├── .github └── workflows │ └── publish.yml ├── .gitignore ├── LICENSE ├── README.md ├── __init__.py ├── examples ├── infinite_you_workflow.json ├── multi_id_infinite_you_workflow.json ├── multi_id_workflow.jpg ├── teaser.jpg └── workflow.jpg ├── infuse_net.py ├── nodes.py ├── pyproject.toml ├── requirements.txt ├── resampler.py └── utils.py /.github/workflows/publish.yml: -------------------------------------------------------------------------------- 1 | name: Publish to Comfy registry 2 | on: 3 | workflow_dispatch: 4 | push: 5 | branches: 6 | - main 7 | - master 8 | paths: 9 | - "pyproject.toml" 10 | 11 | permissions: 12 | issues: write 13 | 14 | jobs: 15 | publish-node: 16 | name: Publish Custom Node to registry 17 | runs-on: ubuntu-latest 18 | if: ${{ github.repository_owner == 'bytedance' }} 19 | steps: 20 | - name: Check out code 21 | uses: actions/checkout@v4 22 | with: 23 | submodules: true 24 | - name: Publish Custom Node 25 | uses: Comfy-Org/publish-node-action@v1 26 | with: 27 | ## Add your own personal access token to your Github Repository secrets and reference it here. 28 | personal_access_token: ${{ secrets.REGISTRY_ACCESS_TOKEN }} 29 | -------------------------------------------------------------------------------- /.gitignore: -------------------------------------------------------------------------------- 1 | .DS_Store 2 | .idea 3 | .ipynb_checkpoints 4 | .gradio 5 | *.swp 6 | *.pyc 7 | __pycache__ 8 | *.tar* 9 | *.zip 10 | *.pkl 11 | *.pyc 12 | *.bak 13 | *.png 14 | *.deb 15 | 16 | .isort.cfg 17 | .pre-commit-config.yaml 18 | 19 | dataset_stats 20 | debug* 21 | locks 22 | checkpoints 23 | pretrained_checkpoint 24 | ./models 25 | models 26 | results 27 | wandb 28 | tmp* 29 | env* 30 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | Apache License 2 | Version 2.0, January 2004 3 | http://www.apache.org/licenses/ 4 | 5 | TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 6 | 7 | 1. Definitions. 8 | 9 | "License" shall mean the terms and conditions for use, reproduction, 10 | and distribution as defined by Sections 1 through 9 of this document. 11 | 12 | "Licensor" shall mean the copyright owner or entity authorized by 13 | the copyright owner that is granting the License. 14 | 15 | "Legal Entity" shall mean the union of the acting entity and all 16 | other entities that control, are controlled by, or are under common 17 | control with that entity. For the purposes of this definition, 18 | "control" means (i) the power, direct or indirect, to cause the 19 | direction or management of such entity, whether by contract or 20 | otherwise, or (ii) ownership of fifty percent (50%) or more of the 21 | outstanding shares, or (iii) beneficial ownership of such entity. 22 | 23 | "You" (or "Your") shall mean an individual or Legal Entity 24 | exercising permissions granted by this License. 25 | 26 | "Source" form shall mean the preferred form for making modifications, 27 | including but not limited to software source code, documentation 28 | source, and configuration files. 29 | 30 | "Object" form shall mean any form resulting from mechanical 31 | transformation or translation of a Source form, including but 32 | not limited to compiled object code, generated documentation, 33 | and conversions to other media types. 34 | 35 | "Work" shall mean the work of authorship, whether in Source or 36 | Object form, made available under the License, as indicated by a 37 | copyright notice that is included in or attached to the work 38 | (an example is provided in the Appendix below). 39 | 40 | "Derivative Works" shall mean any work, whether in Source or Object 41 | form, that is based on (or derived from) the Work and for which the 42 | editorial revisions, annotations, elaborations, or other modifications 43 | represent, as a whole, an original work of authorship. For the purposes 44 | of this License, Derivative Works shall not include works that remain 45 | separable from, or merely link (or bind by name) to the interfaces of, 46 | the Work and Derivative Works thereof. 47 | 48 | "Contribution" shall mean any work of authorship, including 49 | the original version of the Work and any modifications or additions 50 | to that Work or Derivative Works thereof, that is intentionally 51 | submitted to Licensor for inclusion in the Work by the copyright owner 52 | or by an individual or Legal Entity authorized to submit on behalf of 53 | the copyright owner. For the purposes of this definition, "submitted" 54 | means any form of electronic, verbal, or written communication sent 55 | to the Licensor or its representatives, including but not limited to 56 | communication on electronic mailing lists, source code control systems, 57 | and issue tracking systems that are managed by, or on behalf of, the 58 | Licensor for the purpose of discussing and improving the Work, but 59 | excluding communication that is conspicuously marked or otherwise 60 | designated in writing by the copyright owner as "Not a Contribution." 61 | 62 | "Contributor" shall mean Licensor and any individual or Legal Entity 63 | on behalf of whom a Contribution has been received by Licensor and 64 | subsequently incorporated within the Work. 65 | 66 | 2. Grant of Copyright License. Subject to the terms and conditions of 67 | this License, each Contributor hereby grants to You a perpetual, 68 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 69 | copyright license to reproduce, prepare Derivative Works of, 70 | publicly display, publicly perform, sublicense, and distribute the 71 | Work and such Derivative Works in Source or Object form. 72 | 73 | 3. Grant of Patent License. Subject to the terms and conditions of 74 | this License, each Contributor hereby grants to You a perpetual, 75 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 76 | (except as stated in this section) patent license to make, have made, 77 | use, offer to sell, sell, import, and otherwise transfer the Work, 78 | where such license applies only to those patent claims licensable 79 | by such Contributor that are necessarily infringed by their 80 | Contribution(s) alone or by combination of their Contribution(s) 81 | with the Work to which such Contribution(s) was submitted. If You 82 | institute patent litigation against any entity (including a 83 | cross-claim or counterclaim in a lawsuit) alleging that the Work 84 | or a Contribution incorporated within the Work constitutes direct 85 | or contributory patent infringement, then any patent licenses 86 | granted to You under this License for that Work shall terminate 87 | as of the date such litigation is filed. 88 | 89 | 4. Redistribution. You may reproduce and distribute copies of the 90 | Work or Derivative Works thereof in any medium, with or without 91 | modifications, and in Source or Object form, provided that You 92 | meet the following conditions: 93 | 94 | (a) You must give any other recipients of the Work or 95 | Derivative Works a copy of this License; and 96 | 97 | (b) You must cause any modified files to carry prominent notices 98 | stating that You changed the files; and 99 | 100 | (c) You must retain, in the Source form of any Derivative Works 101 | that You distribute, all copyright, patent, trademark, and 102 | attribution notices from the Source form of the Work, 103 | excluding those notices that do not pertain to any part of 104 | the Derivative Works; and 105 | 106 | (d) If the Work includes a "NOTICE" text file as part of its 107 | distribution, then any Derivative Works that You distribute must 108 | include a readable copy of the attribution notices contained 109 | within such NOTICE file, excluding those notices that do not 110 | pertain to any part of the Derivative Works, in at least one 111 | of the following places: within a NOTICE text file distributed 112 | as part of the Derivative Works; within the Source form or 113 | documentation, if provided along with the Derivative Works; or, 114 | within a display generated by the Derivative Works, if and 115 | wherever such third-party notices normally appear. The contents 116 | of the NOTICE file are for informational purposes only and 117 | do not modify the License. You may add Your own attribution 118 | notices within Derivative Works that You distribute, alongside 119 | or as an addendum to the NOTICE text from the Work, provided 120 | that such additional attribution notices cannot be construed 121 | as modifying the License. 122 | 123 | You may add Your own copyright statement to Your modifications and 124 | may provide additional or different license terms and conditions 125 | for use, reproduction, or distribution of Your modifications, or 126 | for any such Derivative Works as a whole, provided Your use, 127 | reproduction, and distribution of the Work otherwise complies with 128 | the conditions stated in this License. 129 | 130 | 5. Submission of Contributions. Unless You explicitly state otherwise, 131 | any Contribution intentionally submitted for inclusion in the Work 132 | by You to the Licensor shall be under the terms and conditions of 133 | this License, without any additional terms or conditions. 134 | Notwithstanding the above, nothing herein shall supersede or modify 135 | the terms of any separate license agreement you may have executed 136 | with Licensor regarding such Contributions. 137 | 138 | 6. Trademarks. This License does not grant permission to use the trade 139 | names, trademarks, service marks, or product names of the Licensor, 140 | except as required for reasonable and customary use in describing the 141 | origin of the Work and reproducing the content of the NOTICE file. 142 | 143 | 7. Disclaimer of Warranty. Unless required by applicable law or 144 | agreed to in writing, Licensor provides the Work (and each 145 | Contributor provides its Contributions) on an "AS IS" BASIS, 146 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or 147 | implied, including, without limitation, any warranties or conditions 148 | of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A 149 | PARTICULAR PURPOSE. You are solely responsible for determining the 150 | appropriateness of using or redistributing the Work and assume any 151 | risks associated with Your exercise of permissions under this License. 152 | 153 | 8. Limitation of Liability. In no event and under no legal theory, 154 | whether in tort (including negligence), contract, or otherwise, 155 | unless required by applicable law (such as deliberate and grossly 156 | negligent acts) or agreed to in writing, shall any Contributor be 157 | liable to You for damages, including any direct, indirect, special, 158 | incidental, or consequential damages of any character arising as a 159 | result of this License or out of the use or inability to use the 160 | Work (including but not limited to damages for loss of goodwill, 161 | work stoppage, computer failure or malfunction, or any and all 162 | other commercial damages or losses), even if such Contributor 163 | has been advised of the possibility of such damages. 164 | 165 | 9. Accepting Warranty or Additional Liability. While redistributing 166 | the Work or Derivative Works thereof, You may choose to offer, 167 | and charge a fee for, acceptance of support, warranty, indemnity, 168 | or other liability obligations and/or rights consistent with this 169 | License. However, in accepting such obligations, You may act only 170 | on Your own behalf and on Your sole responsibility, not on behalf 171 | of any other Contributor, and only if You agree to indemnify, 172 | defend, and hold each Contributor harmless for any liability 173 | incurred by, or claims asserted against, such Contributor by reason 174 | of your accepting any such warranty or additional liability. 175 | 176 | END OF TERMS AND CONDITIONS 177 | 178 | APPENDIX: How to apply the Apache License to your work. 179 | 180 | To apply the Apache License to your work, attach the following 181 | boilerplate notice, with the fields enclosed by brackets "[]" 182 | replaced with your own identifying information. (Don't include 183 | the brackets!) The text should be enclosed in the appropriate 184 | comment syntax for the file format. We also recommend that a 185 | file or class name and description of purpose be included on the 186 | same "printed page" as the copyright notice for easier 187 | identification within third-party archives. 188 | 189 | Copyright [yyyy] [name of copyright owner] 190 | 191 | Licensed under the Apache License, Version 2.0 (the "License"); 192 | you may not use this file except in compliance with the License. 193 | You may obtain a copy of the License at 194 | 195 | http://www.apache.org/licenses/LICENSE-2.0 196 | 197 | Unless required by applicable law or agreed to in writing, software 198 | distributed under the License is distributed on an "AS IS" BASIS, 199 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 200 | See the License for the specific language governing permissions and 201 | limitations under the License. -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | ## Official ComfyUI Support - InfiniteYou: Flexible Photo Recrafting While Preserving Your Identity 2 | 3 |
4 | 5 |   6 |   7 |   8 |   9 |   10 | 11 |
12 | 13 | This repository provides the official ComfyUI native node for [**InfiniteYou**](https://github.com/bytedance/InfiniteYou) with FLUX. 14 | 15 | ![teaser](examples/teaser.jpg) 16 | 17 |
18 | Abstract (click to expand) 19 | 20 | > *Achieving flexible and high-fidelity identity-preserved image generation remains formidable, particularly with advanced Diffusion Transformers (DiTs) like FLUX. We introduce **InfiniteYou (InfU)**, one of the earliest robust frameworks leveraging DiTs for this task. InfU addresses significant issues of existing methods, such as insufficient identity similarity, poor text-image alignment, and low generation quality and aesthetics. Central to InfU is InfuseNet, a component that injects identity features into the DiT base model via residual connections, enhancing identity similarity while maintaining generation capabilities. A multi-stage training strategy, including pretraining and supervised fine-tuning (SFT) with synthetic single-person-multiple-sample (SPMS) data, further improves text-image alignment, ameliorates image quality, and alleviates face copy-pasting. Extensive experiments demonstrate that InfU achieves state-of-the-art performance, surpassing existing baselines. In addition, the plug-and-play design of InfU ensures compatibility with various existing methods, offering a valuable contribution to the broader community.* 21 | 22 |
23 | 24 | 25 | ## 🛠️ Workflow Example 26 | 27 | This node adds InfiniteYou‑FLUX support to ComfyUI. In [infinite_you_workflow.json](./examples/infinite_you_workflow.json), you can find a simple workflow demonstrating its usage with either an empty face‑pose control image or a real face‑pose control image, configured to run using FLUX FP8 precision. It also shows an example running the node with FLUX.1-schnell. 28 | 29 | ![basic workflow](examples/workflow.jpg) 30 | 31 | **Extension:** We also provide an example [multi-id workflow](examples/multi_id_infinite_you_workflow.json) for identity-preserved image generation of two people. This uses the masked multi-region test of single-ID InfiniteYou‑FLUX models with masked residual blending, and is provided for reference only. 32 | 33 | ![basic workflow](examples/multi_id_workflow.jpg) 34 | 35 | 36 | ## 🔧 Requirements and Installation 37 | 38 | ### Dependencies 39 | 40 | 1. Install [ComfyUI](https://github.com/comfyanonymous/ComfyUI?tab=readme-ov-file#get-started). 41 | 42 | 2. Clone this repo under `ComfyUI/custom_nodes` and install the dependencies: 43 | ``` 44 | cd ComfyUI/custom_nodes 45 | git clone https://github.com/bytedance/ComfyUI_InfiniteYou.git 46 | 47 | cd ComfyUI_InfiniteYou 48 | pip install -r requirements.txt 49 | ``` 50 | 51 | * Our InfiniteYou node has been added to the official Comfy Registry to ease installation: https://registry.comfy.org/publishers/yuminjia/nodes/infiniteyou. Therefore, you can also search `ComfyUI_InfiniteYou` in the ComfyUI Node Manager to install this official node. 52 | 53 | 54 | ### Memory Requirements 55 | 56 | The full-performance BF16 model inference requires a peak VRAM of around **43GB**. Running with FP8 precision requires a peak VRAM of around **24GB**. 57 | 58 | 59 | ## 💡 Usage 60 | 61 | 1. Restart ComfyUI. 62 | 63 | 2. Import the [workflow](examples/infinite_you_workflow.json) from the [examples folder](./examples). Please use [multi-id workflow](examples/multi_id_infinite_you_workflow.json) if needed. 64 | 65 | * Some [important usage tips](https://github.com/bytedance/InfiniteYou?tab=readme-ov-file#-important-usage-tips) can be found in our main InfiniteYou repository. 66 | 67 | 68 | ## 🏰 Required Models 69 | 70 | ### InfiniteYou and InsightFace Detection Models 71 | 72 | This node will automatically download the following models at runtime if they not exists. Alternatively, you may download them manually into the following locations. For InfiniteYou, you need at least `image_proj_model.bin` and `infusenet_.*safetensors` of the corresponding model versions. 73 | 74 | | Model | Location | 75 | | ---- | ---- | 76 | | [InfiniteYou](https://huggingface.co/ByteDance/InfiniteYou/tree/main/infu_flux_v1.0) | `ComfyUI/models/infinite_you/` | 77 | | [InsightFace AntelopeV2](https://huggingface.co/ByteDance/InfiniteYou/tree/main/supports/insightface/models/antelopev2) | `ComfyUI/models/insightface/models/antelopev2` | 78 | 79 | You may follow [ComfyUI FLUX examples](https://comfyanonymous.github.io/ComfyUI_examples/flux/) to download FLUX and other models for full-performance inference. 80 | 81 | 82 | ### Other Required Models for Running FP8 Precision 83 | 84 | The FP8 InfiniteYou model can also be downloaded automatically or manually as above. To run with FP8 precision, other models need to be downloaded manually into the following locations. 85 | 86 | | Model | Location | 87 | | ---- | ---- | 88 | | [FLUX FP8](https://huggingface.co/Kijai/flux-fp8/tree/main) | `ComfyUI/models/diffusion_models` | 89 | | [FLUX VAE](https://huggingface.co/black-forest-labs/FLUX.1-schnell/blob/main/ae.safetensors) | `ComfyUI/models/vae` | 90 | | [Text Encoders FP8](https://huggingface.co/comfyanonymous/flux_text_encoders/tree/main) | `ComfyUI/models/text_encoders` | 91 | 92 | 93 | ## 📜 Disclaimer and Licenses 94 | 95 | The images used in this repository and related demos are sourced from consented subjects or generated by the models. These pictures are intended solely to showcase the capabilities of our research. If you have any concerns, please feel free to contact us, and we will promptly remove any inappropriate content. 96 | 97 | The use of the released code, model, and demo must strictly adhere to the respective licenses. Our code is released under the [Apache License 2.0](./LICENSE), and our model is released under the [Creative Commons Attribution-NonCommercial 4.0 International Public License](https://huggingface.co/ByteDance/InfiniteYou/blob/main/LICENSE) for academic research purposes only. Any manual or automatic downloading of the face models from [InsightFace](https://github.com/deepinsight/insightface), the [FLUX.1-dev](https://huggingface.co/black-forest-labs/FLUX.1-dev) base model, LoRAs ([Realism](https://civitai.com/models/631986?modelVersionId=706528) and [Anti-blur](https://civitai.com/models/675581/anti-blur-flux-lora)), *etc.*, must follow their original licenses and be used only for academic research purposes. 98 | 99 | This research aims to positively impact the field of Generative AI. Any usage of this method must be responsible and comply with local laws. The developers do not assume any responsibility for any potential misuse. 100 | 101 | 102 | ## 📖 Citation 103 | 104 | If you find InfiniteYou useful for your research or applications, please cite our paper: 105 | 106 | ```bibtex 107 | @article{jiang2025infiniteyou, 108 | title={{InfiniteYou}: Flexible Photo Recrafting While Preserving Your Identity}, 109 | author={Jiang, Liming and Yan, Qing and Jia, Yumin and Liu, Zichuan and Kang, Hao and Lu, Xin}, 110 | journal={arXiv preprint}, 111 | volume={arXiv:2503.16418}, 112 | year={2025} 113 | } 114 | ``` 115 | 116 | We also appreciate it if you could give a star :star: to this repository and our [main repository](https://github.com/bytedance/InfiniteYou). Thanks a lot! 117 | -------------------------------------------------------------------------------- /__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) 2025 Bytedance Ltd. and/or its affiliates. All rights reserved. 2 | 3 | # Licensed under the Apache License, Version 2.0 (the "License"); 4 | # you may not use this file except in compliance with the License. 5 | # You may obtain a copy of the License at 6 | 7 | # http://www.apache.org/licenses/LICENSE-2.0 8 | 9 | # Unless required by applicable law or agreed to in writing, software 10 | # distributed under the License is distributed on an "AS IS" BASIS, 11 | # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 12 | # See the License for the specific language governing permissions and 13 | # limitations under the License. 14 | 15 | from .nodes import * 16 | 17 | __all__ = ["NODE_CLASS_MAPPINGS", "NODE_DISPLAY_NAME_MAPPINGS"] 18 | -------------------------------------------------------------------------------- /examples/infinite_you_workflow.json: -------------------------------------------------------------------------------- 1 | { 2 | "id": "c1a3d0d1-f0cb-4369-83ae-273696df248c", 3 | "revision": 0, 4 | "last_node_id": 116, 5 | "last_link_id": 185, 6 | "nodes": [ 7 | { 8 | "id": 52, 9 | "type": "CLIPTextEncode", 10 | "pos": [ 11 | -568.8687133789062, 12 | 687.748046875 13 | ], 14 | "size": [ 15 | 397.7880859375, 16 | 150.98748779296875 17 | ], 18 | "flags": { 19 | "collapsed": true 20 | }, 21 | "order": 11, 22 | "mode": 0, 23 | "inputs": [ 24 | { 25 | "name": "clip", 26 | "type": "CLIP", 27 | "link": 94 28 | } 29 | ], 30 | "outputs": [ 31 | { 32 | "name": "CONDITIONING", 33 | "shape": 3, 34 | "type": "CONDITIONING", 35 | "slot_index": 0, 36 | "links": [ 37 | 65, 38 | 132, 39 | 159 40 | ] 41 | } 42 | ], 43 | "title": "CLIP Text Encode (EMPTY)", 44 | "properties": { 45 | "Node name for S&R": "CLIPTextEncode" 46 | }, 47 | "widgets_values": [ 48 | "" 49 | ] 50 | }, 51 | { 52 | "id": 110, 53 | "type": "ExtractIDEmbedding", 54 | "pos": [ 55 | -565.7846069335938, 56 | 778.5155639648438 57 | ], 58 | "size": [ 59 | 367.79998779296875, 60 | 86 61 | ], 62 | "flags": {}, 63 | "order": 13, 64 | "mode": 0, 65 | "inputs": [ 66 | { 67 | "name": "face_detector", 68 | "type": "MODEL", 69 | "link": 172 70 | }, 71 | { 72 | "name": "arcface_model", 73 | "type": "MODEL", 74 | "link": 173 75 | }, 76 | { 77 | "name": "image_proj_model", 78 | "type": "MODEL", 79 | "link": 174 80 | }, 81 | { 82 | "name": "image", 83 | "type": "IMAGE", 84 | "link": 175 85 | } 86 | ], 87 | "outputs": [ 88 | { 89 | "name": "CONDITIONING", 90 | "type": "CONDITIONING", 91 | "links": [ 92 | 176 93 | ] 94 | } 95 | ], 96 | "properties": { 97 | "Node name for S&R": "ExtractIDEmbedding" 98 | }, 99 | "widgets_values": [] 100 | }, 101 | { 102 | "id": 45, 103 | "type": "SaveImage", 104 | "pos": [ 105 | 668.8138427734375, 106 | 288.35491943359375 107 | ], 108 | "size": [ 109 | 688.4625244140625, 110 | 863.3844604492188 111 | ], 112 | "flags": { 113 | "collapsed": false 114 | }, 115 | "order": 15, 116 | "mode": 0, 117 | "inputs": [ 118 | { 119 | "name": "images", 120 | "type": "IMAGE", 121 | "link": 60 122 | } 123 | ], 124 | "outputs": [], 125 | "properties": { 126 | "Node name for S&R": "SaveImage" 127 | }, 128 | "widgets_values": [ 129 | "ComfyUI", 130 | "" 131 | ] 132 | }, 133 | { 134 | "id": 44, 135 | "type": "VAEDecode", 136 | "pos": [ 137 | 315.2570495605469, 138 | 300.6652526855469 139 | ], 140 | "size": [ 141 | 210, 142 | 46 143 | ], 144 | "flags": { 145 | "collapsed": true 146 | }, 147 | "order": 14, 148 | "mode": 0, 149 | "inputs": [ 150 | { 151 | "name": "samples", 152 | "type": "LATENT", 153 | "link": 58 154 | }, 155 | { 156 | "name": "vae", 157 | "type": "VAE", 158 | "link": 59 159 | } 160 | ], 161 | "outputs": [ 162 | { 163 | "name": "IMAGE", 164 | "type": "IMAGE", 165 | "slot_index": 0, 166 | "links": [ 167 | 60 168 | ] 169 | } 170 | ], 171 | "properties": { 172 | "Node name for S&R": "VAEDecode" 173 | }, 174 | "widgets_values": [] 175 | }, 176 | { 177 | "id": 109, 178 | "type": "InfuseNetApply", 179 | "pos": [ 180 | 278.14422607421875, 181 | 803.1834716796875 182 | ], 183 | "size": [ 184 | 315, 185 | 206 186 | ], 187 | "flags": {}, 188 | "order": 17, 189 | "mode": 0, 190 | "inputs": [ 191 | { 192 | "name": "positive", 193 | "type": "CONDITIONING", 194 | "link": 169 195 | }, 196 | { 197 | "name": "negative", 198 | "type": "CONDITIONING", 199 | "link": 159 200 | }, 201 | { 202 | "name": "id_embedding", 203 | "type": "CONDITIONING", 204 | "link": 176 205 | }, 206 | { 207 | "name": "control_net", 208 | "type": "CONTROL_NET", 209 | "link": 166 210 | }, 211 | { 212 | "name": "image", 213 | "shape": 7, 214 | "type": "IMAGE", 215 | "link": 180 216 | }, 217 | { 218 | "name": "vae", 219 | "shape": 7, 220 | "type": "VAE", 221 | "link": 170 222 | } 223 | ], 224 | "outputs": [ 225 | { 226 | "name": "positive", 227 | "type": "CONDITIONING", 228 | "links": [ 229 | 163 230 | ] 231 | }, 232 | { 233 | "name": "negative", 234 | "type": "CONDITIONING", 235 | "links": [ 236 | 164 237 | ] 238 | } 239 | ], 240 | "properties": { 241 | "Node name for S&R": "InfuseNetApply" 242 | }, 243 | "widgets_values": [ 244 | 1, 245 | 0, 246 | 1 247 | ] 248 | }, 249 | { 250 | "id": 43, 251 | "type": "EmptyLatentImage", 252 | "pos": [ 253 | -138.02438354492188, 254 | 463.6690368652344 255 | ], 256 | "size": [ 257 | 242.7444610595703, 258 | 106 259 | ], 260 | "flags": { 261 | "collapsed": false 262 | }, 263 | "order": 0, 264 | "mode": 0, 265 | "inputs": [], 266 | "outputs": [ 267 | { 268 | "name": "LATENT", 269 | "type": "LATENT", 270 | "slot_index": 0, 271 | "links": [ 272 | 66 273 | ] 274 | } 275 | ], 276 | "properties": { 277 | "Node name for S&R": "EmptyLatentImage" 278 | }, 279 | "widgets_values": [ 280 | 864, 281 | 1152, 282 | 1 283 | ] 284 | }, 285 | { 286 | "id": 92, 287 | "type": "PrimitiveNode", 288 | "pos": [ 289 | -140.0155487060547, 290 | 623.6337890625 291 | ], 292 | "size": [ 293 | 268.72747802734375, 294 | 82 295 | ], 296 | "flags": { 297 | "collapsed": false 298 | }, 299 | "order": 1, 300 | "mode": 0, 301 | "inputs": [], 302 | "outputs": [ 303 | { 304 | "name": "INT", 305 | "type": "INT", 306 | "widget": { 307 | "name": "seed" 308 | }, 309 | "links": [ 310 | 127 311 | ] 312 | } 313 | ], 314 | "title": "seed", 315 | "properties": { 316 | "Run widget replace on values": false 317 | }, 318 | "widgets_values": [ 319 | 1117450117680779, 320 | "randomize" 321 | ] 322 | }, 323 | { 324 | "id": 106, 325 | "type": "ExtractFacePoseImage", 326 | "pos": [ 327 | -125.12596130371094, 328 | 1233.2313232421875 329 | ], 330 | "size": [ 331 | 315, 332 | 102 333 | ], 334 | "flags": {}, 335 | "order": 12, 336 | "mode": 0, 337 | "inputs": [ 338 | { 339 | "name": "face_detector", 340 | "type": "MODEL", 341 | "link": 150 342 | }, 343 | { 344 | "name": "image", 345 | "type": "IMAGE", 346 | "link": 151 347 | } 348 | ], 349 | "outputs": [ 350 | { 351 | "name": "IMAGE", 352 | "type": "IMAGE", 353 | "links": [ 354 | 156 355 | ] 356 | } 357 | ], 358 | "properties": { 359 | "Node name for S&R": "ExtractFacePoseImage" 360 | }, 361 | "widgets_values": [ 362 | 864, 363 | 1152 364 | ] 365 | }, 366 | { 367 | "id": 105, 368 | "type": "EmptyImage", 369 | "pos": [ 370 | -125.57537841796875, 371 | 1033.876220703125 372 | ], 373 | "size": [ 374 | 315, 375 | 130 376 | ], 377 | "flags": { 378 | "collapsed": false 379 | }, 380 | "order": 2, 381 | "mode": 0, 382 | "inputs": [], 383 | "outputs": [ 384 | { 385 | "name": "IMAGE", 386 | "type": "IMAGE", 387 | "links": [ 388 | 178 389 | ] 390 | } 391 | ], 392 | "title": "Empty Pose Image", 393 | "properties": { 394 | "Node name for S&R": "EmptyImage" 395 | }, 396 | "widgets_values": [ 397 | 864, 398 | 1152, 399 | 1, 400 | 0 401 | ] 402 | }, 403 | { 404 | "id": 46, 405 | "type": "VAELoader", 406 | "pos": [ 407 | -128.3144073486328, 408 | 269.4505615234375 409 | ], 410 | "size": [ 411 | 210, 412 | 58 413 | ], 414 | "flags": {}, 415 | "order": 3, 416 | "mode": 0, 417 | "inputs": [], 418 | "outputs": [ 419 | { 420 | "name": "VAE", 421 | "shape": 3, 422 | "type": "VAE", 423 | "slot_index": 0, 424 | "links": [ 425 | 59, 426 | 137, 427 | 161, 428 | 170 429 | ] 430 | } 431 | ], 432 | "properties": { 433 | "Node name for S&R": "VAELoader" 434 | }, 435 | "widgets_values": [ 436 | "ae.safetensors" 437 | ] 438 | }, 439 | { 440 | "id": 51, 441 | "type": "KSampler", 442 | "pos": [ 443 | 278.04345703125, 444 | 418.0480041503906 445 | ], 446 | "size": [ 447 | 210, 448 | 238 449 | ], 450 | "flags": {}, 451 | "order": 16, 452 | "mode": 0, 453 | "inputs": [ 454 | { 455 | "name": "model", 456 | "type": "MODEL", 457 | "link": 185 458 | }, 459 | { 460 | "name": "positive", 461 | "type": "CONDITIONING", 462 | "link": 163 463 | }, 464 | { 465 | "name": "negative", 466 | "type": "CONDITIONING", 467 | "link": 164 468 | }, 469 | { 470 | "name": "latent_image", 471 | "type": "LATENT", 472 | "link": 66 473 | }, 474 | { 475 | "name": "seed", 476 | "type": "INT", 477 | "widget": { 478 | "name": "seed" 479 | }, 480 | "link": 127 481 | } 482 | ], 483 | "outputs": [ 484 | { 485 | "name": "LATENT", 486 | "shape": 3, 487 | "type": "LATENT", 488 | "slot_index": 0, 489 | "links": [ 490 | 58 491 | ] 492 | } 493 | ], 494 | "properties": { 495 | "Node name for S&R": "KSampler" 496 | }, 497 | "widgets_values": [ 498 | 1117450117680779, 499 | "fixed", 500 | 4, 501 | 1, 502 | "euler", 503 | "beta", 504 | 1 505 | ] 506 | }, 507 | { 508 | "id": 48, 509 | "type": "DualCLIPLoader", 510 | "pos": [ 511 | -987.113525390625, 512 | 476.66717529296875 513 | ], 514 | "size": [ 515 | 356.2766418457031, 516 | 130 517 | ], 518 | "flags": {}, 519 | "order": 4, 520 | "mode": 0, 521 | "inputs": [], 522 | "outputs": [ 523 | { 524 | "name": "CLIP", 525 | "shape": 3, 526 | "type": "CLIP", 527 | "slot_index": 0, 528 | "links": [ 529 | 93, 530 | 94 531 | ] 532 | } 533 | ], 534 | "properties": { 535 | "Node name for S&R": "DualCLIPLoader" 536 | }, 537 | "widgets_values": [ 538 | "t5xxl_fp8_e4m3fn.safetensors", 539 | "clip_l.safetensors", 540 | "flux", 541 | "default" 542 | ] 543 | }, 544 | { 545 | "id": 108, 546 | "type": "InfuseNetLoader", 547 | "pos": [ 548 | -128.15025329589844, 549 | 920.6913452148438 550 | ], 551 | "size": [ 552 | 315, 553 | 58 554 | ], 555 | "flags": {}, 556 | "order": 5, 557 | "mode": 0, 558 | "inputs": [], 559 | "outputs": [ 560 | { 561 | "name": "CONTROL_NET", 562 | "type": "CONTROL_NET", 563 | "links": [ 564 | 157, 565 | 166 566 | ] 567 | } 568 | ], 569 | "properties": { 570 | "Node name for S&R": "InfuseNetLoader" 571 | }, 572 | "widgets_values": [ 573 | "aes_stage2/infusenet_aes_fp8e4m3fn.safetensors" 574 | ] 575 | }, 576 | { 577 | "id": 97, 578 | "type": "LoadImage", 579 | "pos": [ 580 | -985.9622802734375, 581 | 981.4930419921875 582 | ], 583 | "size": [ 584 | 315, 585 | 314 586 | ], 587 | "flags": {}, 588 | "order": 6, 589 | "mode": 0, 590 | "inputs": [], 591 | "outputs": [ 592 | { 593 | "name": "IMAGE", 594 | "type": "IMAGE", 595 | "links": [ 596 | 139, 597 | 144, 598 | 167, 599 | 175 600 | ] 601 | }, 602 | { 603 | "name": "MASK", 604 | "type": "MASK", 605 | "links": null 606 | } 607 | ], 608 | "title": "Load ID Image", 609 | "properties": { 610 | "Node name for S&R": "LoadImage" 611 | }, 612 | "widgets_values": [ 613 | "woman.jpg", 614 | "image", 615 | "" 616 | ] 617 | }, 618 | { 619 | "id": 107, 620 | "type": "LoadImage", 621 | "pos": [ 622 | -558.5816650390625, 623 | 982.09228515625 624 | ], 625 | "size": [ 626 | 315, 627 | 314 628 | ], 629 | "flags": {}, 630 | "order": 7, 631 | "mode": 0, 632 | "inputs": [], 633 | "outputs": [ 634 | { 635 | "name": "IMAGE", 636 | "type": "IMAGE", 637 | "links": [ 638 | 151 639 | ] 640 | }, 641 | { 642 | "name": "MASK", 643 | "type": "MASK", 644 | "links": null 645 | } 646 | ], 647 | "title": "Load Control Image", 648 | "properties": { 649 | "Node name for S&R": "LoadImage" 650 | }, 651 | "widgets_values": [ 652 | "woman.jpg", 653 | "image", 654 | "" 655 | ] 656 | }, 657 | { 658 | "id": 102, 659 | "type": "IDEmbeddingModelLoader", 660 | "pos": [ 661 | -999.3194580078125, 662 | 746.9200439453125 663 | ], 664 | "size": [ 665 | 374.6639709472656, 666 | 170.36888122558594 667 | ], 668 | "flags": { 669 | "collapsed": false 670 | }, 671 | "order": 8, 672 | "mode": 0, 673 | "inputs": [], 674 | "outputs": [ 675 | { 676 | "name": "FACE_DETECTOR", 677 | "type": "MODEL", 678 | "links": [ 679 | 140, 680 | 150, 681 | 172 682 | ] 683 | }, 684 | { 685 | "name": "ARCFACE_MODEL", 686 | "type": "MODEL", 687 | "links": [ 688 | 141, 689 | 173 690 | ] 691 | }, 692 | { 693 | "name": "IMAGE_PROJ_MODEL", 694 | "type": "MODEL", 695 | "links": [ 696 | 142, 697 | 174 698 | ] 699 | } 700 | ], 701 | "properties": { 702 | "Node name for S&R": "IDEmbeddingModelLoader" 703 | }, 704 | "widgets_values": [ 705 | "aes_stage2/image_proj_model.bin", 706 | 8, 707 | "CUDA", 708 | "AUTO" 709 | ] 710 | }, 711 | { 712 | "id": 50, 713 | "type": "CLIPTextEncodeFlux", 714 | "pos": [ 715 | -572.113525390625, 716 | 265.6330871582031 717 | ], 718 | "size": [ 719 | 357.9499816894531, 720 | 361.7095031738281 721 | ], 722 | "flags": {}, 723 | "order": 10, 724 | "mode": 0, 725 | "inputs": [ 726 | { 727 | "name": "clip", 728 | "type": "CLIP", 729 | "link": 93 730 | } 731 | ], 732 | "outputs": [ 733 | { 734 | "name": "CONDITIONING", 735 | "shape": 3, 736 | "type": "CONDITIONING", 737 | "slot_index": 0, 738 | "links": [ 739 | 64, 740 | 133, 741 | 169 742 | ] 743 | } 744 | ], 745 | "properties": { 746 | "Node name for S&R": "CLIPTextEncodeFlux" 747 | }, 748 | "widgets_values": [ 749 | "A young woman holding a sign with the text \"InfiniteYou\", \"Infinite\" in black and \"You\" in red, pure background", 750 | "A young woman holding a sign with the text \"InfiniteYou\", \"Infinite\" in black and \"You\" in red, pure background", 751 | 3.5 752 | ] 753 | }, 754 | { 755 | "id": 47, 756 | "type": "UNETLoader", 757 | "pos": [ 758 | -980.1280517578125, 759 | 264.6694030761719 760 | ], 761 | "size": [ 762 | 270, 763 | 82 764 | ], 765 | "flags": { 766 | "collapsed": false 767 | }, 768 | "order": 9, 769 | "mode": 0, 770 | "inputs": [], 771 | "outputs": [ 772 | { 773 | "name": "MODEL", 774 | "shape": 3, 775 | "type": "MODEL", 776 | "slot_index": 0, 777 | "links": [ 778 | 149, 779 | 184, 780 | 185 781 | ] 782 | } 783 | ], 784 | "properties": { 785 | "Node name for S&R": "UNETLoader" 786 | }, 787 | "widgets_values": [ 788 | "flux1-schnell-fp8-e4m3fn.safetensors", 789 | "fp8_e4m3fn_fast" 790 | ] 791 | } 792 | ], 793 | "links": [ 794 | [ 795 | 58, 796 | 51, 797 | 0, 798 | 44, 799 | 0, 800 | "LATENT" 801 | ], 802 | [ 803 | 59, 804 | 46, 805 | 0, 806 | 44, 807 | 1, 808 | "VAE" 809 | ], 810 | [ 811 | 60, 812 | 44, 813 | 0, 814 | 45, 815 | 0, 816 | "IMAGE" 817 | ], 818 | [ 819 | 66, 820 | 43, 821 | 0, 822 | 51, 823 | 3, 824 | "LATENT" 825 | ], 826 | [ 827 | 93, 828 | 48, 829 | 0, 830 | 50, 831 | 0, 832 | "CLIP" 833 | ], 834 | [ 835 | 94, 836 | 48, 837 | 0, 838 | 52, 839 | 0, 840 | "CLIP" 841 | ], 842 | [ 843 | 127, 844 | 92, 845 | 0, 846 | 51, 847 | 4, 848 | "INT" 849 | ], 850 | [ 851 | 150, 852 | 102, 853 | 0, 854 | 106, 855 | 0, 856 | "MODEL" 857 | ], 858 | [ 859 | 151, 860 | 107, 861 | 0, 862 | 106, 863 | 1, 864 | "IMAGE" 865 | ], 866 | [ 867 | 159, 868 | 52, 869 | 0, 870 | 109, 871 | 1, 872 | "CONDITIONING" 873 | ], 874 | [ 875 | 163, 876 | 109, 877 | 0, 878 | 51, 879 | 1, 880 | "CONDITIONING" 881 | ], 882 | [ 883 | 164, 884 | 109, 885 | 1, 886 | 51, 887 | 2, 888 | "CONDITIONING" 889 | ], 890 | [ 891 | 166, 892 | 108, 893 | 0, 894 | 109, 895 | 3, 896 | "CONTROL_NET" 897 | ], 898 | [ 899 | 169, 900 | 50, 901 | 0, 902 | 109, 903 | 0, 904 | "CONDITIONING" 905 | ], 906 | [ 907 | 170, 908 | 46, 909 | 0, 910 | 109, 911 | 5, 912 | "VAE" 913 | ], 914 | [ 915 | 172, 916 | 102, 917 | 0, 918 | 110, 919 | 0, 920 | "MODEL" 921 | ], 922 | [ 923 | 173, 924 | 102, 925 | 1, 926 | 110, 927 | 1, 928 | "MODEL" 929 | ], 930 | [ 931 | 174, 932 | 102, 933 | 2, 934 | 110, 935 | 2, 936 | "MODEL" 937 | ], 938 | [ 939 | 175, 940 | 97, 941 | 0, 942 | 110, 943 | 3, 944 | "IMAGE" 945 | ], 946 | [ 947 | 176, 948 | 110, 949 | 0, 950 | 109, 951 | 2, 952 | "CONDITIONING" 953 | ], 954 | [ 955 | 180, 956 | 105, 957 | 0, 958 | 109, 959 | 4, 960 | "IMAGE" 961 | ], 962 | [ 963 | 184, 964 | 47, 965 | 0, 966 | 51, 967 | 0, 968 | "MODEL" 969 | ], 970 | [ 971 | 185, 972 | 47, 973 | 0, 974 | 51, 975 | 0, 976 | "MODEL" 977 | ] 978 | ], 979 | "groups": [], 980 | "config": {}, 981 | "extra": { 982 | "ds": { 983 | "scale": 0.7400249944258194, 984 | "offset": [ 985 | 1149.8946398425187, 986 | 572.9878718971356 987 | ] 988 | } 989 | }, 990 | "version": 0.4 991 | } -------------------------------------------------------------------------------- /examples/multi_id_infinite_you_workflow.json: -------------------------------------------------------------------------------- 1 | { 2 | "id": "c1a3d0d1-f0cb-4369-83ae-273696df248c", 3 | "revision": 0, 4 | "last_node_id": 165, 5 | "last_link_id": 320, 6 | "nodes": [ 7 | { 8 | "id": 158, 9 | "type": "EmptySD3LatentImage", 10 | "pos": [ 11 | -437.91448974609375, 12 | 287.0752258300781 13 | ], 14 | "size": [ 15 | 315, 16 | 106 17 | ], 18 | "flags": {}, 19 | "order": 0, 20 | "mode": 0, 21 | "inputs": [], 22 | "outputs": [ 23 | { 24 | "label": "LATENT", 25 | "name": "LATENT", 26 | "type": "LATENT", 27 | "slot_index": 0, 28 | "links": [ 29 | 266 30 | ] 31 | } 32 | ], 33 | "properties": { 34 | "Node name for S&R": "EmptySD3LatentImage" 35 | }, 36 | "widgets_values": [ 37 | 864, 38 | 1152, 39 | 1 40 | ] 41 | }, 42 | { 43 | "id": 159, 44 | "type": "RandomNoise", 45 | "pos": [ 46 | -448.7533874511719, 47 | 448.651611328125 48 | ], 49 | "size": [ 50 | 315, 51 | 82 52 | ], 53 | "flags": {}, 54 | "order": 1, 55 | "mode": 0, 56 | "inputs": [], 57 | "outputs": [ 58 | { 59 | "label": "NOISE", 60 | "name": "NOISE", 61 | "type": "NOISE", 62 | "links": [ 63 | 262 64 | ] 65 | } 66 | ], 67 | "properties": { 68 | "Node name for S&R": "RandomNoise" 69 | }, 70 | "widgets_values": [ 71 | 145190283448493, 72 | "randomize" 73 | ], 74 | "color": "#2a363b", 75 | "bgcolor": "#3f5159" 76 | }, 77 | { 78 | "id": 154, 79 | "type": "DualCLIPLoader", 80 | "pos": [ 81 | -874.8045043945312, 82 | 63.95768737792969 83 | ], 84 | "size": [ 85 | 315, 86 | 130 87 | ], 88 | "flags": {}, 89 | "order": 2, 90 | "mode": 0, 91 | "inputs": [], 92 | "outputs": [ 93 | { 94 | "label": "CLIP", 95 | "name": "CLIP", 96 | "type": "CLIP", 97 | "slot_index": 0, 98 | "links": [ 99 | 269 100 | ] 101 | } 102 | ], 103 | "properties": { 104 | "Node name for S&R": "DualCLIPLoader" 105 | }, 106 | "widgets_values": [ 107 | "t5xxl_fp8_e4m3fn.safetensors", 108 | "clip_l.safetensors", 109 | "flux", 110 | "default" 111 | ] 112 | }, 113 | { 114 | "id": 152, 115 | "type": "VAEDecode", 116 | "pos": [ 117 | 911.7118530273438, 118 | 82.94072723388672 119 | ], 120 | "size": [ 121 | 210, 122 | 46 123 | ], 124 | "flags": {}, 125 | "order": 24, 126 | "mode": 0, 127 | "inputs": [ 128 | { 129 | "label": "samples", 130 | "name": "samples", 131 | "type": "LATENT", 132 | "link": 315 133 | }, 134 | { 135 | "label": "vae", 136 | "name": "vae", 137 | "type": "VAE", 138 | "link": 268 139 | } 140 | ], 141 | "outputs": [ 142 | { 143 | "label": "IMAGE", 144 | "name": "IMAGE", 145 | "type": "IMAGE", 146 | "slot_index": 0, 147 | "links": [ 148 | 276 149 | ] 150 | } 151 | ], 152 | "properties": { 153 | "Node name for S&R": "VAEDecode" 154 | }, 155 | "widgets_values": [] 156 | }, 157 | { 158 | "id": 155, 159 | "type": "VAELoader", 160 | "pos": [ 161 | 635.755615234375, 162 | 728.4552612304688 163 | ], 164 | "size": [ 165 | 311.81634521484375, 166 | 60.429901123046875 167 | ], 168 | "flags": {}, 169 | "order": 3, 170 | "mode": 0, 171 | "inputs": [], 172 | "outputs": [ 173 | { 174 | "label": "VAE", 175 | "name": "VAE", 176 | "type": "VAE", 177 | "slot_index": 0, 178 | "links": [ 179 | 268, 180 | 296, 181 | 301 182 | ] 183 | } 184 | ], 185 | "properties": { 186 | "Node name for S&R": "VAELoader" 187 | }, 188 | "widgets_values": [ 189 | "ae.safetensors" 190 | ] 191 | }, 192 | { 193 | "id": 150, 194 | "type": "BasicGuider", 195 | "pos": [ 196 | 344.4640197753906, 197 | 730.0437622070312 198 | ], 199 | "size": [ 200 | 222.3482666015625, 201 | 46 202 | ], 203 | "flags": {}, 204 | "order": 22, 205 | "mode": 0, 206 | "inputs": [ 207 | { 208 | "label": "model", 209 | "name": "model", 210 | "type": "MODEL", 211 | "link": 320 212 | }, 213 | { 214 | "label": "conditioning", 215 | "name": "conditioning", 216 | "type": "CONDITIONING", 217 | "link": 304 218 | } 219 | ], 220 | "outputs": [ 221 | { 222 | "label": "GUIDER", 223 | "name": "GUIDER", 224 | "type": "GUIDER", 225 | "slot_index": 0, 226 | "links": [ 227 | 263 228 | ] 229 | } 230 | ], 231 | "properties": { 232 | "Node name for S&R": "BasicGuider" 233 | }, 234 | "widgets_values": [] 235 | }, 236 | { 237 | "id": 160, 238 | "type": "KSamplerSelect", 239 | "pos": [ 240 | -433.69488525390625, 241 | 610.5037231445312 242 | ], 243 | "size": [ 244 | 315, 245 | 58 246 | ], 247 | "flags": {}, 248 | "order": 4, 249 | "mode": 0, 250 | "inputs": [], 251 | "outputs": [ 252 | { 253 | "label": "SAMPLER", 254 | "name": "SAMPLER", 255 | "type": "SAMPLER", 256 | "links": [ 257 | 264 258 | ] 259 | } 260 | ], 261 | "properties": { 262 | "Node name for S&R": "KSamplerSelect" 263 | }, 264 | "widgets_values": [ 265 | "euler" 266 | ] 267 | }, 268 | { 269 | "id": 120, 270 | "type": "LoadImage", 271 | "pos": [ 272 | 965.3735961914062, 273 | 1622.5902099609375 274 | ], 275 | "size": [ 276 | 315, 277 | 314 278 | ], 279 | "flags": {}, 280 | "order": 5, 281 | "mode": 0, 282 | "inputs": [], 283 | "outputs": [ 284 | { 285 | "name": "IMAGE", 286 | "type": "IMAGE", 287 | "links": [ 288 | 187 289 | ] 290 | }, 291 | { 292 | "name": "MASK", 293 | "type": "MASK", 294 | "links": null 295 | } 296 | ], 297 | "title": "Load ID Image", 298 | "properties": { 299 | "Node name for S&R": "LoadImage" 300 | }, 301 | "widgets_values": [ 302 | "man.jpg", 303 | "image" 304 | ] 305 | }, 306 | { 307 | "id": 134, 308 | "type": "LoadImageMask", 309 | "pos": [ 310 | 1344.5068359375, 311 | 1615.87109375 312 | ], 313 | "size": [ 314 | 304.0878601074219, 315 | 330 316 | ], 317 | "flags": {}, 318 | "order": 6, 319 | "mode": 0, 320 | "inputs": [], 321 | "outputs": [ 322 | { 323 | "name": "MASK", 324 | "type": "MASK", 325 | "links": [ 326 | 302, 327 | 306 328 | ] 329 | } 330 | ], 331 | "properties": { 332 | "Node name for S&R": "LoadImageMask" 333 | }, 334 | "widgets_values": [ 335 | "mask_right_half.png", 336 | "red", 337 | "image" 338 | ] 339 | }, 340 | { 341 | "id": 119, 342 | "type": "ExtractFacePoseImage", 343 | "pos": [ 344 | 1716.77392578125, 345 | 1481.45556640625 346 | ], 347 | "size": [ 348 | 210, 349 | 122 350 | ], 351 | "flags": {}, 352 | "order": 16, 353 | "mode": 0, 354 | "inputs": [ 355 | { 356 | "name": "face_detector", 357 | "type": "MODEL", 358 | "link": 203 359 | }, 360 | { 361 | "name": "image", 362 | "type": "IMAGE", 363 | "link": 307 364 | }, 365 | { 366 | "name": "mask", 367 | "shape": 7, 368 | "type": "MASK", 369 | "link": 306 370 | } 371 | ], 372 | "outputs": [ 373 | { 374 | "name": "IMAGE", 375 | "type": "IMAGE", 376 | "links": [ 377 | 300 378 | ] 379 | } 380 | ], 381 | "properties": { 382 | "Node name for S&R": "ExtractFacePoseImage" 383 | }, 384 | "widgets_values": [ 385 | 864, 386 | 1152 387 | ] 388 | }, 389 | { 390 | "id": 121, 391 | "type": "ExtractIDEmbedding", 392 | "pos": [ 393 | 1324.1436767578125, 394 | 1391.9034423828125 395 | ], 396 | "size": [ 397 | 367.79998779296875, 398 | 86 399 | ], 400 | "flags": { 401 | "collapsed": true 402 | }, 403 | "order": 14, 404 | "mode": 0, 405 | "inputs": [ 406 | { 407 | "name": "face_detector", 408 | "type": "MODEL", 409 | "link": 188 410 | }, 411 | { 412 | "name": "arcface_model", 413 | "type": "MODEL", 414 | "link": 189 415 | }, 416 | { 417 | "name": "image_proj_model", 418 | "type": "MODEL", 419 | "link": 190 420 | }, 421 | { 422 | "name": "image", 423 | "type": "IMAGE", 424 | "link": 187 425 | } 426 | ], 427 | "outputs": [ 428 | { 429 | "name": "CONDITIONING", 430 | "type": "CONDITIONING", 431 | "links": [ 432 | 298 433 | ] 434 | } 435 | ], 436 | "properties": { 437 | "Node name for S&R": "ExtractIDEmbedding" 438 | }, 439 | "widgets_values": [] 440 | }, 441 | { 442 | "id": 110, 443 | "type": "ExtractIDEmbedding", 444 | "pos": [ 445 | -742.9761962890625, 446 | 1403.471435546875 447 | ], 448 | "size": [ 449 | 367.79998779296875, 450 | 86 451 | ], 452 | "flags": { 453 | "collapsed": true 454 | }, 455 | "order": 18, 456 | "mode": 0, 457 | "inputs": [ 458 | { 459 | "name": "face_detector", 460 | "type": "MODEL", 461 | "link": 172 462 | }, 463 | { 464 | "name": "arcface_model", 465 | "type": "MODEL", 466 | "link": 173 467 | }, 468 | { 469 | "name": "image_proj_model", 470 | "type": "MODEL", 471 | "link": 174 472 | }, 473 | { 474 | "name": "image", 475 | "type": "IMAGE", 476 | "link": 175 477 | } 478 | ], 479 | "outputs": [ 480 | { 481 | "name": "CONDITIONING", 482 | "type": "CONDITIONING", 483 | "links": [ 484 | 293 485 | ] 486 | } 487 | ], 488 | "properties": { 489 | "Node name for S&R": "ExtractIDEmbedding" 490 | }, 491 | "widgets_values": [] 492 | }, 493 | { 494 | "id": 142, 495 | "type": "LoadImageMask", 496 | "pos": [ 497 | -563.2556762695312, 498 | 1623.5535888671875 499 | ], 500 | "size": [ 501 | 330, 502 | 318 503 | ], 504 | "flags": {}, 505 | "order": 7, 506 | "mode": 0, 507 | "inputs": [], 508 | "outputs": [ 509 | { 510 | "name": "MASK", 511 | "type": "MASK", 512 | "links": [ 513 | 297, 514 | 305 515 | ] 516 | } 517 | ], 518 | "properties": { 519 | "Node name for S&R": "LoadImageMask" 520 | }, 521 | "widgets_values": [ 522 | "mask_left_half.png", 523 | "red", 524 | "image" 525 | ] 526 | }, 527 | { 528 | "id": 106, 529 | "type": "ExtractFacePoseImage", 530 | "pos": [ 531 | -142.10414123535156, 532 | 1576.8348388671875 533 | ], 534 | "size": [ 535 | 210, 536 | 122 537 | ], 538 | "flags": {}, 539 | "order": 15, 540 | "mode": 0, 541 | "inputs": [ 542 | { 543 | "name": "face_detector", 544 | "type": "MODEL", 545 | "link": 150 546 | }, 547 | { 548 | "name": "image", 549 | "type": "IMAGE", 550 | "link": 151 551 | }, 552 | { 553 | "name": "mask", 554 | "shape": 7, 555 | "type": "MASK", 556 | "link": 305 557 | } 558 | ], 559 | "outputs": [ 560 | { 561 | "name": "IMAGE", 562 | "type": "IMAGE", 563 | "links": [ 564 | 156, 565 | 295 566 | ] 567 | } 568 | ], 569 | "properties": { 570 | "Node name for S&R": "ExtractFacePoseImage" 571 | }, 572 | "widgets_values": [ 573 | 864, 574 | 1152 575 | ] 576 | }, 577 | { 578 | "id": 108, 579 | "type": "InfuseNetLoader", 580 | "pos": [ 581 | 569.8663940429688, 582 | 1030.9078369140625 583 | ], 584 | "size": [ 585 | 463.8862609863281, 586 | 58 587 | ], 588 | "flags": {}, 589 | "order": 8, 590 | "mode": 0, 591 | "inputs": [], 592 | "outputs": [ 593 | { 594 | "name": "CONTROL_NET", 595 | "type": "CONTROL_NET", 596 | "links": [ 597 | 157, 598 | 294, 599 | 299 600 | ] 601 | } 602 | ], 603 | "properties": { 604 | "Node name for S&R": "InfuseNetLoader" 605 | }, 606 | "widgets_values": [ 607 | "aes_stage2/infusenet_aes_fp8e4m3fn.safetensors" 608 | ] 609 | }, 610 | { 611 | "id": 102, 612 | "type": "IDEmbeddingModelLoader", 613 | "pos": [ 614 | -869.2985229492188, 615 | 977.8091430664062 616 | ], 617 | "size": [ 618 | 530.486328125, 619 | 170 620 | ], 621 | "flags": { 622 | "collapsed": false 623 | }, 624 | "order": 9, 625 | "mode": 0, 626 | "inputs": [], 627 | "outputs": [ 628 | { 629 | "name": "FACE_DETECTOR", 630 | "type": "MODEL", 631 | "links": [ 632 | 140, 633 | 150, 634 | 172, 635 | 188, 636 | 203 637 | ] 638 | }, 639 | { 640 | "name": "ARCFACE_MODEL", 641 | "type": "MODEL", 642 | "links": [ 643 | 141, 644 | 173, 645 | 189 646 | ] 647 | }, 648 | { 649 | "name": "IMAGE_PROJ_MODEL", 650 | "type": "MODEL", 651 | "links": [ 652 | 142, 653 | 174, 654 | 190 655 | ] 656 | } 657 | ], 658 | "properties": { 659 | "Node name for S&R": "IDEmbeddingModelLoader" 660 | }, 661 | "widgets_values": [ 662 | "aes_stage2/image_proj_model.bin", 663 | 8, 664 | "CUDA", 665 | "AUTO" 666 | ] 667 | }, 668 | { 669 | "id": 149, 670 | "type": "FluxGuidance", 671 | "pos": [ 672 | 8.051407814025879, 673 | 69.1958236694336 674 | ], 675 | "size": [ 676 | 317.4000244140625, 677 | 58 678 | ], 679 | "flags": {}, 680 | "order": 19, 681 | "mode": 0, 682 | "inputs": [ 683 | { 684 | "label": "conditioning", 685 | "name": "conditioning", 686 | "type": "CONDITIONING", 687 | "link": 259 688 | } 689 | ], 690 | "outputs": [ 691 | { 692 | "label": "CONDITIONING", 693 | "name": "CONDITIONING", 694 | "type": "CONDITIONING", 695 | "slot_index": 0, 696 | "links": [ 697 | 292 698 | ] 699 | } 700 | ], 701 | "properties": { 702 | "Node name for S&R": "FluxGuidance" 703 | }, 704 | "widgets_values": [ 705 | 3.5 706 | ], 707 | "color": "#233", 708 | "bgcolor": "#355" 709 | }, 710 | { 711 | "id": 153, 712 | "type": "CLIPTextEncode", 713 | "pos": [ 714 | -481.558837890625, 715 | 68.3829574584961 716 | ], 717 | "size": [ 718 | 422.84503173828125, 719 | 164.31304931640625 720 | ], 721 | "flags": {}, 722 | "order": 13, 723 | "mode": 0, 724 | "inputs": [ 725 | { 726 | "label": "clip", 727 | "name": "clip", 728 | "type": "CLIP", 729 | "link": 269 730 | } 731 | ], 732 | "outputs": [ 733 | { 734 | "label": "CONDITIONING", 735 | "name": "CONDITIONING", 736 | "type": "CONDITIONING", 737 | "slot_index": 0, 738 | "links": [ 739 | 259 740 | ] 741 | } 742 | ], 743 | "title": "CLIP Text Encode (Positive Prompt)", 744 | "properties": { 745 | "Node name for S&R": "CLIPTextEncode" 746 | }, 747 | "widgets_values": [ 748 | "A couple in an classroom, The woman, with long wavy brown hair, wears a white graphic tee, denim mini skirt, black boots, the woman is laughing. The man with short black hair, sports a black bomber jacket, white tee, and headphones, and black pants. Both appear casual and stylish. both are looking at camera" 749 | ], 750 | "color": "#232", 751 | "bgcolor": "#353" 752 | }, 753 | { 754 | "id": 107, 755 | "type": "LoadImage", 756 | "pos": [ 757 | 527.5704956054688, 758 | 1645.22998046875 759 | ], 760 | "size": [ 761 | 315, 762 | 314.0001220703125 763 | ], 764 | "flags": {}, 765 | "order": 10, 766 | "mode": 0, 767 | "inputs": [], 768 | "outputs": [ 769 | { 770 | "name": "IMAGE", 771 | "type": "IMAGE", 772 | "links": [ 773 | 151, 774 | 307 775 | ] 776 | }, 777 | { 778 | "name": "MASK", 779 | "type": "MASK", 780 | "links": null 781 | } 782 | ], 783 | "title": "Load Control Image", 784 | "properties": { 785 | "Node name for S&R": "LoadImage" 786 | }, 787 | "widgets_values": [ 788 | "ComfyUI_00050_.png", 789 | "image" 790 | ] 791 | }, 792 | { 793 | "id": 163, 794 | "type": "InfuseNetApply", 795 | "pos": [ 796 | 136.95687866210938, 797 | 1326.953369140625 798 | ], 799 | "size": [ 800 | 315, 801 | 226 802 | ], 803 | "flags": {}, 804 | "order": 20, 805 | "mode": 0, 806 | "inputs": [ 807 | { 808 | "name": "positive", 809 | "type": "CONDITIONING", 810 | "link": 292 811 | }, 812 | { 813 | "name": "id_embedding", 814 | "type": "CONDITIONING", 815 | "link": 293 816 | }, 817 | { 818 | "name": "control_net", 819 | "type": "CONTROL_NET", 820 | "link": 294 821 | }, 822 | { 823 | "name": "image", 824 | "type": "IMAGE", 825 | "link": 295 826 | }, 827 | { 828 | "name": "negative", 829 | "shape": 7, 830 | "type": "CONDITIONING", 831 | "link": null 832 | }, 833 | { 834 | "name": "vae", 835 | "shape": 7, 836 | "type": "VAE", 837 | "link": 296 838 | }, 839 | { 840 | "name": "control_mask", 841 | "shape": 7, 842 | "type": "MASK", 843 | "link": 297 844 | } 845 | ], 846 | "outputs": [ 847 | { 848 | "name": "positive", 849 | "type": "CONDITIONING", 850 | "links": [ 851 | 303 852 | ] 853 | }, 854 | { 855 | "name": "negative", 856 | "type": "CONDITIONING", 857 | "links": null 858 | } 859 | ], 860 | "properties": { 861 | "Node name for S&R": "InfuseNetApply" 862 | }, 863 | "widgets_values": [ 864 | 1.0000000000000002, 865 | 0, 866 | 1 867 | ] 868 | }, 869 | { 870 | "id": 164, 871 | "type": "InfuseNetApply", 872 | "pos": [ 873 | 2013.8388671875, 874 | 1340.6917724609375 875 | ], 876 | "size": [ 877 | 315, 878 | 226 879 | ], 880 | "flags": {}, 881 | "order": 21, 882 | "mode": 0, 883 | "inputs": [ 884 | { 885 | "name": "positive", 886 | "type": "CONDITIONING", 887 | "link": 303 888 | }, 889 | { 890 | "name": "id_embedding", 891 | "type": "CONDITIONING", 892 | "link": 298 893 | }, 894 | { 895 | "name": "control_net", 896 | "type": "CONTROL_NET", 897 | "link": 299 898 | }, 899 | { 900 | "name": "image", 901 | "type": "IMAGE", 902 | "link": 300 903 | }, 904 | { 905 | "name": "negative", 906 | "shape": 7, 907 | "type": "CONDITIONING", 908 | "link": null 909 | }, 910 | { 911 | "name": "vae", 912 | "shape": 7, 913 | "type": "VAE", 914 | "link": 301 915 | }, 916 | { 917 | "name": "control_mask", 918 | "shape": 7, 919 | "type": "MASK", 920 | "link": 302 921 | } 922 | ], 923 | "outputs": [ 924 | { 925 | "name": "positive", 926 | "type": "CONDITIONING", 927 | "links": [ 928 | 304 929 | ] 930 | }, 931 | { 932 | "name": "negative", 933 | "type": "CONDITIONING", 934 | "links": null 935 | } 936 | ], 937 | "properties": { 938 | "Node name for S&R": "InfuseNetApply" 939 | }, 940 | "widgets_values": [ 941 | 1.0000000000000002, 942 | 0, 943 | 1 944 | ] 945 | }, 946 | { 947 | "id": 151, 948 | "type": "SamplerCustomAdvanced", 949 | "pos": [ 950 | 506.7764587402344, 951 | 102.62291717529297 952 | ], 953 | "size": [ 954 | 272.3617858886719, 955 | 124.53733825683594 956 | ], 957 | "flags": {}, 958 | "order": 23, 959 | "mode": 0, 960 | "inputs": [ 961 | { 962 | "label": "noise", 963 | "name": "noise", 964 | "type": "NOISE", 965 | "link": 262 966 | }, 967 | { 968 | "label": "guider", 969 | "name": "guider", 970 | "type": "GUIDER", 971 | "link": 263 972 | }, 973 | { 974 | "label": "sampler", 975 | "name": "sampler", 976 | "type": "SAMPLER", 977 | "link": 264 978 | }, 979 | { 980 | "label": "sigmas", 981 | "name": "sigmas", 982 | "type": "SIGMAS", 983 | "link": 311 984 | }, 985 | { 986 | "label": "latent_image", 987 | "name": "latent_image", 988 | "type": "LATENT", 989 | "link": 266 990 | } 991 | ], 992 | "outputs": [ 993 | { 994 | "label": "output", 995 | "name": "output", 996 | "type": "LATENT", 997 | "slot_index": 0, 998 | "links": [ 999 | 315 1000 | ] 1001 | }, 1002 | { 1003 | "label": "denoised_output", 1004 | "name": "denoised_output", 1005 | "type": "LATENT", 1006 | "links": [] 1007 | } 1008 | ], 1009 | "properties": { 1010 | "Node name for S&R": "SamplerCustomAdvanced" 1011 | }, 1012 | "widgets_values": [] 1013 | }, 1014 | { 1015 | "id": 148, 1016 | "type": "UNETLoader", 1017 | "pos": [ 1018 | -876.4181518554688, 1019 | 267.08319091796875 1020 | ], 1021 | "size": [ 1022 | 315, 1023 | 82 1024 | ], 1025 | "flags": {}, 1026 | "order": 11, 1027 | "mode": 0, 1028 | "inputs": [], 1029 | "outputs": [ 1030 | { 1031 | "label": "MODEL", 1032 | "name": "MODEL", 1033 | "type": "MODEL", 1034 | "slot_index": 0, 1035 | "links": [ 1036 | 319, 1037 | 320 1038 | ] 1039 | } 1040 | ], 1041 | "properties": { 1042 | "Node name for S&R": "UNETLoader" 1043 | }, 1044 | "widgets_values": [ 1045 | "flux1-schnell-fp8-e4m3fn.safetensors", 1046 | "default" 1047 | ], 1048 | "color": "#223", 1049 | "bgcolor": "#335" 1050 | }, 1051 | { 1052 | "id": 97, 1053 | "type": "LoadImage", 1054 | "pos": [ 1055 | -920.3511962890625, 1056 | 1624.9892578125 1057 | ], 1058 | "size": [ 1059 | 315, 1060 | 314 1061 | ], 1062 | "flags": {}, 1063 | "order": 12, 1064 | "mode": 0, 1065 | "inputs": [], 1066 | "outputs": [ 1067 | { 1068 | "name": "IMAGE", 1069 | "type": "IMAGE", 1070 | "links": [ 1071 | 139, 1072 | 144, 1073 | 167, 1074 | 175 1075 | ] 1076 | }, 1077 | { 1078 | "name": "MASK", 1079 | "type": "MASK", 1080 | "links": null 1081 | } 1082 | ], 1083 | "title": "Load ID Image", 1084 | "properties": { 1085 | "Node name for S&R": "LoadImage" 1086 | }, 1087 | "widgets_values": [ 1088 | "woman.jpg", 1089 | "image" 1090 | ] 1091 | }, 1092 | { 1093 | "id": 45, 1094 | "type": "SaveImage", 1095 | "pos": [ 1096 | 1508.172119140625, 1097 | 26.20151138305664 1098 | ], 1099 | "size": [ 1100 | 688.4625244140625, 1101 | 863.3844604492188 1102 | ], 1103 | "flags": { 1104 | "collapsed": false 1105 | }, 1106 | "order": 25, 1107 | "mode": 0, 1108 | "inputs": [ 1109 | { 1110 | "name": "images", 1111 | "type": "IMAGE", 1112 | "link": 276 1113 | } 1114 | ], 1115 | "outputs": [], 1116 | "properties": { 1117 | "Node name for S&R": "SaveImage" 1118 | }, 1119 | "widgets_values": [ 1120 | "ComfyUI" 1121 | ] 1122 | }, 1123 | { 1124 | "id": 161, 1125 | "type": "BasicScheduler", 1126 | "pos": [ 1127 | 83.75202941894531, 1128 | 511.52001953125 1129 | ], 1130 | "size": [ 1131 | 315, 1132 | 106 1133 | ], 1134 | "flags": {}, 1135 | "order": 17, 1136 | "mode": 0, 1137 | "inputs": [ 1138 | { 1139 | "label": "model", 1140 | "name": "model", 1141 | "type": "MODEL", 1142 | "link": 319 1143 | } 1144 | ], 1145 | "outputs": [ 1146 | { 1147 | "label": "SIGMAS", 1148 | "name": "SIGMAS", 1149 | "type": "SIGMAS", 1150 | "links": [ 1151 | 311 1152 | ] 1153 | } 1154 | ], 1155 | "properties": { 1156 | "Node name for S&R": "BasicScheduler" 1157 | }, 1158 | "widgets_values": [ 1159 | "beta", 1160 | 4, 1161 | 1 1162 | ] 1163 | } 1164 | ], 1165 | "links": [ 1166 | [ 1167 | 150, 1168 | 102, 1169 | 0, 1170 | 106, 1171 | 0, 1172 | "MODEL" 1173 | ], 1174 | [ 1175 | 151, 1176 | 107, 1177 | 0, 1178 | 106, 1179 | 1, 1180 | "IMAGE" 1181 | ], 1182 | [ 1183 | 172, 1184 | 102, 1185 | 0, 1186 | 110, 1187 | 0, 1188 | "MODEL" 1189 | ], 1190 | [ 1191 | 173, 1192 | 102, 1193 | 1, 1194 | 110, 1195 | 1, 1196 | "MODEL" 1197 | ], 1198 | [ 1199 | 174, 1200 | 102, 1201 | 2, 1202 | 110, 1203 | 2, 1204 | "MODEL" 1205 | ], 1206 | [ 1207 | 175, 1208 | 97, 1209 | 0, 1210 | 110, 1211 | 3, 1212 | "IMAGE" 1213 | ], 1214 | [ 1215 | 187, 1216 | 120, 1217 | 0, 1218 | 121, 1219 | 3, 1220 | "IMAGE" 1221 | ], 1222 | [ 1223 | 188, 1224 | 102, 1225 | 0, 1226 | 121, 1227 | 0, 1228 | "MODEL" 1229 | ], 1230 | [ 1231 | 189, 1232 | 102, 1233 | 1, 1234 | 121, 1235 | 1, 1236 | "MODEL" 1237 | ], 1238 | [ 1239 | 190, 1240 | 102, 1241 | 2, 1242 | 121, 1243 | 2, 1244 | "MODEL" 1245 | ], 1246 | [ 1247 | 203, 1248 | 102, 1249 | 0, 1250 | 119, 1251 | 0, 1252 | "MODEL" 1253 | ], 1254 | [ 1255 | 259, 1256 | 153, 1257 | 0, 1258 | 149, 1259 | 0, 1260 | "CONDITIONING" 1261 | ], 1262 | [ 1263 | 262, 1264 | 159, 1265 | 0, 1266 | 151, 1267 | 0, 1268 | "NOISE" 1269 | ], 1270 | [ 1271 | 263, 1272 | 150, 1273 | 0, 1274 | 151, 1275 | 1, 1276 | "GUIDER" 1277 | ], 1278 | [ 1279 | 264, 1280 | 160, 1281 | 0, 1282 | 151, 1283 | 2, 1284 | "SAMPLER" 1285 | ], 1286 | [ 1287 | 266, 1288 | 158, 1289 | 0, 1290 | 151, 1291 | 4, 1292 | "LATENT" 1293 | ], 1294 | [ 1295 | 268, 1296 | 155, 1297 | 0, 1298 | 152, 1299 | 1, 1300 | "VAE" 1301 | ], 1302 | [ 1303 | 269, 1304 | 154, 1305 | 0, 1306 | 153, 1307 | 0, 1308 | "CLIP" 1309 | ], 1310 | [ 1311 | 276, 1312 | 152, 1313 | 0, 1314 | 45, 1315 | 0, 1316 | "IMAGE" 1317 | ], 1318 | [ 1319 | 292, 1320 | 149, 1321 | 0, 1322 | 163, 1323 | 0, 1324 | "CONDITIONING" 1325 | ], 1326 | [ 1327 | 293, 1328 | 110, 1329 | 0, 1330 | 163, 1331 | 1, 1332 | "CONDITIONING" 1333 | ], 1334 | [ 1335 | 294, 1336 | 108, 1337 | 0, 1338 | 163, 1339 | 2, 1340 | "CONTROL_NET" 1341 | ], 1342 | [ 1343 | 295, 1344 | 106, 1345 | 0, 1346 | 163, 1347 | 3, 1348 | "IMAGE" 1349 | ], 1350 | [ 1351 | 296, 1352 | 155, 1353 | 0, 1354 | 163, 1355 | 5, 1356 | "VAE" 1357 | ], 1358 | [ 1359 | 297, 1360 | 142, 1361 | 0, 1362 | 163, 1363 | 6, 1364 | "MASK" 1365 | ], 1366 | [ 1367 | 298, 1368 | 121, 1369 | 0, 1370 | 164, 1371 | 1, 1372 | "CONDITIONING" 1373 | ], 1374 | [ 1375 | 299, 1376 | 108, 1377 | 0, 1378 | 164, 1379 | 2, 1380 | "CONTROL_NET" 1381 | ], 1382 | [ 1383 | 300, 1384 | 119, 1385 | 0, 1386 | 164, 1387 | 3, 1388 | "IMAGE" 1389 | ], 1390 | [ 1391 | 301, 1392 | 155, 1393 | 0, 1394 | 164, 1395 | 5, 1396 | "VAE" 1397 | ], 1398 | [ 1399 | 302, 1400 | 134, 1401 | 0, 1402 | 164, 1403 | 6, 1404 | "MASK" 1405 | ], 1406 | [ 1407 | 303, 1408 | 163, 1409 | 0, 1410 | 164, 1411 | 0, 1412 | "CONDITIONING" 1413 | ], 1414 | [ 1415 | 304, 1416 | 164, 1417 | 0, 1418 | 150, 1419 | 1, 1420 | "CONDITIONING" 1421 | ], 1422 | [ 1423 | 305, 1424 | 142, 1425 | 0, 1426 | 106, 1427 | 2, 1428 | "MASK" 1429 | ], 1430 | [ 1431 | 306, 1432 | 134, 1433 | 0, 1434 | 119, 1435 | 2, 1436 | "MASK" 1437 | ], 1438 | [ 1439 | 307, 1440 | 107, 1441 | 0, 1442 | 119, 1443 | 1, 1444 | "IMAGE" 1445 | ], 1446 | [ 1447 | 311, 1448 | 161, 1449 | 0, 1450 | 151, 1451 | 3, 1452 | "SIGMAS" 1453 | ], 1454 | [ 1455 | 315, 1456 | 151, 1457 | 0, 1458 | 152, 1459 | 0, 1460 | "LATENT" 1461 | ], 1462 | [ 1463 | 319, 1464 | 148, 1465 | 0, 1466 | 161, 1467 | 0, 1468 | "MODEL" 1469 | ], 1470 | [ 1471 | 320, 1472 | 148, 1473 | 0, 1474 | 150, 1475 | 0, 1476 | "MODEL" 1477 | ] 1478 | ], 1479 | "groups": [ 1480 | { 1481 | "id": 1, 1482 | "title": "ID 1", 1483 | "bounding": [ 1484 | -938.527099609375, 1485 | 1232.126953125, 1486 | 1409.8516845703125, 1487 | 723.2393798828125 1488 | ], 1489 | "color": "#3f789e", 1490 | "font_size": 24, 1491 | "flags": {} 1492 | }, 1493 | { 1494 | "id": 2, 1495 | "title": "ID 2", 1496 | "bounding": [ 1497 | 941.7040405273438, 1498 | 1250.2364501953125, 1499 | 1486.1617431640625, 1500 | 706.6896362304688 1501 | ], 1502 | "color": "#3f789e", 1503 | "font_size": 24, 1504 | "flags": {} 1505 | }, 1506 | { 1507 | "id": 4, 1508 | "title": "Flux", 1509 | "bounding": [ 1510 | -898.5185546875, 1511 | -25.370464324951172, 1512 | 2022.9615478515625, 1513 | 921.326416015625 1514 | ], 1515 | "color": "#3f789e", 1516 | "font_size": 24, 1517 | "flags": {} 1518 | } 1519 | ], 1520 | "config": {}, 1521 | "extra": { 1522 | "ds": { 1523 | "scale": 0.32349184307607415, 1524 | "offset": [ 1525 | 1020.3511962890625, 1526 | 73.79848861694336 1527 | ] 1528 | }, 1529 | "frontendVersion": "1.20.4", 1530 | "groupNodes": {} 1531 | }, 1532 | "version": 0.4 1533 | } -------------------------------------------------------------------------------- /examples/multi_id_workflow.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/bytedance/ComfyUI_InfiniteYou/fb507e133b5e6e4a86a9c63a9b58f562aed4906a/examples/multi_id_workflow.jpg -------------------------------------------------------------------------------- /examples/teaser.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/bytedance/ComfyUI_InfiniteYou/fb507e133b5e6e4a86a9c63a9b58f562aed4906a/examples/teaser.jpg -------------------------------------------------------------------------------- /examples/workflow.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/bytedance/ComfyUI_InfiniteYou/fb507e133b5e6e4a86a9c63a9b58f562aed4906a/examples/workflow.jpg -------------------------------------------------------------------------------- /infuse_net.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) 2025 Bytedance Ltd. and/or its affiliates. All rights reserved. 2 | 3 | # Licensed under the Apache License, Version 2.0 (the "License"); 4 | # you may not use this file except in compliance with the License. 5 | # You may obtain a copy of the License at 6 | 7 | # http://www.apache.org/licenses/LICENSE-2.0 8 | 9 | # Unless required by applicable law or agreed to in writing, software 10 | # distributed under the License is distributed on an "AS IS" BASIS, 11 | # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 12 | # See the License for the specific language governing permissions and 13 | # limitations under the License. 14 | 15 | import torch 16 | import math 17 | from torch import Tensor, nn 18 | from einops import rearrange, repeat 19 | 20 | import comfy 21 | from comfy.controlnet import controlnet_config, controlnet_load_state_dict, ControlNet, StrengthType 22 | from comfy.ldm.flux.model import Flux 23 | from comfy.ldm.flux.layers import (timestep_embedding) 24 | import comfy.ldm.common_dit 25 | 26 | class InfuseNet(ControlNet): 27 | def __init__(self, 28 | control_model=None, 29 | id_embedding = None, 30 | global_average_pooling=False, 31 | compression_ratio=8, 32 | latent_format=None, 33 | load_device=None, 34 | manual_cast_dtype=None, 35 | extra_conds=["y"], 36 | strength_type=StrengthType.CONSTANT, 37 | concat_mask=False, 38 | preprocess_image=lambda a: a): 39 | super().__init__(control_model=control_model, 40 | global_average_pooling=global_average_pooling, 41 | compression_ratio=compression_ratio, 42 | latent_format=latent_format, 43 | load_device=load_device, 44 | manual_cast_dtype=manual_cast_dtype, 45 | extra_conds=extra_conds, 46 | strength_type=strength_type, 47 | concat_mask=concat_mask, 48 | preprocess_image=preprocess_image) 49 | self.id_embedding = id_embedding 50 | 51 | def copy(self): 52 | c = InfuseNet(None, global_average_pooling=self.global_average_pooling, load_device=self.load_device, manual_cast_dtype=self.manual_cast_dtype) 53 | c.control_model = self.control_model 54 | c.control_model_wrapped = self.control_model_wrapped 55 | c.id_embedding = self.id_embedding 56 | self.copy_to(c) 57 | return c 58 | 59 | def get_control(self, x_noisy, t, cond, batched_number, transformer_options): 60 | cond = cond.copy() 61 | cond['crossattn_controlnet'] = self.id_embedding 62 | cond['c_crossattn'] = self.id_embedding 63 | return super().get_control(x_noisy, t, cond, batched_number, transformer_options) 64 | 65 | class InfuseNetFlux(Flux): 66 | def __init__(self, latent_input=False, num_union_modes=0, mistoline=False, control_latent_channels=None, image_model=None, dtype=None, device=None, operations=None, **kwargs): 67 | super().__init__(final_layer=False, dtype=dtype, device=device, operations=operations, **kwargs) 68 | 69 | self.main_model_double = 19 70 | self.main_model_single = 38 71 | 72 | self.mistoline = mistoline 73 | # add ControlNet blocks 74 | if self.mistoline: 75 | control_block = lambda : MistolineControlnetBlock(self.hidden_size, dtype=dtype, device=device, operations=operations) 76 | else: 77 | control_block = lambda : operations.Linear(self.hidden_size, self.hidden_size, dtype=dtype, device=device) 78 | 79 | self.controlnet_blocks = nn.ModuleList([]) 80 | for _ in range(self.params.depth): 81 | self.controlnet_blocks.append(control_block()) 82 | 83 | self.controlnet_single_blocks = nn.ModuleList([]) 84 | for _ in range(self.params.depth_single_blocks): 85 | self.controlnet_single_blocks.append(control_block()) 86 | 87 | self.num_union_modes = num_union_modes 88 | self.controlnet_mode_embedder = None 89 | if self.num_union_modes > 0: 90 | self.controlnet_mode_embedder = operations.Embedding(self.num_union_modes, self.hidden_size, dtype=dtype, device=device) 91 | 92 | self.gradient_checkpointing = False 93 | self.latent_input = latent_input 94 | if control_latent_channels is None: 95 | control_latent_channels = self.in_channels 96 | else: 97 | control_latent_channels *= 2 * 2 #patch size 98 | 99 | self.pos_embed_input = operations.Linear(control_latent_channels, self.hidden_size, bias=True, dtype=dtype, device=device) 100 | if not self.latent_input: 101 | if self.mistoline: 102 | self.input_cond_block = MistolineCondDownsamplBlock(dtype=dtype, device=device, operations=operations) 103 | else: 104 | self.input_hint_block = nn.Sequential( 105 | operations.Conv2d(3, 16, 3, padding=1, dtype=dtype, device=device), 106 | nn.SiLU(), 107 | operations.Conv2d(16, 16, 3, padding=1, dtype=dtype, device=device), 108 | nn.SiLU(), 109 | operations.Conv2d(16, 16, 3, padding=1, stride=2, dtype=dtype, device=device), 110 | nn.SiLU(), 111 | operations.Conv2d(16, 16, 3, padding=1, dtype=dtype, device=device), 112 | nn.SiLU(), 113 | operations.Conv2d(16, 16, 3, padding=1, stride=2, dtype=dtype, device=device), 114 | nn.SiLU(), 115 | operations.Conv2d(16, 16, 3, padding=1, dtype=dtype, device=device), 116 | nn.SiLU(), 117 | operations.Conv2d(16, 16, 3, padding=1, stride=2, dtype=dtype, device=device), 118 | nn.SiLU(), 119 | operations.Conv2d(16, 16, 3, padding=1, dtype=dtype, device=device) 120 | ) 121 | 122 | def forward_orig( 123 | self, 124 | img: Tensor, 125 | img_ids: Tensor, 126 | controlnet_cond: Tensor, 127 | txt: Tensor, 128 | txt_ids: Tensor, 129 | timesteps: Tensor, 130 | y: Tensor, 131 | guidance: Tensor = None, 132 | control_type: Tensor = None, 133 | out_mask: Tensor = None 134 | ) -> Tensor: 135 | if img.ndim != 3 or txt.ndim != 3: 136 | raise ValueError("Input img and txt tensors must have 3 dimensions.") 137 | 138 | # running on sequences img 139 | img = self.img_in(img) 140 | 141 | controlnet_cond = self.pos_embed_input(controlnet_cond) 142 | img = img + controlnet_cond 143 | vec = self.time_in(timestep_embedding(timesteps, 256)) 144 | if self.params.guidance_embed: 145 | vec = vec + self.guidance_in(timestep_embedding(guidance, 256)) 146 | vec = vec + self.vector_in(y) 147 | txt = self.txt_in(txt) 148 | 149 | if self.controlnet_mode_embedder is not None and len(control_type) > 0: 150 | control_cond = self.controlnet_mode_embedder(torch.tensor(control_type, device=img.device), out_dtype=img.dtype).unsqueeze(0).repeat((txt.shape[0], 1, 1)) 151 | txt = torch.cat([control_cond, txt], dim=1) 152 | txt_ids = torch.cat([txt_ids[:,:1], txt_ids], dim=1) 153 | 154 | ids = torch.cat((txt_ids, img_ids), dim=1) 155 | pe = self.pe_embedder(ids) 156 | 157 | controlnet_double = () 158 | 159 | for i in range(len(self.double_blocks)): 160 | img, txt = self.double_blocks[i](img=img, txt=txt, vec=vec, pe=pe) 161 | controlnet_double = controlnet_double + (self.controlnet_blocks[i](img),) 162 | 163 | img = torch.cat((txt, img), 1) 164 | 165 | controlnet_single = () 166 | 167 | for i in range(len(self.single_blocks)): 168 | img = self.single_blocks[i](img, vec=vec, pe=pe) 169 | controlnet_single = controlnet_single + (self.controlnet_single_blocks[i](img[:, txt.shape[1] :, ...]),) 170 | 171 | repeat = math.ceil(self.main_model_double / len(controlnet_double)) 172 | if self.latent_input: 173 | out_input = () 174 | for x in controlnet_double: 175 | if out_mask is not None: 176 | out_input += (x * out_mask,) * repeat 177 | else: 178 | out_input += (x,) * repeat 179 | else: 180 | out_input = (controlnet_double * repeat) 181 | 182 | out = {"input": out_input[:self.main_model_double]} 183 | if len(controlnet_single) > 0: 184 | repeat = math.ceil(self.main_model_single / len(controlnet_single)) 185 | out_output = () 186 | if self.latent_input: 187 | for x in controlnet_single: 188 | if out_mask is not None: 189 | out_output += (x * out_mask,) * repeat 190 | else: 191 | out_output += (x,) * repeat 192 | else: 193 | out_output = (controlnet_single * repeat) 194 | out["output"] = out_output[:self.main_model_single] 195 | return out 196 | 197 | def forward(self, x, timesteps, context, y, guidance=None, hint=None, **kwargs): 198 | patch_size = 2 199 | if self.latent_input: 200 | hint = comfy.ldm.common_dit.pad_to_patch_size(hint, (patch_size, patch_size)) 201 | elif self.mistoline: 202 | hint = hint * 2.0 - 1.0 203 | hint = self.input_cond_block(hint) 204 | else: 205 | hint = hint * 2.0 - 1.0 206 | hint = self.input_hint_block(hint) 207 | 208 | hint = rearrange(hint, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=patch_size, pw=patch_size) 209 | 210 | bs, c, h, w = x.shape 211 | x = comfy.ldm.common_dit.pad_to_patch_size(x, (patch_size, patch_size)) 212 | 213 | img = rearrange(x, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=patch_size, pw=patch_size) 214 | 215 | control_mask = kwargs.get("control_mask", None) 216 | out_mask = None 217 | if control_mask is not None: 218 | in_mask = comfy.sampler_helpers.prepare_mask(control_mask, (bs, c, h, w), img.device) 219 | in_mask = comfy.ldm.common_dit.pad_to_patch_size(in_mask, (patch_size, patch_size)) 220 | in_mask = rearrange(in_mask, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=patch_size, pw=patch_size) 221 | # (b, seq_len, _) =>(b, seq_len, pulid.dim) 222 | in_mask = in_mask[..., 0].unsqueeze(-1).repeat(1, 1, img.shape[-1]).to(dtype=img.dtype) 223 | img = img * in_mask 224 | 225 | out_mask = comfy.sampler_helpers.prepare_mask(control_mask, (bs, 226 | self.hidden_size // (patch_size * patch_size), h, w), img.device) 227 | out_mask = comfy.ldm.common_dit.pad_to_patch_size(out_mask, (patch_size, patch_size)) 228 | out_mask = rearrange(out_mask, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=patch_size, pw=patch_size) 229 | 230 | h_len = ((h + (patch_size // 2)) // patch_size) 231 | w_len = ((w + (patch_size // 2)) // patch_size) 232 | img_ids = torch.zeros((h_len, w_len, 3), device=x.device, dtype=x.dtype) 233 | img_ids[..., 1] = img_ids[..., 1] + torch.linspace(0, h_len - 1, steps=h_len, device=x.device, dtype=x.dtype)[:, None] 234 | img_ids[..., 2] = img_ids[..., 2] + torch.linspace(0, w_len - 1, steps=w_len, device=x.device, dtype=x.dtype)[None, :] 235 | img_ids = repeat(img_ids, "h w c -> b (h w) c", b=bs) 236 | 237 | txt_ids = torch.zeros((bs, context.shape[1], 3), device=x.device, dtype=x.dtype) 238 | return self.forward_orig(img, img_ids, hint, context, txt_ids, timesteps, y, guidance, control_type=kwargs.get("control_type", []), out_mask=out_mask) 239 | 240 | def load_infuse_net_flux(ckpt_path, model_options={}): 241 | sd = comfy.utils.load_torch_file(ckpt_path, safe_load=True) 242 | new_sd = comfy.model_detection.convert_diffusers_mmdit(sd, "") 243 | model_config, operations, load_device, unet_dtype, manual_cast_dtype, offload_device = controlnet_config(new_sd, model_options=model_options) 244 | for k in sd: 245 | new_sd[k] = sd[k] 246 | 247 | num_union_modes = 0 248 | union_cnet = "controlnet_mode_embedder.weight" 249 | if union_cnet in new_sd: 250 | num_union_modes = new_sd[union_cnet].shape[0] 251 | 252 | control_latent_channels = new_sd.get("pos_embed_input.weight").shape[1] // 4 253 | concat_mask = False 254 | if control_latent_channels == 17: 255 | concat_mask = True 256 | 257 | control_model = InfuseNetFlux(latent_input=True, num_union_modes=num_union_modes, control_latent_channels=control_latent_channels, operations=operations, device=offload_device, dtype=unet_dtype, **model_config.unet_config) 258 | control_model = controlnet_load_state_dict(control_model, new_sd) 259 | 260 | latent_format = comfy.latent_formats.Flux() 261 | extra_conds = ['y', 'guidance'] 262 | control = InfuseNet(control_model, compression_ratio=1, latent_format=latent_format, concat_mask=concat_mask, load_device=load_device, manual_cast_dtype=manual_cast_dtype, extra_conds=extra_conds) 263 | return control 264 | -------------------------------------------------------------------------------- /nodes.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) 2025 Bytedance Ltd. and/or its affiliates. All rights reserved. 2 | 3 | # Licensed under the Apache License, Version 2.0 (the "License"); 4 | # you may not use this file except in compliance with the License. 5 | # You may obtain a copy of the License at 6 | 7 | # http://www.apache.org/licenses/LICENSE-2.0 8 | 9 | # Unless required by applicable law or agreed to in writing, software 10 | # distributed under the License is distributed on an "AS IS" BASIS, 11 | # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 12 | # See the License for the specific language governing permissions and 13 | # limitations under the License. 14 | 15 | import os 16 | import torch 17 | import folder_paths 18 | import cv2 19 | import numpy as np 20 | from PIL import Image 21 | import comfy 22 | from huggingface_hub import snapshot_download, hf_hub_download 23 | import shutil 24 | import glob 25 | 26 | from facexlib.recognition import init_recognition_model 27 | from insightface.app import FaceAnalysis 28 | 29 | from .utils import extract_arcface_bgr_embedding, tensor_to_np_image, np_image_to_tensor, resize_and_pad_pil_image, draw_kps, escape_path_for_url 30 | from .infuse_net import load_infuse_net_flux 31 | from .resampler import Resampler 32 | 33 | folder_paths.add_model_folder_path("infinite_you", os.path.join(folder_paths.models_dir, "infinite_you")) 34 | 35 | class FaceDetector: 36 | def __init__(self, 37 | det_sizes, 38 | root_dir, 39 | providers) -> None: 40 | self.apps = [] 41 | for det_size in det_sizes: 42 | app = FaceAnalysis(name="antelopev2", root=root_dir, providers=providers) 43 | app.prepare(ctx_id=0, det_size=(det_size, det_size)) 44 | self.apps.append(app) 45 | 46 | def __call__(self, np_image_bgr): 47 | for app in self.apps: 48 | faces = app.get(np_image_bgr) 49 | if len(faces) > 0: 50 | return faces 51 | return [] 52 | 53 | class IDEmbeddingModelLoader: 54 | @classmethod 55 | def INPUT_TYPES(s): 56 | return { 57 | "required": { 58 | "image_proj_model_name": (IDEmbeddingModelLoader.get_image_proj_names(), ), 59 | 'image_proj_num_tokens': ([8, 16], ), 60 | 'face_analysis_provider': (['CUDA', 'CPU'], ), 61 | 'face_analysis_det_size': (["AUTO", "640", "320", "160"], ) 62 | }, 63 | } 64 | 65 | RETURN_NAMES = ("FACE_DETECTOR", "ARCFACE_MODEL", "IMAGE_PROJ_MODEL") 66 | RETURN_TYPES = ("MODEL", "MODEL", "MODEL") 67 | 68 | FUNCTION = "load_insightface" 69 | CATEGORY = "infinite_you" 70 | 71 | def get_image_proj_names(): 72 | names = [ 73 | os.path.join("sim_stage1", "image_proj_model.bin"), 74 | os.path.join("aes_stage2", "image_proj_model.bin"), 75 | *folder_paths.get_filename_list("infinite_you"), 76 | ] 77 | return list(filter(lambda x: x.endswith(".bin"), list(set(names)))) 78 | 79 | def load_insightface(self, image_proj_model_name, image_proj_num_tokens, face_analysis_provider, face_analysis_det_size): 80 | insight_facedir = os.path.join(folder_paths.models_dir, "insightface") 81 | 82 | # Download insightface models 83 | antelopev2_dir = os.path.join(insight_facedir, 'models', 'antelopev2') 84 | if not os.path.exists(antelopev2_dir) or len(glob.glob(os.path.join(antelopev2_dir, "*.onnx"))) == 0: 85 | os.makedirs(antelopev2_dir, exist_ok=True) 86 | snapshot_download(repo_id="MonsterMMORPG/tools", allow_patterns="*.onnx", local_dir=antelopev2_dir) 87 | 88 | # Download infinite you models 89 | infinite_you_dir = os.path.join(folder_paths.models_dir, "infinite_you") 90 | image_proj_model_path = os.path.join(infinite_you_dir, image_proj_model_name) 91 | if not os.path.exists(image_proj_model_path): 92 | dst_dir = os.path.dirname(image_proj_model_path) 93 | os.makedirs(dst_dir, exist_ok=True) 94 | 95 | downloaded_file = hf_hub_download(repo_id="ByteDance/InfiniteYou", 96 | filename=escape_path_for_url(os.path.join("infu_flux_v1.0", image_proj_model_name)), 97 | local_dir=infinite_you_dir) 98 | shutil.move(downloaded_file, image_proj_model_path) 99 | 100 | provider = 'CPUExecutionProvider' 101 | if face_analysis_provider == 'CUDA': 102 | provider = 'CUDAExecutionProvider' 103 | det_sizes = [] 104 | if face_analysis_det_size == 'AUTO': 105 | det_sizes = [640, 320, 160] 106 | else: 107 | det_sizes = [int(face_analysis_det_size)] 108 | face_detector = FaceDetector(det_sizes=det_sizes, root_dir=insight_facedir, providers=[provider]) 109 | 110 | device = comfy.model_management.get_torch_device() 111 | 112 | # Load arcface model 113 | arcface_model = init_recognition_model('arcface', device=device) 114 | 115 | # Load image proj model 116 | image_emb_dim = 512 117 | image_proj_model = Resampler( 118 | dim=1280, 119 | depth=4, 120 | dim_head=64, 121 | heads=20, 122 | num_queries=image_proj_num_tokens, 123 | embedding_dim=image_emb_dim, 124 | output_dim=4096, 125 | ff_mult=4, 126 | ) 127 | ipm_state_dict = torch.load(image_proj_model_path, map_location="cpu") 128 | image_proj_model.load_state_dict(ipm_state_dict['image_proj']) 129 | del ipm_state_dict 130 | image_proj_model.to(device, torch.bfloat16) 131 | image_proj_model.eval() 132 | 133 | return (face_detector, arcface_model, image_proj_model) 134 | 135 | class ExtractFacePoseImage: 136 | @classmethod 137 | def INPUT_TYPES(s): 138 | return { 139 | "required": { 140 | "face_detector": ("MODEL", ), 141 | "image": ("IMAGE", ), 142 | "width": ("INT", {"default": 864, "min": 0, "max": 2048, "step": 1}), 143 | "height": ("INT", {"default": 1152, "min": 0, "max": 2048, "step": 1}), 144 | }, 145 | "optional": { 146 | "mask": ("MASK", ), 147 | } 148 | } 149 | 150 | RETURN_TYPES = ("IMAGE",) 151 | FUNCTION = "extract_face_pose" 152 | CATEGORY = "infinite_you" 153 | 154 | def extract_face_pose(self, face_detector, image, width, height, mask = None): 155 | np_image = tensor_to_np_image(image)[0] 156 | if mask is not None: 157 | np_mask = tensor_to_np_image(mask)[0] 158 | np_mask = cv2.resize(np_mask, (np_image.shape[1], np_image.shape[0]), interpolation=cv2.INTER_NEAREST) 159 | mask_3ch = np.expand_dims(np_mask, axis=-1) 160 | mask_3ch = np.repeat(mask_3ch, 3, axis=-1) # Shape: (H, W, 3) 161 | np_image = np_image * mask_3ch 162 | 163 | pil_image = resize_and_pad_pil_image(Image.fromarray(np_image), (width, height)) 164 | face_info = face_detector(cv2.cvtColor(np.array(pil_image), cv2.COLOR_RGB2BGR)) 165 | if len(face_info) == 0: 166 | raise ValueError('No face detected in the input pose image') 167 | 168 | face_info = sorted(face_info, key=lambda x:(x['bbox'][2]-x['bbox'][0])*(x['bbox'][3]-x['bbox'][1]))[-1] # only use the maximum face 169 | pil_image = draw_kps(pil_image, face_info['kps']) 170 | 171 | return (np_image_to_tensor(np.array(pil_image)).unsqueeze(0), ) 172 | 173 | class ExtractIDEmbedding: 174 | @classmethod 175 | def INPUT_TYPES(s): 176 | return { 177 | "required": { 178 | "face_detector": ("MODEL", ), 179 | "arcface_model": ("MODEL", ), 180 | "image_proj_model": ("MODEL", ), 181 | "image": ("IMAGE", ), 182 | } 183 | } 184 | 185 | RETURN_TYPES = ("CONDITIONING",) 186 | FUNCTION = "extract_id_embedding" 187 | CATEGORY = "infinite_you" 188 | 189 | def extract_id_embedding(self, face_detector, arcface_model, image_proj_model, image): 190 | np_image = tensor_to_np_image(image) 191 | id_image_cv2 = cv2.cvtColor(np_image[0], cv2.COLOR_RGB2BGR) 192 | face_info = face_detector(id_image_cv2) 193 | if len(face_info) == 0: 194 | raise ValueError('No face detected in the input ID image') 195 | 196 | device = comfy.model_management.get_torch_device() 197 | 198 | face_info = sorted(face_info, key=lambda x:(x['bbox'][2]-x['bbox'][0])*(x['bbox'][3]-x['bbox'][1]))[-1] # only use the maximum face 199 | landmark = face_info['kps'] 200 | id_embed = extract_arcface_bgr_embedding(id_image_cv2, landmark, arcface_model) 201 | id_embed = id_embed.clone().unsqueeze(0).float().to(device) 202 | id_embed = id_embed.reshape([1, -1, 512]) 203 | id_embed = id_embed.to(device=device, dtype=torch.bfloat16) 204 | with torch.no_grad(): 205 | id_embed = image_proj_model(id_embed) 206 | bs_embed, seq_len, _ = id_embed.shape 207 | id_embed = id_embed.repeat(1, 1, 1) 208 | id_embed = id_embed.view(bs_embed * 1, seq_len, -1) 209 | id_embed = id_embed.to(device=device, dtype=torch.bfloat16) 210 | 211 | return ({'id_embedding': id_embed}, ) 212 | 213 | class InfuseNetLoader: 214 | @classmethod 215 | def INPUT_TYPES(s): 216 | return {"required": { "controlnet_name": (InfuseNetLoader.get_controlnet_names(), )}} 217 | 218 | def get_controlnet_names(): 219 | names = [ 220 | os.path.join("sim_stage1", "infusenet_sim_bf16.safetensors"), 221 | os.path.join("sim_stage1", "infusenet_sim_fp8e4m3fn.safetensors"), 222 | os.path.join("aes_stage2", "infusenet_aes_bf16.safetensors"), 223 | os.path.join("aes_stage2", "infusenet_aes_fp8e4m3fn.safetensors"), 224 | *folder_paths.get_filename_list("infinite_you"), 225 | ] 226 | return list(filter(lambda x: x.endswith(".safetensors"), list(set(names)))) 227 | 228 | RETURN_TYPES = ("CONTROL_NET",) 229 | FUNCTION = "load_controlnet" 230 | 231 | CATEGORY = "infinite_you" 232 | 233 | def load_controlnet(self, controlnet_name): 234 | infinite_you_dir = os.path.join(folder_paths.models_dir, "infinite_you") 235 | controlnet_path = os.path.join(infinite_you_dir, controlnet_name) 236 | 237 | if not os.path.exists(controlnet_path): 238 | dst_dir = os.path.dirname(controlnet_path) 239 | os.makedirs(dst_dir, exist_ok=True) 240 | downloaded_file = hf_hub_download(repo_id="ByteDance/InfiniteYou", 241 | filename=escape_path_for_url(os.path.join("infu_flux_v1.0", controlnet_name)), 242 | local_dir=infinite_you_dir) 243 | 244 | shutil.move(downloaded_file, controlnet_path) 245 | 246 | controlnet = load_infuse_net_flux(controlnet_path) 247 | return (controlnet,) 248 | 249 | class InfuseNetApply: 250 | @classmethod 251 | def INPUT_TYPES(s): 252 | return {"required": {"positive": ("CONDITIONING", ), 253 | "id_embedding": ("CONDITIONING", ), 254 | "control_net": ("CONTROL_NET", ), 255 | "image": ("IMAGE", ), 256 | "strength": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 10.0, "step": 0.01}), 257 | "start_percent": ("FLOAT", {"default": 0.0, "min": 0.0, "max": 1.0, "step": 0.001}), 258 | "end_percent": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.001}) 259 | }, 260 | "optional": { 261 | "negative": ("CONDITIONING", ), 262 | "vae": ("VAE", ), 263 | "control_mask": ("MASK", ), 264 | } 265 | } 266 | 267 | RETURN_TYPES = ("CONDITIONING","CONDITIONING") 268 | RETURN_NAMES = ("positive", "negative") 269 | FUNCTION = "apply_controlnet" 270 | 271 | CATEGORY = "infinite_you" 272 | 273 | def apply_controlnet(self, positive, id_embedding, control_net, image, strength, start_percent, end_percent, negative = None, vae=None, control_mask=None, extra_concat=[]): 274 | if strength == 0: 275 | return (positive, negative) 276 | 277 | if control_mask is not None: 278 | if control_mask.dim() > 3: 279 | control_mask = control_mask.squeeze(-1) 280 | elif control_mask.dim() < 3: 281 | control_mask = control_mask.unsqueeze(0) 282 | 283 | control_hint = image.movedim(-1,1) 284 | cnets = {} 285 | 286 | out = [] 287 | for conditioning in [positive, negative]: 288 | c = [] 289 | if conditioning is None: 290 | out.append(None) 291 | continue 292 | 293 | for t in conditioning: 294 | d = t[1].copy() 295 | 296 | prev_cnet = d.get('control', None) 297 | if prev_cnet in cnets: 298 | c_net = cnets[prev_cnet] 299 | else: 300 | c_net = control_net.copy().set_cond_hint(control_hint, strength, (start_percent, end_percent), vae=vae, extra_concat=extra_concat) 301 | c_net.id_embedding = id_embedding['id_embedding'] 302 | c_net.set_previous_controlnet(prev_cnet) 303 | c_net.set_extra_arg("control_mask", control_mask) 304 | cnets[prev_cnet] = c_net 305 | 306 | d['control'] = c_net 307 | d['control_apply_to_uncond'] = False 308 | n = [t[0], d] 309 | c.append(n) 310 | out.append(c) 311 | return (out[0], out[1]) 312 | 313 | NODE_CLASS_MAPPINGS = { 314 | "IDEmbeddingModelLoader": IDEmbeddingModelLoader, 315 | "ExtractIDEmbedding": ExtractIDEmbedding, 316 | "ExtractFacePoseImage": ExtractFacePoseImage, 317 | "InfuseNetApply": InfuseNetApply, 318 | "InfuseNetLoader": InfuseNetLoader, 319 | } 320 | 321 | NODE_DISPLAY_NAME_MAPPINGS = { 322 | "IDEmbeddingModelLoader": "ID Embedding Model Loader", 323 | "ExtractIDEmbedding": "Extract ID Embedding", 324 | "ExtractFacePoseImage": "Extract Face Pose Image", 325 | "InfuseNetApply": "Apply InfuseNet", 326 | "InfuseNetLoader": "Load InfuseNet", 327 | } 328 | -------------------------------------------------------------------------------- /pyproject.toml: -------------------------------------------------------------------------------- 1 | [project] 2 | name = "infiniteyou" 3 | description = "Official ComfyUI Support - InfiniteYou: Flexible Photo Recrafting While Preserving Your Identity" 4 | version = "1.0.1" 5 | license = {file = "LICENSE"} 6 | dependencies = ["facexlib>=0.3.0", "onnxruntime>=1.19.2", "insightface>=0.7.3", "opencv-python>=4.11.0.86", "huggingface_hub"] 7 | 8 | [project.urls] 9 | Repository = "https://github.com/bytedance/ComfyUI_InfiniteYou" 10 | # Used by Comfy Registry https://comfyregistry.org 11 | 12 | [tool.comfy] 13 | PublisherId = "yuminjia" 14 | DisplayName = "ComfyUI_InfiniteYou" 15 | Icon = "" 16 | -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- 1 | facexlib>=0.3.0 2 | onnxruntime>=1.19.2 3 | insightface>=0.7.3 4 | opencv-python>=4.11.0.86 5 | huggingface_hub 6 | -------------------------------------------------------------------------------- /resampler.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) 2023 Anas Awadalla, Irena Gao, Joshua Gardner, Jack Hessel, Yusuf Hanafy, 2 | # Wanrong Zhu, Kalyani Marathe, Yonatan Bitton, Samir Gadre, Jenia Jitsev, Simon Kornblith, 3 | # Pang Wei Koh, Gabriel Ilharco, Mitchell Wortsman, Ludwig Schmidt. 4 | # SPDX-License-Identifier: MIT 5 | 6 | # Modified from https://github.com/mlfoundations/open_flamingo/blob/main/open_flamingo/src/helpers.py 7 | 8 | import math 9 | 10 | import torch 11 | import torch.nn as nn 12 | 13 | 14 | # FFN 15 | def FeedForward(dim, mult=4): 16 | inner_dim = int(dim * mult) 17 | return nn.Sequential( 18 | nn.LayerNorm(dim), 19 | nn.Linear(dim, inner_dim, bias=False), 20 | nn.GELU(), 21 | nn.Linear(inner_dim, dim, bias=False), 22 | ) 23 | 24 | 25 | def reshape_tensor(x, heads): 26 | bs, length, width = x.shape 27 | #(bs, length, width) --> (bs, length, n_heads, dim_per_head) 28 | x = x.view(bs, length, heads, -1) 29 | # (bs, length, n_heads, dim_per_head) --> (bs, n_heads, length, dim_per_head) 30 | x = x.transpose(1, 2) 31 | # (bs, n_heads, length, dim_per_head) --> (bs*n_heads, length, dim_per_head) 32 | x = x.reshape(bs, heads, length, -1) 33 | return x 34 | 35 | 36 | class PerceiverAttention(nn.Module): 37 | def __init__(self, *, dim, dim_head=64, heads=8): 38 | super().__init__() 39 | self.scale = dim_head**-0.5 40 | self.dim_head = dim_head 41 | self.heads = heads 42 | inner_dim = dim_head * heads 43 | 44 | self.norm1 = nn.LayerNorm(dim) 45 | self.norm2 = nn.LayerNorm(dim) 46 | 47 | self.to_q = nn.Linear(dim, inner_dim, bias=False) 48 | self.to_kv = nn.Linear(dim, inner_dim * 2, bias=False) 49 | self.to_out = nn.Linear(inner_dim, dim, bias=False) 50 | 51 | def forward(self, x, latents): 52 | """ 53 | Args: 54 | x (torch.Tensor): image features 55 | shape (b, n1, D) 56 | latent (torch.Tensor): latent features 57 | shape (b, n2, D) 58 | """ 59 | x = self.norm1(x) 60 | latents = self.norm2(latents) 61 | 62 | b, l, _ = latents.shape 63 | 64 | q = self.to_q(latents) 65 | kv_input = torch.cat((x, latents), dim=-2) 66 | k, v = self.to_kv(kv_input).chunk(2, dim=-1) 67 | 68 | q = reshape_tensor(q, self.heads) 69 | k = reshape_tensor(k, self.heads) 70 | v = reshape_tensor(v, self.heads) 71 | 72 | # attention 73 | scale = 1 / math.sqrt(math.sqrt(self.dim_head)) 74 | weight = (q * scale) @ (k * scale).transpose(-2, -1) # More stable with f16 than dividing afterwards 75 | weight = torch.softmax(weight.float(), dim=-1).type(weight.dtype) 76 | out = weight @ v 77 | 78 | out = out.permute(0, 2, 1, 3).reshape(b, l, -1) 79 | 80 | return self.to_out(out) 81 | 82 | 83 | class Resampler(nn.Module): 84 | def __init__( 85 | self, 86 | dim=1024, 87 | depth=8, 88 | dim_head=64, 89 | heads=16, 90 | num_queries=8, 91 | embedding_dim=768, 92 | output_dim=1024, 93 | ff_mult=4, 94 | ): 95 | super().__init__() 96 | 97 | self.latents = nn.Parameter(torch.randn(1, num_queries, dim) / dim**0.5) 98 | 99 | self.proj_in = nn.Linear(embedding_dim, dim) 100 | 101 | self.proj_out = nn.Linear(dim, output_dim) 102 | self.norm_out = nn.LayerNorm(output_dim) 103 | 104 | self.layers = nn.ModuleList([]) 105 | for _ in range(depth): 106 | self.layers.append( 107 | nn.ModuleList( 108 | [ 109 | PerceiverAttention(dim=dim, dim_head=dim_head, heads=heads), 110 | FeedForward(dim=dim, mult=ff_mult), 111 | ] 112 | ) 113 | ) 114 | 115 | def forward(self, x): 116 | 117 | latents = self.latents.repeat(x.size(0), 1, 1) 118 | 119 | x = self.proj_in(x) 120 | 121 | for attn, ff in self.layers: 122 | latents = attn(x, latents) + latents 123 | latents = ff(latents) + latents 124 | 125 | latents = self.proj_out(latents) 126 | return self.norm_out(latents) 127 | -------------------------------------------------------------------------------- /utils.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) 2025 Bytedance Ltd. and/or its affiliates. All rights reserved. 2 | 3 | # Licensed under the Apache License, Version 2.0 (the "License"); 4 | # you may not use this file except in compliance with the License. 5 | # You may obtain a copy of the License at 6 | 7 | # http://www.apache.org/licenses/LICENSE-2.0 8 | 9 | # Unless required by applicable law or agreed to in writing, software 10 | # distributed under the License is distributed on an "AS IS" BASIS, 11 | # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 12 | # See the License for the specific language governing permissions and 13 | # limitations under the License. 14 | 15 | import torch 16 | import numpy as np 17 | from insightface.utils import face_align 18 | from PIL import Image 19 | import math 20 | import cv2 21 | 22 | def extract_arcface_bgr_embedding(in_image, landmark, arcface_model, in_settings=None): 23 | kps = landmark 24 | arc_face_image = face_align.norm_crop(in_image, landmark=np.array(kps), image_size=112) 25 | arc_face_image = torch.from_numpy(arc_face_image).unsqueeze(0).permute(0,3,1,2) / 255. 26 | arc_face_image = 2 * arc_face_image - 1 27 | arc_face_image = arc_face_image.cuda().contiguous() 28 | face_emb = arcface_model(arc_face_image)[0] # [512], normalized 29 | return face_emb 30 | 31 | def tensor_to_np_image(tensor): 32 | return tensor.mul(255).clamp(0, 255).byte().cpu().numpy() 33 | 34 | def np_image_to_tensor(image): 35 | return torch.clamp(torch.from_numpy(image).float() / 255., 0, 1) 36 | 37 | def resize_and_pad_pil_image(source_img, target_img_size): 38 | # Get original and target sizes 39 | source_img_size = source_img.size 40 | target_width, target_height = target_img_size 41 | 42 | # Determine the new size based on the shorter side of target_img 43 | if target_width <= target_height: 44 | new_width = target_width 45 | new_height = int(target_width * (source_img_size[1] / source_img_size[0])) 46 | else: 47 | new_height = target_height 48 | new_width = int(target_height * (source_img_size[0] / source_img_size[1])) 49 | 50 | # Resize the source image using LANCZOS interpolation for high quality 51 | resized_source_img = source_img.resize((new_width, new_height), Image.LANCZOS) 52 | 53 | # Compute padding to center resized image 54 | pad_left = (target_width - new_width) // 2 55 | pad_top = (target_height - new_height) // 2 56 | 57 | # Create a new image with white background 58 | padded_img = Image.new("RGB", target_img_size, (255, 255, 255)) 59 | padded_img.paste(resized_source_img, (pad_left, pad_top)) 60 | 61 | return padded_img 62 | 63 | # modified from https://github.com/instantX-research/InstantID/blob/main/pipeline_stable_diffusion_xl_instantid.py 64 | def draw_kps(image_pil, kps, color_list=[(255,0,0), (0,255,0), (0,0,255), (255,255,0), (255,0,255)]): 65 | stickwidth = 4 66 | limbSeq = np.array([[0, 2], [1, 2], [3, 2], [4, 2]]) 67 | kps = np.array(kps) 68 | 69 | w, h = image_pil.size 70 | out_img = np.zeros([h, w, 3]) 71 | 72 | for i in range(len(limbSeq)): 73 | index = limbSeq[i] 74 | color = color_list[index[0]] 75 | 76 | x = kps[index][:, 0] 77 | y = kps[index][:, 1] 78 | length = ((x[0] - x[1]) ** 2 + (y[0] - y[1]) ** 2) ** 0.5 79 | angle = math.degrees(math.atan2(y[0] - y[1], x[0] - x[1])) 80 | polygon = cv2.ellipse2Poly((int(np.mean(x)), int(np.mean(y))), (int(length / 2), stickwidth), int(angle), 0, 360, 1) 81 | out_img = cv2.fillConvexPoly(out_img.copy(), polygon, color) 82 | out_img = (out_img * 0.6).astype(np.uint8) 83 | 84 | for idx_kp, kp in enumerate(kps): 85 | color = color_list[idx_kp] 86 | x, y = kp 87 | out_img = cv2.circle(out_img.copy(), (int(x), int(y)), 10, color, -1) 88 | 89 | out_img_pil = Image.fromarray(out_img.astype(np.uint8)) 90 | return out_img_pil 91 | 92 | def escape_path_for_url(path): 93 | return path.replace("\\", "/") 94 | --------------------------------------------------------------------------------