Sdxl comfy ui workflow A lower 次に2つ目のメリットは、SDXLのrefinerモデルを既に正式にサポートしている点です。 執筆時点ではStable Diffusion web UIのほうはrefinerモデルにまだ完全に対応していないのですが、ComfyUIは既にSDXLに対応済みで簡単にrefinerモデルを使うことができます。 動作が I modified a simple workflow to include the freshly released Controlnet Canny. To use ReVision, you must enable it in the “Functions” section. The workflow files and examples are from the ComfyUI Blog. © Civitai 2025. json: Text-to-image workflow for SDXL Turbo This workflow is just something fun I put together while testing SDXL models and LoRAs that made some cool picture so I am sharing it here. x, SDXL, Stable Video Diffusion, Stable Cascade, SD3 and Stable Audio, Flux. If your model is based on SD 1. 4. ai/workflows/openart/basic-sdxl-workflow SDXL Ultimate Workflow is the best and most complete single workflow that exists for SDXL 1. ComfyUI was created by comfyanonymous, who made the tool to understand how Stable Diffusion works. ComfyUI Academy. Sort by: You might be able to add in another LORA through a Loader but i Understanding SDXL Model Types SDXL comes in several variants: Base SDXL 1. to the corresponding Comfy folders, as discussed in ComfyUI manual ワークフロー（Workflow）とは？ ワークフローは ComfyUI の最も重要な概念です。 簡単に言えば、複数のノード（Node）を接続して構成されるグラフィカルインターフェースで、AI画像生成の全処理過程を表現します。 Hello there and thanks for checking out this workflow! Disclaimer: Glitch-vid workflow sold separately; batteries not included; no refunds — Purpose — This workflow provides a simple, solid and reliable way to efficiently inpaint images to p e r f e c t i o n!. json. To begin we remove the default layout to make room for our personalized workflow. As of writing of this it is in its beta phase, but I am sure some are eager to test Comfy Deploy. Here is an example workflow that can be dragged or loaded into The interface for using this new workflow is also designed in a different way, with all parameters that are usually tweaked to generate images tightly packed together. Here is my way of merging BASE models and applying LORAs to them in non-conflicting way using the ComfyUI (grab the workflow itself in the attachment to this article): All the images in this repo contain metadata which means they can be loaded into ComfyUI with the Load button (or dragged onto the window) to get the full workflow that was used to create the image. 0 Base SDXL 1. It is made by the same people who made the SD 1. They can be used with any SDXL SDXL Ultimate Workflow is the best and most complete single workflow that exists for SDXL 1. My Workflows. Part 3 (this post) - we will add an SDXL refiner for the full SDXL process This ComfyUI workflow is here to help! This workflow will assist you in converting simple interior sketches into realistic room images with ease. Thanks in advance. Employing Controlnet for enhanced image control and precision Comfy UI is the most powerful and modular stable diffusion GUI and backend. 5B parameter base model and a 6. Part 2 - we added SDXL-specific conditioning implementation + tested the impact of conditioning parameters on the generated images. AP Workflow 4. Useful for bigger labelling of nodes, helpful for wider screen captures or Thanks for the tips on Comfy! I'm enjoying it a lot so far. 9 I was using some ComfyUI workflow shared here where the refiner was always an improved version versus the base. All (20) Img2img Text2img Upscale (2) Inpaint Lora ControlNet Embeddings Model merging Sdxl Cascade UnCLIP Hypernetworks 3d Video Lcm Turbo. Install ControlNet-aux custom nodes;. 0 and SD 1. Workflow Templates. 0. However, for the best results, it’s important to set the resolution to 1024×1024 or other resolutions with an equivalent pixel count but a different aspect ratio. if you don't want that then i will send you two workflows one with upscale and other with no upscale Make your images more random in a This SDXL workflow allows you to create images with the SDXL base model and the refiner and adds a LoRA to the image generation. x, SD2. ICU. ComfyUI SDXL Turbo Workflow. 在最左侧的这两个load checkpoint模块分别添加base模型和refiner模型。 I have updated the workflow submitted last week, cleaning up a bit the layout and adding many functions I wanted to learn better. s to generate . If you have the SDXL 1. There are two CLIP positive Stability. Le checkpoint de base SDXL peut être utilisé comme n’importe quel checkpoint régulier dans ComfyUI. 3/Run the step 2 Workflow as many times as you need- you need to input the location of the original frames and the dimensions as before. Fully supports SD1. 2. Lesson 3: Latent Upscaling in ComfyUI - Comfy Academy; View all 11 lessons. 0 the refiner is Discovery, share and run thousands of ComfyUI Workflows on OpenArt. 0 is “built on an innovative new architecture composed of a 3. 1. In part 1 , we implemented the simplest SDXL Base workflow and generated our first images. 6B parameter refiner The base model generates (noisy) latent, which are then further processed Preview of my workflow – download via the link below. If you’d like a way to make someone un-comfy perhaps by making the lips of their face move with different dialogue, check out my post using Wav2Lip. 3 Exemples SDXL. 5 model files Workflow is in the attachment json file in the top right. SDXL LCM LoRA SVD Workflow (31661 downloads ) Select the image you want to animate, define the SDXL dimensions you want eg. ; Link the Created by: Adel AI: This approach uses the merging technique to convert the used model into its inpaint version, as well as the new InpaintModelConditioning node (You need to update ComfyUI and Manager). This step sets the foundational framework for incorporating LoRA. Custom nodes for easier use of SDXL in ComfyUI including an img2img workflow that utilizes both the base and refiner checkpoints. Following Workflows In part 1 (this post), we will implement the simplest SDXL Base workflow and generate our first images. 5. The proper way to use it is with the new SDTurboScheduler node but it might also work with the regular schedulers. Low-Rank Adaptation (LoRA) is a method of fine tuning the SDXL model with additional training, and is implemented via a a small “patch” to the model, without having to re-build the model from scratch. About SDXL 1. Important elements include loading checkpoints using SDXL and loading the VAE. Simply Comfy. This section will introduce the installation of the official version models and the download of workflow files. All Workflows / Morph SDXL. The SDXL base checkpoint can be used like any regular checkpoint in ComfyUI. An all-in-one workflow that supports various tasks like txt2img, img2img, and inpainting. This SDXL Examples. SDXL Turbo 示例. I then recommend enabling Extra Options -> Auto Queue in the interface. A good place to start if you have no idea how any of this works is the: It'll load a basic SDXL workflow that includes a bunch of notes explaining things. 0 Refiner Automatic calculation of the steps required for both the Base and the Refiner models Quick selection of image width and height based on the SDXL training set SDXL Workflow. Inpainting a cat with the v2 inpainting model: Inpainting a woman with the v2 inpainting model: It also works with non inpainting models. I hope this proves to be a useful tool to you! — Features — Custom built group nodes with focus on efficiency Created by: OpenArt: What this workflow does This workflows is a very simple workflow to use IPAdapter IP-Adapter is an effective and lightweight adapter to achieve image prompt capability for stable diffusion models. Per the announcement, SDXL 1. Now in Comfy, from the Img2img workflow, let’s duplicate Load Image and Upscale Image Nodes. Some key capabilities of the Comfy Workflow include: Generating high-quality images using SDXL, Kolors, and Flux samplers. It's since What this workflow does. 2. The prompt word will affect the combination effect(Image link to WD14 in this workflow). Any model, any VAE, any LoRAs. It incorporates Nodes/graph/flowchart interface to experiment and create complex Stable Diffusion workflows without needing to code anything. What it's great for: This is a great starting point to These workflow templates are intended as multi-purpose templates for use on a wide variety of projects. You signed out in another tab or window. This repository contains a handful of SDXL workflows I use, make sure to check the usefull links as some of these models, and/or plugins are required to use these in ComfyUI. Morph SDXL. This a workflow to fix hands. How to use. Sure, it's not 2. If you choise SDXL model, make sure to load appropriate SDXL ControlNet model. This resource has been removed by its owner. Go to OpenArt main site. As an alternative to the SDXL Base+Refiner models, or the Base/Fine-Tuned SDXL model, you can generate images with the ReVision method. MimicPC simplifies this by Then move it to the “\ComfyUI\models\controlnet” folder. Contribute to fabiomb/Comfy-Workflow-sdxl development by creating an account on GitHub. Join the largest ComfyUI community. ai has released Stable Diffusion XL (SDXL) 1. You signed in with another tab or window. Download hand_yolo_8s model and put it in 💡Education. Comfy Academy; 9:23. 0: The standard model offering excellent image quality; SDXL Turbo: Optimized for speed with slightly lower quality; SDXL Lightning: A balanced option between speed and quality; Eg. Download the . Nobody needs all that, LOL. 0 - What is an official workflow for Comfy UI? It was a secret before. No, because it's not there yet. You need the model from here, put it in comfyUI (yourpath\ComfyUI\models\controlnet), and you are ready to go: 1. Comfy Summit Workflows (Los Angeles, US & Shenzhen, China) Challenges. 官方的SDXL工作流的使用介绍. I am using a base SDXL Zavychroma as my base model then using Juggernaut In this guide, we'll set up SDXL v1. I am using a base SDXL Zavychroma as my base model then using Juggernaut Lightning to stylize the image . It can generate variants in a similar style based on the input image without the need for text prompts. With SDXL 0. Following Workflows. By default, the workflow is setup to create 25 frames and create a 6 frame per second (FPS) GIF. IPAdapter can be bypassed. This workflow adds a refiner model on topic of the basic SDXL workflow ( https://openart. Run ComfyUI workflows online and deploy APIs with one click In ComfyUI the saved checkpoints contain the full workflow used to generate them so they can be loaded in the UI just like images to get the full workflow that was used to create them. Image Models LoRAs etc. Contents text_to_image. pth and . を参照）を別途適用するよう、ワークフローを修正しました。変更はvaeの接続のみです。 例えば、良いと思う構図で出力された場合は、その画像をuiにドラッグ＆ドロップしてからseedを固定に 約5GBの「diffusion_pytorch_model. New. And it doesn't just work for images, it also has a good effect on SVD models. Img2img. 5 Original FP16 Version ComfyUI Workflow. 5. It's since become the In this guide, we'll set up SDXL v1. Flux Redux is an adapter model specifically designed for generating image variants. FaceID SDXL: The SDXL version of FaceID, maintaining the same InsightFace model as the v1. It somewhat works. ai/workflows/openart/basic-sdxl-workflow/P8VEtDSQGYf4pOugtnvO). 0 with the node-based Stable Diffusion user interface ComfyUI. You switched accounts on another tab or window. It also has full inpainting SDXL 示例. Download Download Comfy UI, the most powerful and modular stable diffusion GUI and 0-1. Run and deploy your ComfyUI workflows. safetensors」などに名前を変えて、「\models\controlnet」のフォル Then you can load this image in ComfyUI to get the workflow that shows how to use the LCM SDXL lora with the SDXL base model: The important parts are to use a low cfg, use the “lcm” sampler and the “sgm_uniform” or “simple” im working on basic SDXL workflow Also you need to install an upscaling node to create 4K images. SDXL most definitely doesn't work with the old control net. Install Impact Pack custom nodes;. Creators Created by: KoalaNation: Tips Workflow development and tutorials not only take part of my time, but also consume resources. A collection of workflow templates for use with Comfy UI These workflow templates are intended as multi-purpose templates for use on a wide variety of projects. Study this workflow and notes to understand the basics of ComfyUI, SDXL, and Refiner workflow. 5 Model Files. This workflow use the Impact-Pack and the Reactor-Node. The SDXL base checkpoint functions like any standard checkpoint in ComfyUI. AnimateDiff for SDXL is a motion module which is used with SDXL to create animations. Created by: CgTopTips: ControlNet++: All-in-one ControlNet for image generations and editing! The controlnet-union-sdxl-1. As evident by the name, this workflow is intended for Stable Diffusion 1. Next, integrate the LoRA node into your workflow: Position the Node: Place the LoRA node between the diffusion model and the CLIP nodes in your workflow. 3. Workflow Templates Liked Workflows. Download Stable Diffusion 3. 4. (more advanced UI elements right next to the main UI) Best (simple) SDXL Inpaint Workflow. Run ComfyUI workflows online and deploy APIs with one click. Contest Winners. Liked Workflows. Make your images more random in a good way or more precise wit this Comfy nodes! 6. To enhance SDXL Turbo is a SDXL model that can generate consistent images in a single step. ICU Serverless cloud for running ComfyUI workflows with an API Searge-SDXL: EVOLVED. 24 KB. Another special thanks to PseudoTerminalX, Caith, ThrottleKitty, ComfyAnonymous, HumbleMikey, CaptnSeraph, and Joe Penna for the support and help working on this project. once you download the file drag and drop it into ComfyUI and it will populate the workflow. Workflow Initialization. The Comfy Online. 5 models and is a very beginner-friendly workflow allowing anyone to The following images can be loaded in ComfyUI to get the full workflow. This should make it easier to have every important element on the screen at the same time without scrolling. 首先當然要下載 SDXL 1. The Ultimate SD upscale is one of the nicest things in Auto11, it first upscales your image using GAN or any other old school upscaler, then cuts it into tiles small enough to be digestable by SD, typically 512x512, the pieces are overlapping each other and can be bigger. SDXL Turbo是一个SDXL模型，能够在单步中生成一致的图像。您可以使用更多步骤来提高质量。使用它的正确方式是与新的SDTurboScheduler节点一起使用，但它也可能与常规调度器一起工作。 这是下载官方SDXL Turbo检查点的链接：官方SDXL Turbo检查点 A simple combination of PuLID and IPAdapter StyleTransfer. It's a super simple custom node for Comfy UI, to generate text, with a font size option. Part 2 - (coming in 48 hours) we will add SDXL-specific conditioning implementation + test what impact that About LoRAs. Reload to refresh your session. 0 Created by: CgTopTips: Since the specific ControlNet model for FLUX has not been released yet, we can use a trick to utilize the SDXL ControlNet models in FLUX, which will help you achieve almost what you want. 5 Best AI UI Design Tools Download ControlNet Openpose model (both . Use ComfyUI This workflow is just something fun I put together while testing SDXL models and LoRAs that made some cool picture so I am sharing it here. Hidden Faces. Open ComfyUI and navigate to the "Clear" button. Here is an example of how to create a CosXL model from a regular SDXL model Thanks to the creators of ComfyUI for creating a flexible and powerful UI. It also has full inpainting If you want the ultimate SDXL workflow for ComfyUI then look no further than Searge-SDXL: EVOLVED which is by far the most advanced workflow for ComfyUI. 我们导入官方SDXL的标准流程，我在每个板块在旁边添加了中文备注，解释了每个板块的用途。接下来我演示一下如何使用SDXL模型在comfyui上的工作流. it will change the image into an animated video using Animate-Diff and ip adapter in ComfyUI. The only important thing is that for optimal performance the resolution should be set to 1024x1024 or other resolutions with the same amount of pixels but a different aspect ratio. Introduction. Model conversion Created by: Reverent Elusarca: This workflow uses SDXL or SD 1. Upload workflow. Here is the link to download the official SDXL turbo checkpoint. You can experiment between Flux and SDXL (along with Lora 8 Steps ) with the click of Step 5: Connect the LoRA Node. 5 model as a base image generations, using ControlNet Pose and IPAdapter for style. It now includes: SDXL 1. 0_fp16. You also need to go to the comfy output and find the blendframes folder and input the location of that in here too. The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. By integrating it with tools like SD, SDXL & Flux ControlNet, it can convert simple sketches into high-quality images, providing creative flexibility and artistic enhancement to your work. 0 Base and Refiners models downloaded and saved in the right place, it should work out of the box. Once inside the ComfyUI environment, start by loading the default workflow. Provide a source picture and a face and the workflow will do the rest. Detailed Tutorial on Flux Redux Workflow. All Workflows. My stuff. Here’s an example with the anythingV3 model: Outpainting. Home. 0 includes the following advanced functions: ReVision. I tried to find a good Inpaint workflow and just found a bunch of wild workflows that wanted a million nodes and had a bunch of different functions. - comfyanonymous/ComfyUI Nodes/graph/flowchart interface to experiment and create complex Stable Diffusion workflows without needing to code anything. How to use this workflow. 0 (26 July 2023)! Time to test it out using a no-code GUI called ComfyUI!. Created by: CgTopTips: The Painter Node in ComfyUI transforms your drawings into stunning artworks using AI models. Resources. Comfy Online. This basic workflow runs the base SDXL model with some optimization for SDXL. json file to import the workflow. SDXLがリリースされた時にStable Diffusion Web UIより速く対応し、話題になっていたのがComfyUIです。 UIの外観で敬遠してしまう方も多いと思いますが、自分でカスタマイズするというよりはワークフローを取り込ん Share, discover, & run thousands of ComfyUI workflows. yaml files), and put it into "\comfy\ComfyUI\models\controlnet"; Download and open this workflow. Connect the upscale node’s Stable Diffusion 3. Upscale. Leaderboard. Make sure to adjust prompts accordingly. ThinkDiffusion - SDXL_Default. Then press “Queue Prompt” once and This repo contains the workflows and Gradio UI from the "How to Use SDXL Turbo in Comfy UI for Fast Image Generation" video tutorial. How to install. You can use more steps to increase the quality. Comfy. Our journey starts by setting up a workflow. 5 時大了足足一整倍，而且訓練數據也增加了3倍，所以最終出來的 Checkpoint Comfy. This is a ComfyUI workflow to swap faces from an image. ComfyUI 工作流基础概念以及如何使用 什么是工作流（Workflow）？ 工作流是 ComfyUI 中最核心的概念，简单来说就是由多个节点（Node）连接组成的图形化界面，用来描述 AI 绘图的整个处理过程。 The workflow also includes advanced features such as Controlnet, various optimization tools, and the ability to upscale your images with ease. This can be useful for systems with limited resources as the refiner takes another 6GB or ram. How to use this workflow The IPAdapter model has to match the CLIP vision encoder and of course the main checkpoint. OpenArt Workflows. The AnimateDiff node integrates model and context options to adjust animation dynamics. A beginner-friendly workflow that simplifies the process of using Stable Diffusion by allowing the loading of a single checkpoint with minimal settings, perfect for those new to ComfyUI. Your ControlNet pose reference image should be like in this workflow. 5 but scaled for Ah, ComfyUI SDXL model merging for AI-generated art! That's exciting! Merging different Stable Diffusion models opens up a vast playground for creative exploration. Navigate to the "Load" button. Download Required Files The first one on the list is the SD1. Created by: 358 op: NVIDIA released a giant cowhide project Align Your Steps a few days ago, which can greatly improve the effect of SD low inference steps to generate images. 2 Pass Txt2Img. ComfyUI already supports this algorithm natively, and it works pretty well after trying it for 10 steps. Upscale to unlimited resolution using SDXL Tile regardless with no VRAM limitations. This workflow also contains 2 up scaler workflows. Select the downloaded . This workflow creates two outputs with two different sets of settings. safetensors」をダウンロードして、他のControlNetのモデルと区別するために、例えば、「controlnet-canny-sdxl-1. 5 models. safetensors model is a combined model that integrates several ControlNet models, saving you from having to download each model individually, such as canny, lineart, depth, and others. Instead of SDXL 1. 5 Template Workflows for ComfyUI which is a multi-purpose workflow that comes with three templates. Here is a workflow for using it: Save this image then load it or drag it on ComfyUI to get the workflow. safetensors”. This ComfyUI workflow is designed for creating animations from reference images by using AnimateDiff and IP-Adapter. The quality of SDXL Turbo is relatively good, though it may not always be stable. This step is crucial because it establishes the foundation of our workflow ensuring we have all the tools to us. La seule chose importante est que pour une performance optimale, la résolution doit être réglée à 1024x1024 ou à d’autres résolutions avec le même nombre de pixels mais un rapport d’aspect différent. 0 的 Checkpoint Model，由於 SDXL 在訓練時圖片用上了 1024 x 1024 的圖片，解像度比 SD 1. It has many upscaling options, such as img2img upscaling and Ultimate SD Upscale upscaling. Write prompt and negative for background. example of the variants: Installation Steps 1. Discover the Ultimate Workflow with ComfyUI in this hands-on tutorial, where I guide you through integrating custom nodes, refining images with advanced tool Q: What is the optimal size for SDXL conversions? A: The optimal size for SDXL conversions is identified as 1024, which is the recommended train size for achieving the best results. Since Flux doesn't support ControlNet and IPAdapte yet, this is the current method. They can be used with any SDLX checkpoint model. Img2Img. The only important thing is that for optimal performance the resolution should be set to 1024x1024 or other resolutions Workflow for ComfyUI and SDXL 1. Share Add a Comment. attached is a workflow for ComfyUI to convert an image into a video. にて追記しています。 ※ (12-1-2024) sdxlのvae（追加した1-5. SDXL Turbo synthesizes image outputs in a single step and generates real-time text-to-image outputs. Does anyone have a workflow I can have to use loras with SDXL? All workflows I have found don't have the option of adding loras. SDXL can indeed generate a nude body, and the model itself doesn't stop you from fine-tuning it towards whatever spicy stuff there is with a dataset, at least by the looks of it. 1316 x 832px which will be the dimensions for the final animated video. Q: How can I adjust the level of transformation in the image-to-image process? A: The level of transformation can be adjusted using the denoise parameter. json file from this repository. Text2img. 🛟 Support 下載 SDXL 1. But now in SDXL 1. We name the file “canny-sdxl-1. If you want to play with Step 3: Load the Default Workflow in ComfyUI. 1. . 5, use In this guide, we are aiming to collect a list of 10 cool ComfyUI workflows that you can simply download and try out for yourself. SDXL基础检查点可以在 ComfyUI 中像任何常规检查点一样使用。唯一重要的是，为了获得最佳性能，分辨率应设置为1024x1024或具有相同像素数量但不同宽高比的其他分辨率。 例如：896x1152或1536x640是好的分辨率。 SDXL Examples. You can also use similar workflows for outpainting. SDXL FaceID Models. Workflow for ComfyUI and SDXL 1. All SD15 models and all models ending ComfyUI は、画像生成AIである Stable Diffusionを操作するためのツールの一つ です。 特に、ノードベースのUIを採用しており、さまざまなパーツをつなぐことで画像生成の流れを制御します。 Stable Diffusionの画像生成web UIとしては、AUTOMATIC1111が有名ですが、 「ComfyUI」はSDXLへの対応の速さや、低 This is a basic img2img workflow on top of our basic SDXL workflow ( https://openart. bhyy fidq tbncybx wpt ghfxtyl wwyeg mxssw kvmes fckzvm twehew qmck uhleol dcn gmu hck