{ "id": "9ae50e0e-7618-4b4d-a5ef-5fff699d254c", "revision": 0, "last_node_id": 101, "last_link_id": 219, "nodes": [ { "id": 8, "type": "VAEDecode", "pos": [ 1578.4067385408707, 167.36048381577803 ], "size": [ 225, 46 ], "flags": {}, "order": 17, "mode": 0, "inputs": [ { "name": "samples", "type": "LATENT", "link": 35 }, { "name": "vae", "type": "VAE", "link": 76 } ], "outputs": [ { "name": "IMAGE", "type": "IMAGE", "slot_index": 0, "links": [ 101 ] } ], "properties": { "Node name for S&R": "VAEDecode", "cnr_id": "comfy-core", "ver": "0.3.33" } }, { "id": 37, "type": "UNETLoader", "pos": [ 876.8100510941658, 34.805270385742176 ], "size": [ 305.3782043457031, 82 ], "flags": {}, "order": 0, "mode": 0, "inputs": [], "outputs": [ { "name": "MODEL", "type": "MODEL", "slot_index": 0, "links": [ 105 ] } ], "properties": { "Node name for S&R": "UNETLoader", "cnr_id": "comfy-core", "ver": "0.3.33" }, "widgets_values": [ "flux3_image_klein_sd_int8_convrot.safetensors", "default" ], "widgets_values_named": { "unet_name": "flux3_image_klein_sd_int8_convrot.safetensors", "weight_dtype": "default" }, "color": "#323", "bgcolor": "#535" }, { "id": 38, "type": "CLIPLoader", "pos": [ 67.38904068787144, 186.7020461700888 ], "size": [ 301.3524169921875, 106 ], "flags": {}, "order": 1, "mode": 0, "inputs": [], "outputs": [ { "name": "CLIP", "type": "CLIP", "slot_index": 0, "links": [ 74 ] } ], "properties": { "Node name for S&R": "CLIPLoader", "cnr_id": "comfy-core", "ver": "0.3.33" }, "widgets_values": [ "qwen3vl_4b_int8_convrot.safetensors", "flux3", "default" ], "widgets_values_named": { "clip_name": "qwen3vl_4b_int8_convrot.safetensors", "type": "flux3", "device": "default" }, "color": "#432", "bgcolor": "#653" }, { "id": 3, "type": "KSampler", "pos": [ 1224.7286989900895, 167.36048381577803 ], "size": [ 315, 262 ], "flags": {}, "order": 16, "mode": 0, "inputs": [ { "name": "model", "type": "MODEL", "link": 105 }, { "name": "positive", "type": "CONDITIONING", "link": 162 }, { "name": "negative", "type": "CONDITIONING", "link": 163 }, { "name": "latent_image", "type": "LATENT", "link": 164 } ], "outputs": [ { "name": "LATENT", "type": "LATENT", "slot_index": 0, "links": [ 35 ] } ], "properties": { "Node name for S&R": "KSampler", "cnr_id": "comfy-core", "ver": "0.3.33" }, "widgets_values": [ 1234, "fixed", 15, 1, "euler", "simple", 1 ], "widgets_values_named": { "seed": 1234, "control_after_generate": "fixed", "steps": 15, "cfg": 1, "sampler_name": "euler", "scheduler": "simple", "denoise": 1 } }, { "id": 6, "type": "CLIPTextEncode", "pos": [ 415.7513025457208, 186.7020461700888 ], "size": [ 418.5183049982245, 177.7917699892659 ], "flags": {}, "order": 13, "mode": 0, "inputs": [ { "name": "clip", "type": "CLIP", "link": 74 }, { "name": "text", "type": "STRING", "widget": { "name": "text" }, "link": 217 } ], "outputs": [ { "name": "CONDITIONING", "type": "CONDITIONING", "slot_index": 0, "links": [ 113, 160 ] } ], "title": "CLIP Text Encode (Positive Prompt)", "properties": { "Node name for S&R": "CLIPTextEncode", "cnr_id": "comfy-core", "ver": "0.3.33" }, "widgets_values": [ "Add a shallow water puddle on the ground around her bare feet. Change her red shirt and gray plaid trousers to a one-piece swimsuit. " ], "widgets_values_named": { "text": "Add a shallow water puddle on the ground around her bare feet. Change her red shirt and gray plaid trousers to a one-piece swimsuit. " } }, { "id": 56, "type": "SaveImage", "pos": [ 1848.0136535092056, 167.36048381577803 ], "size": [ 794.7253433548476, 660.1491545655919 ], "flags": {}, "order": 18, "mode": 0, "inputs": [ { "name": "images", "type": "IMAGE", "link": 101 } ], "outputs": [ { "name": "images", "type": "IMAGE", "links": null } ], "properties": { "cnr_id": "comfy-core", "ver": "0.3.75" }, "widgets_values": [ "ComfyUI" ], "widgets_values_named": { "filename_prefix": "ComfyUI" } }, { "id": 55, "type": "MarkdownNote", "pos": [ -925.6987758042737, -308.79051990273126 ], "size": [ 435.72716346592114, 545.840161067743 ], "flags": {}, "order": 2, "mode": 0, "inputs": [], "outputs": [], "properties": {}, "widgets_values": [ "## models\n\n> [black-forest-labs/flux-3-image-klein-comfy](https://huggingface.co/black-forest-labs/flux-3-image-klein-comfy) is gated. Log in to Hugging Face and accept the terms on the repository page before downloading.\n\n- diffusion_models\n - flux3_image_klein_sd_int8_convrot.safetensors (5.83 GB)\n- text_encoders\n - qwen3vl_4b_int8_convrot.safetensors (4.86 GB)\n - qwen3.5_9b_flux3_prompt_upsampler_int8_convrot.safetensors (9.75 GB, prompt upsampler / optional)\n- vae\n - flux2-vae.safetensors (336 MB)\n\n```text\nπŸ“‚ComfyUI/\n└── πŸ“‚models/\n β”œβ”€β”€ πŸ“‚diffusion_models/\n β”‚ └── flux3_image_klein_sd_int8_convrot.safetensors\n β”œβ”€β”€ πŸ“‚text_encoders/\n β”‚ β”œβ”€β”€ qwen3vl_4b_int8_convrot.safetensors\n β”‚ └── qwen3.5_9b_flux3_prompt_upsampler_int8_convrot.safetensors\n └── πŸ“‚vae/\n └── flux2-vae.safetensors\n```\n\n## guide\n\n- [ζ—₯本θͺž](https://comfyui.nomadoor.net/ja/basic-workflows/flux-3-image-klein/)\n- [English](https://comfyui.nomadoor.net/en/basic-workflows/flux-3-image-klein/)\n- [δΈ­ζ–‡](https://comfyui.nomadoor.net/zh/basic-workflows/flux-3-image-klein/)" ], "widgets_values_named": { "text": "## models\n\n> [black-forest-labs/flux-3-image-klein-comfy](https://huggingface.co/black-forest-labs/flux-3-image-klein-comfy) is gated. Log in to Hugging Face and accept the terms on the repository page before downloading.\n\n- diffusion_models\n - flux3_image_klein_sd_int8_convrot.safetensors (5.83 GB)\n- text_encoders\n - qwen3vl_4b_int8_convrot.safetensors (4.86 GB)\n - qwen3.5_9b_flux3_prompt_upsampler_int8_convrot.safetensors (9.75 GB, prompt upsampler / optional)\n- vae\n - flux2-vae.safetensors (336 MB)\n\n```text\nπŸ“‚ComfyUI/\n└── πŸ“‚models/\n β”œβ”€β”€ πŸ“‚diffusion_models/\n β”‚ └── flux3_image_klein_sd_int8_convrot.safetensors\n β”œβ”€β”€ πŸ“‚text_encoders/\n β”‚ β”œβ”€β”€ qwen3vl_4b_int8_convrot.safetensors\n β”‚ └── qwen3.5_9b_flux3_prompt_upsampler_int8_convrot.safetensors\n └── πŸ“‚vae/\n └── flux2-vae.safetensors\n```\n\n## guide\n\n- [ζ—₯本θͺž](https://comfyui.nomadoor.net/ja/basic-workflows/flux-3-image-klein/)\n- [English](https://comfyui.nomadoor.net/en/basic-workflows/flux-3-image-klein/)\n- [δΈ­ζ–‡](https://comfyui.nomadoor.net/zh/basic-workflows/flux-3-image-klein/)" }, "color": "#323", "bgcolor": "#535" }, { "id": 39, "type": "VAELoader", "pos": [ 519.2696075439453, 538.1055745563773 ], "size": [ 315, 58 ], "flags": {}, "order": 3, "mode": 0, "inputs": [], "outputs": [ { "name": "VAE", "type": "VAE", "slot_index": 0, "links": [ 76, 165 ] } ], "properties": { "Node name for S&R": "VAELoader", "cnr_id": "comfy-core", "ver": "0.3.33" }, "widgets_values": [ "flux2-vae.safetensors" ], "widgets_values_named": { "vae_name": "flux2-vae.safetensors" }, "color": "#322", "bgcolor": "#533" }, { "id": 63, "type": "ConditioningZeroOut", "pos": [ 609.2696075439453, 436.8646670878917 ], "size": [ 225, 26 ], "flags": {}, "order": 14, "mode": 0, "inputs": [ { "name": "conditioning", "type": "CONDITIONING", "link": 113 } ], "outputs": [ { "name": "CONDITIONING", "type": "CONDITIONING", "links": [ 161 ] } ], "properties": { "Node name for S&R": "ConditioningZeroOut" } }, { "id": 74, "type": "ReferenceMultiLatent", "pos": [ 876.8100510941658, 186.7020461700888 ], "size": [ 305.3782043457031, 158 ], "flags": {}, "order": 15, "mode": 0, "inputs": [ { "name": "positive", "type": "CONDITIONING", "link": 160 }, { "name": "negative", "shape": 7, "type": "CONDITIONING", "link": 161 }, { "name": "vae", "shape": 7, "type": "VAE", "link": 165 }, { "name": "images.image0", "shape": 7, "type": "IMAGE", "link": 210 }, { "name": "images.image1", "shape": 7, "type": "IMAGE", "link": null } ], "outputs": [ { "name": "positive", "type": "CONDITIONING", "links": [ 162 ] }, { "name": "negative", "type": "CONDITIONING", "links": [ 163 ] }, { "name": "empty_latent", "type": "LATENT", "links": [ 164 ] } ], "properties": { "Node name for S&R": "ReferenceMultiLatent" }, "widgets_values": [ "keep resolution" ], "widgets_values_named": { "resolution": "keep resolution" }, "color": "#232", "bgcolor": "#353" }, { "id": 75, "type": "LoadImage", "pos": [ -1419.5671439068783, 711.1398618720623 ], "size": [ 639.3421212561118, 554.6128815252974 ], "flags": {}, "order": 4, "mode": 0, "inputs": [], "outputs": [ { "name": "IMAGE", "type": "IMAGE", "links": [ 208, 219 ] }, { "name": "MASK", "type": "MASK", "links": null } ], "properties": { "Node name for S&R": "LoadImage" }, "widgets_values": [ "flux_3_image_klein_image_edit_bbox_input.webp", "image" ], "widgets_values_named": { "image": "flux_3_image_klein_image_edit_bbox_input.webp", "upload": "image" } }, { "id": 81, "type": "ResizeImageMaskNode", "pos": [ -717.3098839950081, 711.1398618720623 ], "size": [ 270, 106 ], "flags": {}, "order": 7, "mode": 0, "inputs": [ { "name": "input", "type": "IMAGE,MASK", "link": 208 } ], "outputs": [ { "name": "resized", "type": "IMAGE", "links": [ 178, 209 ] } ], "properties": { "Node name for S&R": "ResizeImageMaskNode" }, "widgets_values": [ "scale to multiple", 16, "bicubic" ], "widgets_values_named": { "resize_type": "scale to multiple", "resize_type.multiple": 16, "scale_method": "bicubic" } }, { "id": 83, "type": "TextGenerate", "pos": [ 21.683716116319715, -308.79051990273126 ], "size": [ 347.0577415637392, 288 ], "flags": {}, "order": 11, "mode": 0, "showAdvanced": false, "inputs": [ { "name": "clip", "type": "CLIP", "link": 177 }, { "name": "image", "shape": 7, "type": "IMAGE", "link": 178 }, { "name": "video", "shape": 7, "type": "IMAGE", "link": null }, { "name": "audio", "shape": 7, "type": "AUDIO", "link": null }, { "name": "system_prompt", "shape": 7, "type": "STRING", "link": 179 }, { "name": "prompt", "type": "STRING", "widget": { "name": "prompt" }, "link": 214 } ], "outputs": [ { "name": "generated_text", "type": "STRING", "links": [ 181, 217 ] }, { "name": "thinking", "type": "STRING", "links": null } ], "properties": { "Node name for S&R": "TextGenerate" }, "widgets_values": [ "change her hair to silver long straight", 3584, "off", false, true, "auto" ], "widgets_values_named": { "prompt": "change her hair to silver long straight", "max_length": 3584, "sampling_mode": "off", "thinking": false, "use_default_template": true, "mtp": "auto" }, "color": "#332922", "bgcolor": "#593930" }, { "id": 84, "type": "CLIPLoader", "pos": [ -330.64423306738433, -308.79051990273126 ], "size": [ 301.3524169921875, 106 ], "flags": {}, "order": 5, "mode": 0, "inputs": [], "outputs": [ { "name": "CLIP", "type": "CLIP", "links": [ 177 ] } ], "properties": { "Node name for S&R": "CLIPLoader" }, "widgets_values": [ "qwen3.5_9b_flux3_prompt_upsampler_int8_convrot.safetensors", "flux3", "default" ], "widgets_values_named": { "clip_name": "qwen3.5_9b_flux3_prompt_upsampler_int8_convrot.safetensors", "type": "flux3", "device": "default" }, "color": "#332922", "bgcolor": "#593930" }, { "id": 85, "type": "PrimitiveStringMultiline", "pos": [ -429.29181607519683, -142.33512540860602 ], "size": [ 400, 200 ], "flags": {}, "order": 6, "mode": 0, "inputs": [], "outputs": [ { "name": "STRING", "type": "STRING", "links": [ 179 ] } ], "properties": { "Node name for S&R": "PrimitiveStringMultiline" }, "widgets_values": [ "You are the annotation system an image-editing model was trained on. You see the source image ref_image_0 (and any further references ref_image_1, ref_image_2, ...) and the user's message: a short editing request in any language, followed by ONE JSON array of user boxes, each {\"id\", \"bbox\", \"desc\"} with bbox as [y_min, x_min, y_max, x_max] integers on a 0-1000 grid, top-left origin. The task is always an edit of ref_image_0: the image stays, and only the requested changes are made. Determine the edit precisely, imagine the edited result consistently with the source, and output on one line the instruction paragraph, one space, then ONE JSON array, the element table. Output nothing else.\n\nThe paragraph: English, imperative editing voice (\"Change the ...\", \"Remove the ...\", \"Add a ...\"); a whole-image transformation opens with its scope (\"Convert the entire image into ...\"). Never open with \"Create\", \"Generate\" or \"Redraw\", never use observer phrasing, and never re-describe the whole scene. Cite every table item at its first mention with its id in its own angle brackets, every table id exactly once and no other id; mention any other item by its appearance alone; cite a further reference by its token when content is taken from it (\"the cup from \"). The paragraph contains no coordinates or numbers taken from the boxes; describe places in words, and keep every statement about position, size and arrangement consistent with the boxes. State every end state concretely and close with one compact clause that everything else stays identical to the source; people keep their identity, face, clothing and accessories unless asked otherwise.\n\nPlanning: keep every explicit edit, constraint, count, style cue and piece of wording, and decide what is left open as one concrete plan. Ground everything in what is actually visible in the references: identify each referent by its real appearance and never mention content that is not there; if the request names a subject that is absent but a similar one is visible, edit that visible subject. Give new content enough appearance to be generated, at real-world size. Cover the knock-on effects (shadows, reflections, occlusion, perspective, scale, lighting) and say what fills any region the edit empties. Transcribe on-image text verbatim in its original language inside double quotes, with line breaks as \\n, and match the original font, size and color when replacing it. Use no evaluative or quality words. Typical paragraph length is 50 to 150 words.\n\nThe table: rows {\"id\", \"from\", \"src_bbox\", \"tgt_bbox\", \"desc\"}, one row per item or region the edit involves, and no rows for untouched scenery. from: \"ref_image_N\" for an item visible in reference N, \"target\" for newly created content. src_bbox: the item's box in its from-image, null for target rows. tgt_bbox: the item's box in the edited result, null when it is gone from the result. Boxes are [y_min, x_min, y_max, x_max] integers on the 0-1000 grid.\n\nUser boxes: each user box marks a place the user chose. Read the request and the box's desc to decide what the box refers to and whether it marks where an item is now, where it ends up, or both, then put its coordinates into the matching src_bbox and tgt_bbox fields, copied character for character; never move, resize or merge away the coordinates of a user box. When several user boxes refer to the same item, write one row for that item that carries each of their coordinates in the matching field, using the id of the box that marks the result. Every user box's coordinates appear in the table. Keep the user's ids. Any field no user box covers is read from the images as integers on the same grid. desc: one English paragraph of the item's static appearance (as seen in its from-image, or the decided appearance for target rows), opening with its identity; a user desc may be a label, a note or an instruction, so keep every attribute it states and move any action into the paragraph. People get the richest descs (ethnicity, gender, age range, skin tone, hair, face, garments, accessories, pose, expression).\n\nOutput only the paragraph and the table." ], "widgets_values_named": { "value": "You are the annotation system an image-editing model was trained on. You see the source image ref_image_0 (and any further references ref_image_1, ref_image_2, ...) and the user's message: a short editing request in any language, followed by ONE JSON array of user boxes, each {\"id\", \"bbox\", \"desc\"} with bbox as [y_min, x_min, y_max, x_max] integers on a 0-1000 grid, top-left origin. The task is always an edit of ref_image_0: the image stays, and only the requested changes are made. Determine the edit precisely, imagine the edited result consistently with the source, and output on one line the instruction paragraph, one space, then ONE JSON array, the element table. Output nothing else.\n\nThe paragraph: English, imperative editing voice (\"Change the ...\", \"Remove the ...\", \"Add a ...\"); a whole-image transformation opens with its scope (\"Convert the entire image into ...\"). Never open with \"Create\", \"Generate\" or \"Redraw\", never use observer phrasing, and never re-describe the whole scene. Cite every table item at its first mention with its id in its own angle brackets, every table id exactly once and no other id; mention any other item by its appearance alone; cite a further reference by its token when content is taken from it (\"the cup from \"). The paragraph contains no coordinates or numbers taken from the boxes; describe places in words, and keep every statement about position, size and arrangement consistent with the boxes. State every end state concretely and close with one compact clause that everything else stays identical to the source; people keep their identity, face, clothing and accessories unless asked otherwise.\n\nPlanning: keep every explicit edit, constraint, count, style cue and piece of wording, and decide what is left open as one concrete plan. Ground everything in what is actually visible in the references: identify each referent by its real appearance and never mention content that is not there; if the request names a subject that is absent but a similar one is visible, edit that visible subject. Give new content enough appearance to be generated, at real-world size. Cover the knock-on effects (shadows, reflections, occlusion, perspective, scale, lighting) and say what fills any region the edit empties. Transcribe on-image text verbatim in its original language inside double quotes, with line breaks as \\n, and match the original font, size and color when replacing it. Use no evaluative or quality words. Typical paragraph length is 50 to 150 words.\n\nThe table: rows {\"id\", \"from\", \"src_bbox\", \"tgt_bbox\", \"desc\"}, one row per item or region the edit involves, and no rows for untouched scenery. from: \"ref_image_N\" for an item visible in reference N, \"target\" for newly created content. src_bbox: the item's box in its from-image, null for target rows. tgt_bbox: the item's box in the edited result, null when it is gone from the result. Boxes are [y_min, x_min, y_max, x_max] integers on the 0-1000 grid.\n\nUser boxes: each user box marks a place the user chose. Read the request and the box's desc to decide what the box refers to and whether it marks where an item is now, where it ends up, or both, then put its coordinates into the matching src_bbox and tgt_bbox fields, copied character for character; never move, resize or merge away the coordinates of a user box. When several user boxes refer to the same item, write one row for that item that carries each of their coordinates in the matching field, using the id of the box that marks the result. Every user box's coordinates appear in the table. Keep the user's ids. Any field no user box covers is read from the images as integers on the same grid. desc: one English paragraph of the item's static appearance (as seen in its from-image, or the decided appearance for target rows), opening with its identity; a user desc may be a label, a note or an instruction, so keep every attribute it states and move any action into the paragraph. People get the richest descs (ethnicity, gender, age range, skin tone, hair, face, garments, accessories, pose, expression).\n\nOutput only the paragraph and the table." }, "color": "#332922", "bgcolor": "#593930" }, { "id": 86, "type": "PreviewAny", "pos": [ 414.7628296047744, -308.79051990273126 ], "size": [ 387.9444297800972, 321.71285303719196 ], "flags": {}, "order": 12, "mode": 0, "inputs": [ { "name": "source", "type": "*", "link": 181 } ], "outputs": [ { "name": "STRING", "type": "STRING", "links": null } ], "properties": { "Node name for S&R": "PreviewAny" }, "widgets_values": [], "widgets_values_named": {}, "color": "#332922", "bgcolor": "#593930" }, { "id": 95, "type": "Reroute", "pos": [ 759.2696075439453, 711.1398618720623 ], "size": [ 75, 26 ], "flags": {}, "order": 9, "mode": 0, "inputs": [ { "name": "", "type": "*", "link": 209 } ], "outputs": [ { "name": "", "type": "IMAGE", "links": [ 210 ] } ], "properties": { "showOutputText": false, "horizontal": false } }, { "id": 97, "type": "CreateBoundingBoxes", "pos": [ -718.3050904715104, 887.3065336478876 ], "size": [ 420, 560 ], "flags": {}, "order": 8, "mode": 0, "inputs": [ { "name": "background", "shape": 7, "type": "IMAGE", "link": 219 }, { "name": "bboxes", "shape": 7, "type": "BOUNDING_BOX,ARRAY,STRING", "link": null } ], "outputs": [ { "name": "preview", "type": "IMAGE", "links": null }, { "name": "bboxes", "type": "BOUNDING_BOX", "links": null }, { "name": "elements", "type": "ARRAY", "links": [ 213 ] } ], "properties": { "Node name for S&R": "CreateBoundingBoxes", "bboxGrid": false }, "widgets_values": [ 1408, 768, [ { "x": 616, "y": 60, "width": 550, "height": 425, "metadata": { "type": "obj", "text": "", "desc": "src", "palette": [] } }, { "x": 0, "y": 51, "width": 645, "height": 452, "metadata": { "type": "obj", "text": "", "desc": "tgt", "palette": [] } } ], [] ], "widgets_values_named": { "width": 1408, "height": 768, "editor_state": [ { "x": 616, "y": 60, "width": 550, "height": 425, "metadata": { "type": "obj", "text": "", "desc": "src", "palette": [] } }, { "x": 0, "y": 51, "width": 645, "height": 452, "metadata": { "type": "obj", "text": "", "desc": "tgt", "palette": [] } } ], "last_incoming": [] } }, { "id": 100, "type": "Flux3LayoutPrompt", "pos": [ -251.84675514367922, 927.2813900064144 ], "size": [ 306.9057121824424, 154.8260928090063 ], "flags": {}, "order": 10, "mode": 0, "inputs": [ { "name": "element", "type": "ARRAY", "link": 213 } ], "outputs": [ { "name": "prompt", "type": "STRING", "links": [ 214 ] } ], "properties": { "Node name for S&R": "Flux3LayoutPrompt" }, "widgets_values": [ "Move the bird from src to tgt." ], "widgets_values_named": { "prompt": "Move the bird from src to tgt." } } ], "links": [ [ 35, 3, 0, 8, 0, "LATENT" ], [ 74, 38, 0, 6, 0, "CLIP" ], [ 76, 39, 0, 8, 1, "VAE" ], [ 101, 8, 0, 56, 0, "IMAGE" ], [ 105, 37, 0, 3, 0, "MODEL" ], [ 113, 6, 0, 63, 0, "CONDITIONING" ], [ 162, 74, 0, 3, 1, "CONDITIONING" ], [ 163, 74, 1, 3, 2, "CONDITIONING" ], [ 164, 74, 2, 3, 3, "LATENT" ], [ 181, 83, 0, 86, 0, "STRING" ], [ 209, 81, 0, 95, 0, "IMAGE" ], [ 213, 97, 2, 100, 0, "ARRAY" ], [ 217, 83, 0, 6, 1, "STRING" ], [ 219, 75, 0, 97, 0, "IMAGE" ], [ 160, 6, 0, 74, 0, "CONDITIONING" ], [ 161, 63, 0, 74, 1, "CONDITIONING" ], [ 165, 39, 0, 74, 2, "VAE" ], [ 210, 95, 0, 74, 3, "IMAGE" ], [ 208, 75, 0, 81, 0, "IMAGE" ], [ 177, 84, 0, 83, 0, "CLIP" ], [ 178, 81, 0, 83, 1, "IMAGE" ], [ 179, 85, 0, 83, 4, "STRING" ], [ 214, 100, 0, 83, 5, "STRING" ] ], "groups": [], "config": {}, "extra": { "ds": { "scale": 0.5209868481924903, "offset": [ 1873.938948650083, 676.6682095353456 ] }, "frontendVersion": "1.55.17", "VHS_latentpreview": false, "VHS_latentpreviewrate": 0, "VHS_MetadataImage": true, "VHS_KeepIntermediate": true }, "version": 0.4 }