Support for SD 1. Link to my workflows:super easy to do inpainting in the Stable Diffu. Space Composition and Inpainting: ComfyUI supplies space composition and inpainting options with regular and inpainting fashions, considerably boosting image enhancing abilities. . Provides a browser UI for generating images from text prompts and images. ComfyUI: Area Composition or Outpainting? Area Composition: I couldn't get this to work without making the images look like they are stretched specially for landscape long-width-wise images, faster run time wrt atleast to Out painting. Hello, recent comfyUI adopter looking for help with facedetailer or an alternative. For example, you can remove or replace: Power lines and other obstructions. please let me know. 222 added a new inpaint preprocessor: inpaint_only+lama. VAE Encode (for Inpainting) Transform Transform Crop Latent Flip Latent Rotate Latent Loaders. Second thoughts, heres. Stable Diffusion XL (SDXL) 1. For this editor we've integrated Jack Qiao's excellent custom inpainting model from the glid-3-xl-sd project instead. As long as you're running the latest ControlNet and models, the inpainting method should just work. These are examples demonstrating how to do img2img. You can Load these images in ComfyUI to get the full workflow. The text was updated successfully, but these errors were encountered: All reactions. New Features. The pixel images to be upscaled. In part 1 (this post), we will implement the simplest SDXL Base workflow and generate our first images. ComfyShop has been introduced to the ComfyI2I family. ，Comfyui-提示词自动翻译插件来了，告别复制来复制去！，ComfyUI+Roop单张照片换脸，comfyUI使用者神器！comfyUI插件节点使用者册推荐！，整理并总结了B站和C站上现有ComfyUI的相关视频和插件。仍然是学什么和在哪学的省流讲解。Use the "Set Latent Noise Mask" and a lower denoise value in the KSampler, after that you need the "ImageCompositeMasked" to paste the inpainted masked area into the original image, because the VAEEncode don't keep all the details of the original image, that is the equivalent process of the A1111 inpainting, and for better results around the mask you. Notably, it contains a " Mask by Text " node that allows dynamic creation of a mask. Launch ComfyUI by running python main. The settings I used are. Support for FreeU has been added and is included in the v4. "Want to master inpainting in ComfyUI and make your AI Images pop? 🎨 Join me in this video where I'll take you through not just. This means the inpainting is often going to be significantly compromized as it has nothing to go off and uses none of the original image as a clue for generating an adjusted area. aiimag. Part 5: Scale and Composite Latents with SDXL. So I would probably try three of those nodes in sequence, with original conditioning going to the outer two, and your controlnet conditioning going to the middle sampler, then you might be able to add steps. VAE Encode (for Inpainting) Transform Transform Crop Latent Flip Latent Rotate Latent Loaders. 25:01 How to install and use ComfyUI on a free. Creating an inpaint mask. Join. Assuming ComfyUI is already working, then all you need are two more dependencies. something of an advantage comfyUI has over other interfaces is that the user has full control over every step of the process which allows you to load and unload models, images and use stuff entirely in latent space if you want. Troubleshootings: Occasionally, when a new parameter is created in an update, the values of nodes created in the previous version can be shifted to different fields. Imagine that ComfyUI is a factory that produces an image. yaml conda activate hft. an alternative is Impact packs detailer node which can do upscaled inpainting to give you more resolution but this can easily end up giving you more detail than the rest of. yeah ps will work fine, just cut out the image to transparent where you want to inpaint and load it as a separate image as mask. New Features. In researching InPainting using SDXL 1. fills the mask with random unrelated stuff. The SDXL base checkpoint can be used like any regular checkpoint in ComfyUI. If you uncheck and hide a layer, it will be excluded from the inpainting process. okolenmion Sep 1. This feature combines img2img, inpainting and outpainting in a single convenient digital artist-optimized user interface. Using Controlnet with Inpainting models Question | Help Is it possible to use ControlNet with inpainting models? Whenever I try to use them together, the ControlNet component seems to be ignored. As an alternative to the automatic installation, you can install it manually or use an existing installation. Please share your tips, tricks, and workflows for using this software to create your AI art. 0 with ComfyUI. It has an almost uncanny ability. This in-depth tutorial will guide you to set up repositories, prepare datasets, optimize training parameters, and leverage techniques like LoRA and inpainting to achieve photorealistic results. . HELP WITH "LoRa" in XL (colab) r/comfyui. . Prior to adoption I generated an image in A1111, auto-detected and masked the face, inpainted the face only (not whole image), which improved the face rendering 99% of the time. best place to start is here. ComfyUI . • 2 mo. Replace supported tags (with quotation marks) Reload webui to refresh workflows. Now you slap on a new photo to inpaint. Inpaint + Controlnet Workflow. Another neat trick you can do with. All improvements are made INTERMEDIATELY in this one workflow. . Thanks in advanced. Support for FreeU has been added and is included in the v4. Info. . Stable Diffusion Inpainting is a latent text-to-image diffusion model capable of generating photo-realistic images given any text input, with the extra capability of inpainting the pictures by using a mask. The RunwayML Inpainting Model v1. Inpainting with both regular and inpainting models. Inpainting a cat with the v2 inpainting model: Inpainting a woman with the v2 inpainting model: It also works with non inpainting models. Also if you want better quality inpaint I would recommend the impactpack SEGSdetailer node. right. Use the paintbrush tool to create a mask over the area you want to regenerate. Where people create machine learning projects. continue to run the process. Note that these custom nodes cannot be installed together – it’s one or the other. 5 by default, and usually this value works quite well. I'm finding that I have no idea how to make this work with the inpainting workflow I am used to using in Automatic1111. The only way to use Inpainting model in ComfyUI right now is to use "VAE Encode (for inpainting)", however, this only works correctly with the denoising value of 1. Inpainting Process. 0-inpainting-0. Area Composition Examples | ComfyUI_examples (comfyanonymous. ComfyUI has an official tutorial in the. It works pretty well in my tests within the limits of. Especially Latent Images can be used in very creative ways. Where people create machine learning projects. Dust spots and scratches. Sytan SDXL ComfyUI: Very nice workflow showing how to connect the base model with the refiner and include an upscaler. lite stable nightly Info - Token - Model Page; stable_diffusion_comfyui_colab CompVis/stable-diffusion-v-1-4-original: waifu_diffusion_comfyui_colabIf you're running on Linux, or non-admin account on windows you'll want to ensure /ComfyUI/custom_nodes, ComfyUI_I2I, and ComfyI2I. Img2Img works by loading an image like this example image, converting it to latent space with the VAE and then sampling on it with a denoise lower than 1. Images can be uploaded by starting the file dialog or by dropping an image onto the node. This approach is more technically challenging but also allows for unprecedented flexibility. Inpainting with both regular and inpainting models. masquerade nodes are awesome, I use some of them. • 3 mo. Yes, you can add the mask yourself, but the inpainting would still be done with the amount of pixels that are currently in the masked area. When comparing openOutpaint and ComfyUI you can also consider the following projects: stable-diffusion-ui - Easiest 1-click way to install and use Stable Diffusion on your computer. Copy the update-v3. 9模型下载和上传云空间. Inpaint area: Only masked. Loaders GLIGEN Loader Hypernetwork Loader. Contribute to LiuFengHuiXueYYY/ComfyUi development by creating an account on GitHub. addandsubtract • 7 mo. Step 1: Create an inpaint mask; Step 2: Open the inpainting workflow; Step 3: Upload the image; Step 4: Adjust parameters; Step 5:. This node based UI can do a lot more than you might think. The Load Image (as Mask) node can be used to load a channel of an image to use as a mask. 3. 0 with an inpainting model. Is there any way to fix this issue? And is the "inpainting"-version really so much better than the standard 1. . Outpainting: SD-infinity, auto-sd-krita extension. Realistic Vision V6. 1 of the workflow, to use FreeU load the newInpainting. The Pad Image for Outpainting node can be used to to add padding to an image for outpainting. Place the models you downloaded in the previous. Mask is a pixel image that indicates which parts of the input image are missing or. Can anyone add the ability to use the new enhanced inpainting method to ComfyUI which is discussed here Mikubill/sd-webui-controlnet#1464. I'm a newbie to ComfyUI and I'm loving it so far. Master the power of the ComfyUI User Interface! From beginner to advanced levels, this guide will help you navigate the complex node system with ease. I change probably 85% of the image with latent nothing and inpainting models 1. A tutorial that covers some of the processes and techniques used for making art in SD but specific for how to do them in comfyUI using 3rd party programs in. . You can Load these images in ComfyUI to get the full workflow. I reused my original prompt most of the time but edited it when it came to redoing the. I've seen a lot of comments about people having trouble with inpainting and some saying that inpainting is useless. I have about a decade of blender node experience, so I figured that this would be a perfect match for me. 0 based on the effect you want) 3. Question about Detailer (from ComfyUI Impact pack) for inpainting hands. AI, is designed for text-based image creation. Contribute to camenduru/comfyui-colab by creating an account on DagsHub. json file for inpainting or outpainting. Loaders GLIGEN Loader Hypernetwork Loader Load CLIP Load CLIP Vision Load Checkpoint Load ControlNet Model. If you're happy with your inpainting without using any of the controlnet methods to condition your request then you don't need to use it. 5 inpainting model, and separately processing it (with different prompts) by both SDXL base and refiner models:. Join. Maybe I am doing it wrong, but ComfyUI inpainting is a bit awkward to use. ComfyUI Manager: Plugin for CompfyUI that helps detect and install missing plugins. Txt2Img is achieved by passing an empty image to the sampler node with maximum denoise. First we create a mask on a pixel image, then encode it into a latent image. Inpainting models are only for inpaint and outpaint, not txt2img or mixing. bat to update and or install all of you needed dependencies. Although the 'inpaint' function is still in the development phase, the results from the 'outpaint' function remain quite. ComfyUI系统性. i think, its hard to tell what you think is wrong. It looks like I need at least 6GB VRAM to pass VAE Encode (for inpainting) step on 1920*1080 image. Right click menu to add/remove/swap layers. this will open the live painting thing you are looking for. Optional: Custom ComfyUI Server. 5 is a specialized version of Stable Diffusion v1. Here's how the flow looks rn: Yeah, I stole adopted most of it from some example on inpainting a face. ではここからComfyUIの基本的な使い方についてご説明していきます。 ComfyUIは他のツールとは画面の使い方がかなり違う ので最初は少し戸惑うかもしれませんが、慣れればとても便利なのでぜひマスターしてみてください。Launch ComfyUI by running python main. lordpuddingcup. Methods overview "Naive" inpaint : The most basic workflow just masks an area and generates new content for it. Other features include embeddings/textual inversion, area composition, inpainting with both regular and inpainting models, ControlNet and T2I-Adapter, upscale models, unCLIP models, and more. Fernicles SDTools V3 - ComfyUI nodes. r/comfyui. inpainting. You have to draw a mask, save the image with the mask, then upload to the UI again to inpaint. 2 workflow. The Set Latent Noise Mask node can be used to add a mask to the latent images for inpainting. Not hidden in a sub menu. Here you can find the documentation for InvokeAI's various features. Part 6: SDXL 1. 17:38 How to use inpainting with SDXL with ComfyUI. In addition to a whole image inpainting and mask only inpainting, I also have workflows that upscale the masked region to do an inpaint and then downscale it back to the original resolution when pasting it back in. The core idea behind IA is. Chaos Reactor: a community & Open Source modular tool for synthetic media creators. 8. Note that if force_inpaint is turned off, inpainting might not occur due to the guide_size. strength is normalized before mixing multiple noise predictions from the diffusion model. Video tutorial on how to use ComfyUI, a powerful and modular Stable Diffusion GUI and backend, is here . Open a command line window in the custom_nodes directory. Improved AnimateDiff integration for ComfyUI, initially adapted from sd-webui-animatediff but changed greatly since then. Stable Diffusion will redraw the masked area based on your prompt. A GIMP plugin that makes it a facility for ComfyUI. ControlNet and T2I-Adapter; Upscale Models (ESRGAN, ESRGAN variants, SwinIR, Swin2SR, etc. 0. The order of LORA and IPadapter seems to be crucial: Workflow: Time KSampler only: 17s IPadapter->KSampler: 20s LORA->KSampler: 21s Optional: Custom ComfyUI Server. There is a latent workflow and a pixel space ESRGAN workflow in the examples. start sampling at 20 Steps. If you installed from a zip file. safetensors node, And the model output is wired up to the KSampler node instead of using the model output from the previous CheckpointLoaderSimple node. Stability. Text prompt: "a teddy bear on a bench". 0 for ComfyUI. 4K views 2 months ago ComfyUI. Unlike other Stable Diffusion tools that have basic text fields where you enter values and information for generating an image, a node-based interface is different in the sense that you’d have to create nodes to build a workflow to. 3K Members. The area of the mask can be increased using grow_mask_by to provide the inpainting process with some. Using a remote server is also possible this way. So, there is a lot of value of allowing us to use Inpainting model with "Set Latent Noise Mask". Inpainting (with auto-generated transparency masks). Two of the most popular repos. If for some reason you cannot install missing nodes with the Comfyui manager, here are the nodes used in this workflow: ComfyLiterals, Masquerade Nodes, Efficiency Nodes for ComfyUI, pfaeff-comfyui, MTB Nodes. Workflow examples can be found on the Examples page. add a 'load mask' node, and add an vae for inpainting node, plug the mask into that. Inpainting is very effective in Stable Diffusion and the workflow in ComfyUI is really simple. A recent change in ComfyUI conflicted with my implementation of inpainting, this is now fixed and inpainting should work again. Done! FAQ. Works fully offline: will never download anything. CUI can do a batch of 4 and stay within the 12 GB. This is a node pack for ComfyUI, primarily dealing with masks. An example of Inpainting+Controlnet from the controlnet. Yet, it’s ComfyUI. Inpainting a cat with the v2 inpainting model: Inpainting a woman with the v2 inpainting model: It also works with non inpainting models. (custom node) 2. 20:57 How to use LoRAs with SDXL. In this endeavor, I've employed the Impact Pack extension and Con. I have found that the inpainting check point actually without any problems, however just as a single model, there are a couple that did not. edit: this was my fault, updating comfyui, isnt a bad idea i guess. Inpainting can be a very useful tool for. Upload the image to the inpainting canvas. Just an FYI. Interface NodeOptions Save File Formatting Shortcuts Text Prompts Utility Nodes. MoonMoon82on May 2. inpainting, and model mixing all within a single UI. Part 2 - (coming in 48 hours) we will add SDXL-specific conditioning implementation + test what impact that conditioning has on the generated images. 24:47 Where is the ComfyUI support channel. img2img → inpaint, open the script and set the parameters as follows: 23. ago. The extracted folder will be called ComfyUI_windows_portable. height. This is the answer, we need to wait for controlnetXL comfyUI nodes, and then a whole new world opens up. Run update-v3. Since a few days there is IP-Adapter and a corresponding ComfyUI node which allow to guide SD via images rather than text. Please keep posted images SFW. Please share your tips, tricks, and workflows for using this software to create your AI art. For some workflow examples and see what ComfyUI can do you can check out: ComfyUI Examples FeaturesUse LatentKeyframe and TimestampKeyframe from ComfyUI-Advanced-ControlNet to apply diffrent weights for each latent index. Controlnet + img2img workflow. Enjoy a comfortable and intuitive painting app. crop. es: free, easy to install windows program. UI changes Ready to take your image editing skills to the next level? Join me in this journey as we uncover the most mind-blowing inpainting techniques you won't believ. 0 comfyui ControlNet and img2img working alright but inpainting seems like it doesn't even listen to my prompt 8/9 times. 完成ComfyUI界面汉化，并新增ZHO主题配色 ，代码详见：ComfyUI 简体中文版界面 ; 完成ComfyUI Manager汉化 ，代码详见：ComfyUI Manager 简体中文版 . ago. Colab Notebook:. AnimateDiff ComfyUI. upscale_method. 🦙 LaMa: Resolution-robust Large Mask Inpainting with Fourier Convolutions. Outputs will not be saved. bat file to the same directory as your ComfyUI installation. It's just another control net, this one is trained to fill in masked parts of images. ai & PPA Master Professional PhotographerGreetings! I am the lead QA at Stability. amount to pad above the image. A node suite for ComfyUI with many new nodes, such as image processing, text processing, and more. 5 based model and then do it. It looks like this: For AMD (Linux only) or Mac, check the beginner's guide to ComfyUI. Contribute to camenduru/comfyui-colab by creating an account on DagsHub. SDXL-Inpainting. Diffusion Bee: MacOS UI for SD. ComfyUI promises to be an invaluable tool in your creative path, regardless of whether you’re an experienced professional or an inquisitive newbie. ago • Edited 1 yr. はStable Diffusionを簡単に使えるツールに関する話題で 便利なノードベースのウェブUI「ComfyUI」のインストール方法や使い方 を一通りまとめてみるという内容になっています。 Stable Diffusionを. From inpainting, which allows you to make internal edits, to outpainting for extending the canvas, and image-to-image transformations, the platform is designed for flexibility. Yes, you would. Masquerade Nodes. It's also available as a standalone UI (still needs access to Automatic1111 API though). Within the factory there are a variety of machines that do various things to create a complete image, just like you might have multiple machines in a factory that produces cars. Capable of blending blurs but hard to use to enhance quality of objects as there's a tendency for the preprocessor to erase portions of the object instead. (ComfyUI, A1111) - the name (reference) of an great photographer or. The method used for resizing. This project strives to positively impact the domain of AI-driven. edit your mannequin image in photopea to superpose the hand you are using as a pose model to the hand you are fixing in the editet image. . The inpaint + Lama preprocessor doesn't show up. It offers artists all of the available Stable Diffusion generation modes (Text To Image, Image To Image, Inpainting, and Outpainting) as a single unified workflow. I decided to do a short tutorial about how I use it. workflows " directory and replace tags. This is for anyone that wants to make complex workflows with SD or that wants to learn more how SD works. (early and not finished) Here are some more advanced examples: "Hires Fix" aka 2 Pass Txt2Img. Launch ComfyUI by running python main. Users can drag and drop nodes to design advanced AI art pipelines, and also take advantage of libraries of existing workflows. Trying to encourage you to keep moving forward. 5 i thought that the inpanting controlnet was much more useful than the inpaining fine-tuned models. ai is your go-to platform for discovering and comparing the best AI tools. Change your prompt to describe the dress and when you generate a new image it will only change the masked parts. Inpainting large images in comfyui. ComfyUI A powerful and modular stable diffusion GUI and backend. Part 4: Two Text Prompts (Text Encoders) in SDXL 1. Explanation. If your end goal is generating pictures (e. Use the paintbrush tool to create a mask. So in this workflow each of them will run on your input image and you. 5 gives me consistently amazing results (better that trying to convert a regular model to inpainting through controlnet, by the way). The Stable-Diffusion-Inpainting was initialized with the weights of the Stable-Diffusion-v-1-2. Here’s a basic example of how you might code this using a hypothetical inpaint function: In simple terms, inpainting is an image editing process that involves masking a select area and then having Stable Diffusion redraw the area based on user input. json file. Just enter your text prompt, and see the generated image. 6. When the noise mask is set a sampler node will only operate on the masked area. For inpainting, I adjusted the denoise as needed and reused the model, steps, and sampler that I used in txt2img. We will inpaint both the right arm and the face at the same time. ComfyUI ControlNet aux: Plugin with preprocessors for ControlNet, so you can generate images directly from ComfyUI. 0 to create AI artwork. Adjust the value slightly or change the seed to get a different generation. CLIPSeg Plugin for ComfyUI. Then, the output is passed to the inpainting XL pipeline which uses the refiner model to convert the image into a compatible latent format for the final pipeline. Available at HF and Civitai. 1. Examples shown here will also often make use of these helpful sets of nodes: Follow the ComfyUI manual installation instructions for Windows and Linux. by Roman Suvorov, Elizaveta Logacheva, Anton Mashikhin, Anastasia Remizova, Arsenii Ashukha, Aleksei Silvestrov, Naejin Kong, Harshith Goka, Kiwoong Park, Victor Lempitsky. workflows" directory. Here I modified it from the official ComfyUI site, just a simple effort to make it fit perfectly on a 16:9 monitor. It would be great if there was a simple tidy UI workflow the ComfyUI for SDXL. by default images will be uploaded to the input folder of ComfyUI. Normal models work, but they dont't integrate as nicely in the picture. This looks like someone inpainted at full resolution. pip install -U transformers pip install -U accelerate. While it can do regular txt2img and img2img, it really shines when filling in missing regions. I find the results interesting for comparison; hopefully others will too. . If the server is already running locally before starting Krita, the plugin will automatically try to connect. Outpainting just uses a normal model. Stable Diffusion is an AI model able to generate images from text instructions written in natural language (text-to-image. Sadly, I can't use inpaint on images 1. SDXL 1. Barbie play! To achieve this effect, follow these steps: install ddetailer in the extention tab. And that means we can not use underlying image(e. I have a workflow that works. Requirements: WAS Suit [Text List, Text Concatenate] : ( Master Tutorial - Stable Diffusion XL (SDXL) - Install On PC, Google Colab (Free) & RunPod, SDXL LoRA, SDXL InPainting. r/StableDiffusion. Show more. VAE Encode (for Inpainting) Transform Transform Crop Latent Flip Latent Rotate Latent Loaders. Navigate to your ComfyUI/custom_nodes/ directory. g. ComfyUI Community Manual Getting Started Interface. Note that if force_inpaint is turned off, inpainting might not occur due to the guide_size. , Stable Diffusion) fill the "hole" according to the text. For this I used RPGv4 inpainting. MultiLatentComposite 1. The node-based workflow builder makes it. Loaders GLIGEN Loader Hypernetwork Loader. Jattoe. With SD 1. The idea here is th. Original v1 description: After a lot of tests I'm finally releasing my mix model. A series of tutorials about fundamental comfyUI skills This tutorial covers masking, inpainting and image. Inpainting is very effective in Stable Diffusion and the workflow in ComfyUI is really simple. 0. (early and not finished) Here are some more advanced examples: “Hires Fix” aka 2 Pass Txt2Img. AITool. Extract the zip file. In the case of ComfyUI and Stable Diffusion, you have a few different "machines," or nodes. When comparing ComfyUI and stable-diffusion-webui you can also consider the following projects: stable-diffusion-ui - Easiest 1-click way to install and use Stable Diffusion on your computer. Simply download this file and extract it with 7-Zip. This node decodes latents in tiles allowing it to decode larger latent images than the regular VAE Decode node. Euchale asked this question in Q&A. Also come with a ConditioningUpscale node. Maybe I am using it wrong so I have a few questions: When using ControlNet Inpaint (Inpaint_only+lama, ControlNet is more important) should I use an inpaint model or a normal one. What Auto1111 does with "only masked" inpainting is it inpaints the masked area at the resolution you set (so 1024x1024 for examples) and then it downscales it back to stitch it into the picture. io) Also it can be very diffcult to get. 0. comfyUI采用的是workflow体系来运行Stable Diffusion的各种模型和参数，有点类似于桌面软件. 2 workflow. lowering the denoising settings simply shifts the output towards the neutral grey that replaces the masked area. workflows " directory and replace tags. And another general difference is that A1111 when you set 20 steps 0. We've curated some example workflows for you to get started with Workflows in InvokeAI. Take the image out to a 1. As a backend, ComfyUI has some advantages over Auto1111 at the moment, but it never implemented the image-guided ControlNet mode (as far as I know), and results with just regular inpaint ControlNet are not good enough. Setting the crop_factor to 1 considers only the masked area for inpainting, while increasing the crop_factor incorporates context relative to the mask for inpainting. Here’s an example with the anythingV3 model: Outpainting. InvokeAI Architecture. 35 or so. This might be useful for example in batch processing with inpainting so you don't have to manually mask every image. 0 has been out for just a few weeks now, and already we're getting even more SDXL 1. Therefore, unless dealing with small areas like facial enhancements, it's recommended. @taabata There. If you want to do. Results are generally better with fine-tuned models. load your image to be inpainted into the mask node then right click on it and go to edit mask. Forgot to mention, you will have to download this inpaint model from huggingface and put it in your comfyUI "Unet" folder that can be found in the models folder. MultiLatentComposite 1. If you caught the stability. 4 or. In the case of ComfyUI and Stable Diffusion, you have a few different "machines," or nodes. I use nodes from Comfyui-Impact-Pack to automatically segment image, detect hands, create masks and inpaint. PS内直接跑图，模型可自由控制！. I desire: Img2img + Inpaint workflow. But these improvements do come at a cost; SDXL 1. thibaud_xl_openpose also. LaMa: Resolution-robust Large Mask Inpainting with Fourier Convolutions (Apache-2. Inpainting or other method? I found that none of the checkpoints know what a "eye monocle" is, they also struggle with "cigar" I wondered what the best way to get the dude with the eye monocle in this. Config file to set the search paths for models. If a single mask is provided, all the latents in the batch will use this mask. Inpainting relies on a mask to determine which regions of an image to fill in; the area to inpaint is represented by white pixels. 17:38 How to use inpainting with SDXL with ComfyUI. Part 1: Stable Diffusion SDXL 1. With this plugin, you'll be able to take advantage of ComfyUI's best features while working on a canvas. Trying to use b/w image to make impaintings - it is not working at all. ComfyUI - コーディング不要なノードベースUIでStable Diffusionワークフローを構築し実験可能なオープンソースインターフェイス！ControlNET、T2I、Lora、Img2Img、Inpainting、Outpaintingなどもサポート.