S&R": "VAELoader",
570 "cnr_id": "comfy-core",
571 "ver": "0.33.0",
572 "models": [
573 {
574 "name": "minimax_h3_audio_vae_fp32.safetensors",
575 "url": "https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/vae/minimax_h3_audio_vae_fp32.safetensors",
576 "directory": "vae"
577 }
578 ]
579 },
580 "widgets_values": [
581 "minimax_h3_audio_vae_fp32.safetensors"
582 ],
583 "widgets_values_named": {
584 "vae_name": "minimax_h3_audio_vae_fp32.safetensors"
585 }
586 },
587 {
588 "id": 23,
589 "type": "VAEDecodeAudio",
590 "pos": [
591 10,
592 4990
593 ],
594 "size": [
595 230,
596 80
597 ],
598 "flags": {
599 "collapsed": false
600 },
601 "order": 9,
602 "mode": 0,
603 "inputs": [
604 {
605 "localized_name": "samples",
606 "name": "samples",
607 "type": "LATENT",
608 "link": 226
609 },
610 {
611 "localized_name": "vae",
612 "name": "vae",
613 "type": "VAE",
614 "link": 23
615 }
616 ],
617 "outputs": [
618 {
619 "localized_name": "AUDIO",
620 "name": "AUDIO",
621 "type": "AUDIO",
622 "links": [
623 166
624 ]
625 }
626 ],
627 "properties": {
628 "Node name for S&R": "VAEDecodeAudio",
629 "cnr_id": "comfy-core",
630 "ver": "0.33.0"
631 }
632 },
633 {
634 "id": 10,
635 "type": "VAEDecode",
636 "pos": [
637 10,
638 4850
639 ],
640 "size": [
641 230,
642 80
643 ],
644 "flags": {
645 "collapsed": false
646 },
647 "order": 3,
648 "mode": 0,
649 "inputs": [
650 {
651 "localized_name": "samples",
652 "name": "samples",
653 "type": "LATENT",
654 "link": 225
655 },
656 {
657 "localized_name": "vae",
658 "name": "vae",
659 "type": "VAE",
660 "link": 8
661 }
662 ],
663 "outputs": [
664 {
665 "localized_name": "IMAGE",
666 "name": "IMAGE",
667 "type": "IMAGE",
668 "links": [
669 167
670 ]
671 }
672 ],
673 "properties": {
674 "Node name for S&R": "VAEDecode",
675 "cnr_id": "comfy-core",
676 "ver": "0.33.0"
677 }
678 },
679 {
680 "id": 17,
681 "type": "KSamplerSelect",
682 "pos": [
683 -760,
684 4950
685 ],
686 "size": [
687 370,
688 90
689 ],
690 "flags": [],
691 "order": 0,
692 "mode": 0,
693 "inputs": [],
694 "outputs": [
695 {
696 "localized_name": "SAMPLER",
697 "name": "SAMPLER",
698 "type": "SAMPLER",
699 "links": [
700 16
701 ]
702 }
703 ],
704 "properties": {
705 "Node name for S&R": "KSamplerSelect",
706 "cnr_id": "comfy-core",
707 "ver": "0.33.0"
708 },
709 "widgets_values": [
710 "res_multistep"
711 ],
712 "widgets_values_named": {
713 "sampler_name": "res_multistep"
714 }
715 },
716 {
717 "id": 9,
718 "type": "BasicScheduler",
719 "pos": [
720 -760,
721 5100
722 ],
723 "size": [
724 370,
725 150
726 ],
727 "flags": [],
728 "order": 2,
729 "mode": 0,
730 "inputs": [
731 {
732 "localized_name": "model",
733 "name": "model",
734 "type": "MODEL",
735 "link": 285
736 }
737 ],
738 "outputs": [
739 {
740 "localized_name": "SIGMAS",
741 "name": "SIGMAS",
742 "type": "SIGMAS",
743 "links": [
744 18
745 ]
746 }
747 ],
748 "properties": {
749 "Node name for S&R": "BasicScheduler",
750 "cnr_id": "comfy-core",
751 "ver": "0.33.0"
752 },
753 "widgets_values": [
754 "simple",
755 8,
756 1
757 ],
758 "widgets_values_named": {
759 "scheduler": "simple",
760 "steps": 8,
761 "denoise": 1
762 }
763 },
764 {
765 "id": 14,
766 "type": "SamplerCustomAdvanced",
767 "pos": [
768 -340,
769 4780
770 ],
771 "size": [
772 230,
773 390
774 ],
775 "flags": [],
776 "order": 6,
777 "mode": 0,
778 "inputs": [
779 {
780 "localized_name": "noise",
781 "name": "noise",
782 "type": "NOISE",
783 "link": 40
784 },
785 {
786 "localized_name": "guider",
787 "name": "guider",
788 "type": "GUIDER",
789 "link": 12
790 },
791 {
792 "localized_name": "sampler",
793 "name": "sampler",
794 "type": "SAMPLER",
795 "link": 16
796 },
797 {
798 "localized_name": "sigmas",
799 "name": "sigmas",
800 "type": "SIGMAS",
801 "link": 18
802 },
803 {
804 "localized_name": "latent_image",
805 "name": "latent_image",
806 "type": "LATENT",
807 "link": 188
808 }
809 ],
810 "outputs": [
811 {
812 "localized_name": "output",
813 "name": "output",
814 "type": "LATENT",
815 "links": [
816 225,
817 226
818 ]
819 },
820 {
821 "localized_name": "denoised_output",
822 "name": "denoised_output",
823 "type": "LATENT",
824 "links": null
825 }
826 ],
827 "properties": {
828 "Node name for S&R": "SamplerCustomAdvanced",
829 "cnr_id": "comfy-core",
830 "ver": "0.33.0"
831 }
832 },
833 {
834 "id": 16,
835 "type": "BasicGuider",
836 "pos": [
837 -760,
838 4820
839 ],
840 "size": [
841 360,
842 80
843 ],
844 "flags": [],
845 "order": 8,
846 "mode": 0,
847 "inputs": [
848 {
849 "localized_name": "model",
850 "name": "model",
851 "type": "MODEL",
852 "link": 286
853 },
854 {
855 "localized_name": "conditioning",
856 "name": "conditioning",
857 "type": "CONDITIONING",
858 "link": 187
859 }
860 ],
861 "outputs": [
862 {
863 "localized_name": "GUIDER",
864 "name": "GUIDER",
865 "type": "GUIDER",
866 "links": [
867 12
868 ]
869 }
870 ],
871 "properties": {
872 "Node name for S&R": "BasicGuider",
873 "cnr_id": "comfy-core",
874 "ver": "0.33.0"
875 }
876 },
877 {
878 "id": 6,
879 "type": "UNETLoader",
880 "pos": [
881 -2030,
882 4690
883 ],
884 "size": [
885 640,
886 120
887 ],
888 "flags": [],
889 "order": 1,
890 "mode": 0,
891 "inputs": [
892 {
893 "localized_name": "unet_name",
894 "name": "unet_name",
895 "type": "COMBO",
896 "widget": {
897 "name": "unet_name"
898 },
899 "link": 221
900 }
901 ],
902 "outputs": [
903 {
904 "localized_name": "MODEL",
905 "name": "MODEL",
906 "type": "MODEL",
907 "links": [
908 287
909 ]
910 }
911 ],
912 "properties": {
913 "Node name for S&R": "UNETLoader",
914 "cnr_id": "comfy-core",
915 "ver": "0.33.0",
916 "models": [
917 {
918 "name": "fastvideo_fasth3_8step_v2_pruned_int8_convrot.safetensors",
919 "url": "https://huggingface.co/FastVideo/FastVideo-FastH3-Comfy/resolve/main/diffusion_models/fastvideo_fasth3_8step_v2_pruned_int8_convrot.safetensors",
920 "directory": "diffusion_models"
921 }
922 ]
923 },
924 "widgets_values": [
925 "fastvideo_fasth3_8step_v2_pruned_int8_convrot.safetensors",
926 "default"
927 ],
928 "widgets_values_named": {
929 "unet_name": "fastvideo_fasth3_8step_v2_pruned_int8_convrot.safetensors",
930 "weight_dtype": "default"
931 }
932 },
933 {
934 "id": 13,
935 "type": "CLIPLoader",
936 "pos": [
937 -2030,
938 4880
939 ],
940 "size": [
941 640,
942 150
943 ],
944 "flags": [],
945 "order": 5,
946 "mode": 0,
947 "inputs": [
948 {
949 "localized_name": "clip_name",
950 "name": "clip_name",
951 "type": "COMBO",
952 "widget": {
953 "name": "clip_name"
954 },
955 "link": 227
956 }
957 ],
958 "outputs": [
959 {
960 "localized_name": "CLIP",
961 "name": "CLIP",
962 "type": "CLIP",
963 "links": [
964 189
965 ]
966 }
967 ],
968 "properties": {
969 "Node name for S&R": "CLIPLoader",
970 "cnr_id": "comfy-core",
971 "ver": "0.33.0",
972 "models": [
973 {
974 "name": "qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors",
975 "url": "https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/text_encoders/qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors",
976 "directory": "text_encoders"
977 }
978 ]
979 },
980 "widgets_values": [
981 "qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors",
982 "minimax",
983 "default"
984 ],
985 "widgets_values_named": {
986 "clip_name": "qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors",
987 "type": "minimax",
988 "device": "default"
989 }
990 },
991 {
992 "id": 15,
993 "type": "RandomNoise",
994 "pos": [
995 -760,
996 4660
997 ],
998 "size": [
999 360,
1000 90
1001 ],
1002 "flags": [],
1003 "order": 7,
1004 "mode": 0,
1005 "inputs": [
1006 {
1007 "localized_name": "noise_seed",
1008 "name": "noise_seed",
1009 "type": "INT",
1010 "widget": {
1011 "name": "noise_seed"
1012 },
1013 "link": 207
1014 }
1015 ],
1016 "outputs": [
1017 {
1018 "localized_name": "NOISE",
1019 "name": "NOISE",
1020 "type": "NOISE",
1021 "links": [
1022 40
1023 ]
1024 }
1025 ],
1026 "properties": {
1027 "Node name for S&R": "RandomNoise",
1028 "cnr_id": "comfy-core",
1029 "ver": "0.33.0"
1030 },
1031 "widgets_values": [
1032 1,
1033 "randomize"
1034 ],
1035 "widgets_values_named": {
1036 "noise_seed": 1,
1037 "control_after_generate": "randomize"
1038 }
1039 },
1040 {
1041 "id": 91,
1042 "type": "CreateVideo",
1043 "pos": [
1044 340,
1045 4890
1046 ],
1047 "size": [
1048 270,
1049 170
1050 ],
1051 "flags": [],
1052 "order": 11,
1053 "mode": 0,
1054 "inputs": [
1055 {
1056 "localized_name": "images",
1057 "name": "images",
1058 "type": "IMAGE",
1059 "link": 167
1060 },
1061 {
1062 "localized_name": "audio",
1063 "name": "audio",
1064 "shape": 7,
1065 "type": "AUDIO",
1066 "link": 166
1067 }
1068 ],
1069 "outputs": [
1070 {
1071 "localized_name": "VIDEO",
1072 "name": "VIDEO",
1073 "type": "VIDEO",
1074 "links": [
1075 168
1076 ]
1077 }
1078 ],
1079 "properties": {
1080 "Node name for S&R": "CreateVideo",
1081 "cnr_id": "comfy-core",
1082 "ver": "0.33.0"
1083 },
1084 "widgets_values": [
1085 24,
1086 "auto",
1087 "sRGB",
1088 "none"
1089 ],
1090 "widgets_values_named": {
1091 "fps": 24,
1092 "bit_depth": "auto",
1093 "color_space": "sRGB",
1094 "codec": "none"
1095 }
1096 },
1097 {
1098 "id": 104,
1099 "type": "MiniMaxH3ImageToVideo",
1100 "pos": [
1101 -1290,
1102 4740
1103 ],
1104 "size": [
1105 430,
1106 550
1107 ],
1108 "flags": [],
1109 "order": 12,
1110 "mode": 0,
1111 "inputs": [
1112 {
1113 "localized_name": "clip",
1114 "name": "clip",
1115 "type": "CLIP",
1116 "link": 189
1117 },
1118 {
1119 "localized_name": "vae",
1120 "name": "vae",
1121 "type": "VAE",
1122 "link": 190
1123 },
1124 {
1125 "localized_name": "first_frame",
1126 "name": "first_frame",
1127 "shape": 7,
1128 "type": "IMAGE",
1129 "link": 195
1130 },
1131 {
1132 "localized_name": "last_frame",
1133 "name": "last_frame",
1134 "shape": 7,
1135 "type": "IMAGE",
1136 "link": 196
1137 },
1138 {
1139 "localized_name": "prompt",
1140 "name": "prompt",
1141 "type": "STRING",
1142 "widget": {
1143 "name": "prompt"
1144 },
1145 "link": 197
1146 },
1147 {
1148 "localized_name": "width",
1149 "name": "width",
1150 "type": "INT",
1151 "widget": {
1152 "name": "width"
1153 },
1154 "link": 200
1155 },
1156 {
1157 "localized_name": "height",
1158 "name": "height",
1159 "type": "INT",
1160 "widget": {
1161 "name": "height"
1162 },
1163 "link": 201
1164 },
1165 {
1166 "localized_name": "length",
1167 "name": "length",
1168 "type": "INT",
1169 "widget": {
1170 "name": "length"
1171 },
1172 "link": 199
1173 }
1174 ],
1175 "outputs": [
1176 {
1177 "localized_name": "positive",
1178 "name": "positive",
1179 "type": "CONDITIONING",
1180 "links": [
1181 187
1182 ]
1183 },
1184 {
1185 "localized_name": "LATENT",
1186 "name": "LATENT",
1187 "type": "LATENT",
1188 "links": [
1189 188
1190 ]
1191 }
1192 ],
1193 "properties": {
1194 "Node name for S&R": "MiniMaxH3ImageToVideo",
1195 "cnr_id": "comfy-core",
1196 "ver": "0.33.0"
1197 },
1198 "widgets_values": [
1199 "Vaporwave title sequence look: pink and blue gradient palette, VHS tracking artifacts, Greek statue motifs, chrome palm trees, RGB chromatic aberration, lo-fi retro atmosphere, mood languid and nostalgic.\n\nTimeline:\n[0s-1s] VHS static opens the frame, the title \"COMFYUI\" appears with RGB split and a slight horizontal jitter.\n[1s-2.5s] Hard cut, a Greek plaster bust close-up, pink-purple gradient sky, a pixelated sun.\n[2.5s-4s] Clean \"STARRING\" credits appear, \"LATENT\" and \"CONTROLNET\" each shown exactly once.\n[4s-5s] Final card \"DIRECTED BY COMFYUI\" holds, one VHS tracking glitch settling into stability.\n\nHard cuts only, transitions landing with tape jumps, no push-ins, no dissolves.\n\nAudio: lo-fi vaporwave score, slow drum machine with soft bass, VHS tape-noise sample joins at 2.5s, melody fading for the last 1s.\n\nAll text must be clearly legible, do not misspell English, no Chinese characters, do not repeat names or job titles, no soft dissolves, no subtitle bars.",
1200 1344,
1201 768,
1202 73
1203 ],
1204 "widgets_values_named": {
1205 "prompt": "Vaporwave title sequence look: pink and blue gradient palette, VHS tracking artifacts, Greek statue motifs, chrome palm trees, RGB chromatic aberration, lo-fi retro atmosphere, mood languid and nostalgic.\n\nTimeline:\n[0s-1s] VHS static opens the frame, the title \"COMFYUI\" appears with RGB split and a slight horizontal jitter.\n[1s-2.5s] Hard cut, a Greek plaster bust close-up, pink-purple gradient sky, a pixelated sun.\n[2.5s-4s] Clean \"STARRING\" credits appear, \"LATENT\" and \"CONTROLNET\" each shown exactly once.\n[4s-5s] Final card \"DIRECTED BY COMFYUI\" holds, one VHS tracking glitch settling into stability.\n\nHard cuts only, transitions landing with tape jumps, no push-ins, no dissolves.\n\nAudio: lo-fi vaporwave score, slow drum machine with soft bass, VHS tape-noise sample joins at 2.5s, melody fading for the last 1s.\n\nAll text must be clearly legible, do not misspell English, no Chinese characters, do not repeat names or job titles, no soft dissolves, no subtitle bars.",
1206 "width": 1344,
1207 "height": 768,
1208 "length": 73
1209 }
1210 },
1211 {
1212 "id": 107,
1213 "type": "ComfyMathExpression",
1214 "pos": [
1215 -1720,
1216 5600
1217 ],
1218 "size": [
1219 360,
1220 170
1221 ],
1222 "flags": {
1223 "collapsed": false
1224 },
1225 "order": 13,
1226 "mode": 0,
1227 "inputs": [
1228 {
1229 "label": "a",
1230 "localized_name": "a",
1231 "name": "values.a",
1232 "type": "FLOAT,INT,BOOLEAN",
1233 "link": 205
1234 },
1235 {
1236 "localized_name": "b",
1237 "name": "values.b",
1238 "shape": 7,
1239 "type": "FLOAT,INT,BOOLEAN",
1240 "link": null
1241 }
1242 ],
1243 "outputs": [
1244 {
1245 "localized_name": "FLOAT",
1246 "name": "FLOAT",
1247 "type": "FLOAT",
1248 "links": null
1249 },
1250 {
1251 "localized_name": "INT",
1252 "name": "INT",
1253 "type": "INT",
1254 "links": [
1255 199
1256 ]
1257 },
1258 {
1259 "localized_name": "BOOL",
1260 "name": "BOOL",
1261 "type": "BOOLEAN",
1262 "links": null
1263 }
1264 ],
1265 "properties": {
1266 "Node name for S&R": "ComfyMathExpression",
1267 "cnr_id": "comfy-core",
1268 "ver": "0.33.0"
1269 },
1270 "widgets_values": [
1271 "max(5, round(a * 24)) + (5 - (max(5, round(a * 24)) % 17)) % 17"
1272 ],
1273 "widgets_values_named": {
1274 "expression": "max(5, round(a * 24)) + (5 - (max(5, round(a * 24)) % 17)) % 17"
1275 }
1276 },
1277 {
1278 "id": 111,
1279 "type": "PrimitiveFloat",
1280 "pos": [
1281 -2030,
1282 5600
1283 ],
1284 "size": [
1285 270,
1286 90
1287 ],
1288 "flags": [],
1289 "order": 14,
1290 "mode": 0,
1291 "inputs": [
1292 {
1293 "localized_name": "value",
1294 "name": "value",
1295 "type": "FLOAT",
1296 "widget": {
1297 "name": "value"
1298 },
1299 "link": 206
1300 }
1301 ],
1302 "outputs": [
1303 {
1304 "localized_name": "FLOAT",
1305 "name": "FLOAT",
1306 "type": "FLOAT",
1307 "links": [
1308 205
1309 ]
1310 }
1311 ],
1312 "title": "Float (duration)",
1313 "properties": {
1314 "Node name for S&R": "PrimitiveFloat",
1315 "cnr_id": "comfy-core",
1316 "ver": "0.33.0"
1317 },
1318 "widgets_values": [
1319 2
1320 ],
1321 "widgets_values_named": {
1322 "value": 2
1323 }
1324 },
1325 {
1326 "id": 127,
1327 "type": "BlockSparseAttention",
1328 "pos": [
1329 -770,
1330 4240
1331 ],
1332 "size": [
1333 270,
1334 250
1335 ],
1336 "flags": [],
1337 "order": 15,
1338 "mode": 0,
1339 "inputs": [
1340 {
1341 "localized_name": "model",
1342 "name": "model",
1343 "type": "MODEL",
1344 "link": 249
1345 }
1346 ],
1347 "outputs": [
1348 {
1349 "localized_name": "model",
1350 "name": "model",
1351 "type": "MODEL",
1352 "links": [
1353 285,
1354 286
1355 ]
1356 }
1357 ],
1358 "properties": {
1359 "Node name for S&R": "BlockSparseAttention"
1360 },
1361 "widgets_values": [
1362 "vsa",
1363 10,
1364 0.2,
1365 1,
1366 "",
1367 12288,
1368 256,
1369 "exact_kv_and_rows",
1370 false
1371 ],
1372 "widgets_values_named": {
1373 "selection": "vsa",
1374 "selection.keep_percent": 10,
1375 "start_percent": 0.2,
1376 "end_percent": 1,
1377 "dense_blocks": "",
1378 "min_tokens": 12288,
1379 "extra_tokens": 256,
1380 "sink_conditioning": "exact_kv_and_rows",
1381 "verbose": false
1382 }
1383 },
1384 {
1385 "id": 128,
1386 "type": "ModelAttentionBackend",
1387 "pos": [
1388 -1120,
1389 4240
1390 ],
1391 "size": [
1392 300,
1393 130
1394 ],
1395 "flags": [],
1396 "order": 16,
1397 "mode": 0,
1398 "inputs": [
1399 {
1400 "localized_name": "model",
1401 "name": "model",
1402 "type": "MODEL",
1403 "link": 288
1404 },
1405 {
1406 "localized_name": "backend",
1407 "name": "attention",
1408 "type": "COMBO",
1409 "widget": {
1410 "name": "attention"
1411 },
1412 "link": 258
1413 }
1414 ],
1415 "outputs": [
1416 {
1417 "localized_name": "model",
1418 "name": "model",
1419 "type": "MODEL",
1420 "links": [
1421 249
1422 ]
1423 }
1424 ],
1425 "properties": {
1426 "Node name for S&R": "ModelAttentionBackend"
1427 },
1428 "widgets_values": [
1429 "comfy kitchen attention"
1430 ],
1431 "widgets_values_named": {
1432 "attention": "comfy kitchen attention"
1433 }
1434 },
1435 {
1436 "id": 148,
1437 "type": "MiniMaxH3SigmaShift",
1438 "pos": [
1439 -1740,
1440 4250
1441 ],
1442 "size": [
1443 280,
1444 90
1445 ],
1446 "flags": [],
1447 "order": 17,
1448 "mode": 0,
1449 "inputs": [
1450 {
1451 "localized_name": "model",
1452 "name": "model",
1453 "type": "MODEL",
1454 "link": 287
1455 }
1456 ],
1457 "outputs": [
1458 {
1459 "localized_name": "MODEL",
1460 "name": "MODEL",
1461 "type": "MODEL",
1462 "links": [
1463 288
1464 ]
1465 }
1466 ],
1467 "properties": {
1468 "Node name for S&R": "MiniMaxH3SigmaShift"
1469 },
1470 "widgets_values": [
1471 10,
1472 3
1473 ],
1474 "widgets_values_named": {
1475 "shift_video": 10,
1476 "shift_audio": 3
1477 }
1478 }
1479 ],
1480 "groups": [
1481 {
1482 "id": 1,
1483 "title": "Models",
1484 "bounding": [
1485 -2050,
1486 4540,
1487 700,
1488 980
1489 ],
1490 "color": "#3f789e",
1491 "flags": []
1492 },
1493 {
1494 "id": 2,
1495 "title": "Sampling",
1496 "bounding": [
1497 -790,
1498 4540,
1499 700,
1500 980
1501 ],
1502 "color": "#3f789e",
1503 "flags": []
1504 },
1505 {
1506 "id": 3,
1507 "title": "Conditioning",
1508 "bounding": [
1509 -1310,
1510 4540,
1511 480,
1512 980
1513 ],
1514 "color": "#3f789e",
1515 "flags": []
1516 },
1517 {
1518 "id": 4,
1519 "title": "Decoding and create video",
1520 "bounding": [
1521 -50,
1522 4540,
1523 700,
1524 980
1525 ],
1526 "color": "#3f789e",
1527 "flags": []
1528 },
1529 {
1530 "id": 7,
1531 "title": "Sparse Attention (VSA)",
1532 "bounding": [
1533 -1310,
1534 4100,
1535 1220,
1536 420
1537 ],
1538 "color": "#3f789e",
1539 "flags": []
1540 }
1541 ],
1542 "links": [
1543 {
1544 "id": 23,
1545 "origin_id": 24,
1546 "origin_slot": 0,
1547 "target_id": 23,
1548 "target_slot": 1,
1549 "type": "VAE"
1550 },
1551 {
1552 "id": 8,
1553 "origin_id": 11,
1554 "origin_slot": 0,
1555 "target_id": 10,
1556 "target_slot": 1,
1557 "type": "VAE"
1558 },
1559 {
1560 "id": 40,
1561 "origin_id": 15,
1562 "origin_slot": 0,
1563 "target_id": 14,
1564 "target_slot": 0,
1565 "type": "NOISE"
1566 },
1567 {
1568 "id": 12,
1569 "origin_id": 16,
1570 "origin_slot": 0,
1571 "target_id": 14,
1572 "target_slot": 1,
1573 "type": "GUIDER"
1574 },
1575 {
1576 "id": 16,
1577 "origin_id": 17,
1578 "origin_slot": 0,
1579 "target_id": 14,
1580 "target_slot": 2,
1581 "type": "SAMPLER"
1582 },
1583 {
1584 "id": 18,
1585 "origin_id": 9,
1586 "origin_slot": 0,
1587 "target_id": 14,
1588 "target_slot": 3,
1589 "type": "SIGMAS"
1590 },
1591 {
1592 "id": 188,
1593 "origin_id": 104,
1594 "origin_slot": 1,
1595 "target_id": 14,
1596 "target_slot": 4,
1597 "type": "LATENT"
1598 },
1599 {
1600 "id": 187,
1601 "origin_id": 104,
1602 "origin_slot": 0,
1603 "target_id": 16,
1604 "target_slot": 1,
1605 "type": "CONDITIONING"
1606 },
1607 {
1608 "id": 167,
1609 "origin_id": 10,
1610 "origin_slot": 0,
1611 "target_id": 91,
1612 "target_slot": 0,
1613 "type": "IMAGE"
1614 },
1615 {
1616 "id": 166,
1617 "origin_id": 23,
1618 "origin_slot": 0,
1619 "target_id": 91,
1620 "target_slot": 1,
1621 "type": "AUDIO"
1622 },
1623 {
1624 "id": 189,
1625 "origin_id": 13,
1626 "origin_slot": 0,
1627 "target_id": 104,
1628 "target_slot": 0,
1629 "type": "CLIP"
1630 },
1631 {
1632 "id": 190,
1633 "origin_id": 11,
1634 "origin_slot": 0,
1635 "target_id": 104,
1636 "target_slot": 1,
1637 "type": "VAE"
1638 },
1639 {
1640 "id": 168,
1641 "origin_id": 91,
1642 "origin_slot": 0,
1643 "target_id": -20,
1644 "target_slot": 0,
1645 "type": "VIDEO"
1646 },
1647 {
1648 "id": 195,
1649 "origin_id": -10,
1650 "origin_slot": 0,
1651 "target_id": 104,
1652 "target_slot": 2,
1653 "type": "IMAGE"
1654 },
1655 {
1656 "id": 196,
1657 "origin_id": -10,
1658 "origin_slot": 1,
1659 "target_id": 104,
1660 "target_slot": 3,
1661 "type": "IMAGE"
1662 },
1663 {
1664 "id": 197,
1665 "origin_id": -10,
1666 "origin_slot": 2,
1667 "target_id": 104,
1668 "target_slot": 4,
1669 "type": "STRING"
1670 },
1671 {
1672 "id": 199,
1673 "origin_id": 107,
1674 "origin_slot": 1,
1675 "target_id": 104,
1676 "target_slot": 7,
1677 "type": "INT"
1678 },
1679 {
1680 "id": 200,
1681 "origin_id": -10,
1682 "origin_slot": 3,
1683 "target_id": 104,
1684 "target_slot": 5,
1685 "type": "INT"
1686 },
1687 {
1688 "id": 201,
1689 "origin_id": -10,
1690 "origin_slot": 4,
1691 "target_id": 104,
1692 "target_slot": 6,
1693 "type": "INT"
1694 },
1695 {
1696 "id": 205,
1697 "origin_id": 111,
1698 "origin_slot": 0,
1699 "target_id": 107,
1700 "target_slot": 0,
1701 "type": "FLOAT"
1702 },
1703 {
1704 "id": 206,
1705 "origin_id": -10,
1706 "origin_slot": 5,
1707 "target_id": 111,
1708 "target_slot": 0,
1709 "type": "FLOAT"
1710 },
1711 {
1712 "id": 207,
1713 "origin_id": -10,
1714 "origin_slot": 6,
1715 "target_id": 15,
1716 "target_slot": 0,
1717 "type": "INT"
1718 },
1719 {
1720 "id": 221,
1721 "origin_id": -10,
1722 "origin_slot": 7,
1723 "target_id": 6,
1724 "target_slot": 0,
1725 "type": "COMBO"
1726 },
1727 {
1728 "id": 223,
1729 "origin_id": -10,
1730 "origin_slot": 9,
1731 "target_id": 11,
1732 "target_slot": 0,
1733 "type": "COMBO"
1734 },
1735 {
1736 "id": 224,
1737 "origin_id": -10,
1738 "origin_slot": 10,
1739 "target_id": 24,
1740 "target_slot": 0,
1741 "type": "COMBO"
1742 },
1743 {
1744 "id": 225,
1745 "origin_id": 14,
1746 "origin_slot": 0,
1747 "target_id": 10,
1748 "target_slot": 0,
1749 "type": "LATENT"
1750 },
1751 {
1752 "id": 226,
1753 "origin_id": 14,
1754 "origin_slot": 0,
1755 "target_id": 23,
1756 "target_slot": 0,
1757 "type": "LATENT"
1758 },
1759 {
1760 "id": 227,
1761 "origin_id": -10,
1762 "origin_slot": 8,
1763 "target_id": 13,
1764 "target_slot": 0,
1765 "type": "COMBO"
1766 },
1767 {
1768 "id": 249,
1769 "origin_id": 128,
1770 "origin_slot": 0,
1771 "target_id": 127,
1772 "target_slot": 0,
1773 "type": "MODEL"
1774 },
1775 {
1776 "id": 258,
1777 "origin_id": -10,
1778 "origin_slot": 11,
1779 "target_id": 128,
1780 "target_slot": 1,
1781 "type": "COMBO"
1782 },
1783 {
1784 "id": 285,
1785 "origin_id": 127,
1786 "origin_slot": 0,
1787 "target_id": 9,
1788 "target_slot": 0,
1789 "type": "MODEL"
1790 },
1791 {
1792 "id": 286,
1793 "origin_id": 127,
1794 "origin_slot": 0,
1795 "target_id": 16,
1796 "target_slot": 0,
1797 "type": "MODEL"
1798 },
1799 {
1800 "id": 287,
1801 "origin_id": 6,
1802 "origin_slot": 0,
1803 "target_id": 148,
1804 "target_slot": 0,
1805 "type": "MODEL"
1806 },
1807 {
1808 "id": 288,
1809 "origin_id": 148,
1810 "origin_slot": 0,
1811 "target_id": 128,
1812 "target_slot": 0,
1813 "type": "MODEL"
1814 }
1815 ],
1816 "extra": []
1817 }
1818 ]
1819 },
1820 "config": [],
1821 "extra": {
1822 "ds": {
1823 "scale": 0.4319568878088475,
1824 "offset": [
1825 2654.2688833338193,
1826 -4336.063111969756
1827 ]
1828 },
1829 "frontendVersion": "1.52.7",
1830 "VHS_KeepIntermediate": true,
1831 "VHS_MetadataImage": true,
1832 "VHS_latentpreview": false,
1833 "VHS_latentpreviewrate": 0
1834 },
1835 "version": 0.4
1836}