Call/text us anytime to book a tour - (323) 639-7228!
The Intersection
of Gateway and
Getaway.
Comfyui sdxl download
Comfyui sdxl download. json file which is easily loadable into the ComfyUI environment. 71 GB: February 2023: Download Link (opens in a new tab): control_sd15_depth. example to ComfyUI/extra_model_paths. 5 workflows with SD1. Yeah, if I’m being entirely honest, I’m going to download the leak and poke around at it. SDXL Prompt Styler. Features. x, SDXL, LoRA, and upscaling makes ComfyUI flexible. I then This is a comprehensive tutorial on understanding the Basics of ComfyUI for Stable Diffusion. The node specifically replaces a {prompt} placeholder in the 'prompt' field of each template with provided positive text. Save this image then load it or drag it on ComfyUI to get the workflow. LoraInfo Aug 16, 2024 · ComfyUI Impact Pack. Click Queue Prompt to generate an image. Loader SDXL. Keep reading and download AP Workflow 10. LoraInfo The chart above evaluates user preference for SDXL (with and without refinement) over SDXL 0. 9; Install/Upgrade AUTOMATIC1111. Our robust file management capabilities enable easy upload and download of ComfyUI models, nodes, and output results. 1. AUTOMATIC1111 does not have official support for LCM-LoRA yet. There some Custom Nodes utilised so if you get an error, just install the Custom Nodes using ComfyUI Manager. (early and not Update ComfyUI_frontend to 1. 5 and 2. - ltdrdata/ComfyUI-Manager 下載 SDXL 1. SDXL Base in-painting Download the second text encoder from here and place it in ComfyUI/models/t5 - rename it to "mT5-xl. ComfyMath. Nodes that can load & cache Checkpoint, VAE, & LoRA type models. Also, can't find Loras with the names that he has on comfyUI either. Support for SD 1. 6 GB) (8 GB VRAM) (Alternative download link) Put it in ComfyUI > models > checkpoints. 9 and Stable Diffusion 1. Use python entry_with_update. Prerequisites Before you can use this workflow, you need to have ComfyUI installed. How to install ComfyUI. Feb 7, 2024 · To use SDXL, you’ll need to download the two SDXL models and place them in your ComfyUI models folder. 35 in SD1. (Note that the model is called ip_adapter as it is based on the IPAdapter). 5 as I download them from respective repos / branches / whatever those are. 0は、標準で1024×1024ピクセルの画像を生成可能です。 既存のモデルより、光源と影の処理などが改善しており、手や画像中の文字の表現、3次元的な奥行きのある構図などの画像生成aiが苦手とする画像も上手く生成できます。 After you download, you uncompress, use update. The SDXL base model performs significantly better than the previous variants, and the model combined with the refinement module achieves the best overall performance. Generate an image. 首先當然要下載 SDXL 1. Version 4. Furthermore, this extension provides a hub feature and convenience functions to access a wide range of information within ComfyUI. SD 3 Medium (10. 0 is “built on an innovative new architecture composed of a 3. Either manager and install from git, or clone this repo to custom_nodes and run: pip install -r requirements. 40 by @huchenlei in #4691; Add download_path for model downloading progress report. MTB Nodes. com/comfyanonymous/ComfyUIDownload a model https://civitai. Step 5: Start ComfyUI. It can generate high-quality 1024px images in a few steps. Start with strength 0. Merge 2 images together with this ComfyUI workflow. I made a convenient install script that can install the extension and workflow, the python dependencies, and it also offer the option to download the required models. In this guide, I'll use the popular Sytan SDXL workflow and provide a couple of other recommendations. You will discover the principles and techniques behind latent diffusion models, a new class of generative models that can produce high-quality images in seconds. txt. It allows users to apply predefined styling templates stored in JSON files to their prompts effortlessly. safetensors is that the lora? I should then just rename it for SDXL and SD1. bat to update, and use run. 并且comfyui轻量化的特点,使用SDXL模型还能有着更低的显存要求和更快的加载速度,最低支持4G显存的显卡使用。可以说不论是自由度、专业性还是易用性,comfyui在使用SDXL模型上的优势开始越来越明显。 SDXL Default ComfyUI workflow . json file already contains a set of resolutions considered optimal for training in SDXL. Step 4. This was the base for my In this course, you will learn how to use Stable Diffusion, ComfyUI, and SDXL, three powerful and open-source tools that can generate realistic and artistic images from any text prompt. 🧨 Diffusers This guide is tailored towards AUTOMATIC1111 and Invoke AI users, but ComfyUI is also a great choice for SDXL, we’ve published an installation guide for ComfyUI, too! Let’s get started: Step 1: Downloading the SDXL v1. Preview of my workflow – download via the link Hello everybody! I am sure a lot of you saw my post about the workflow I am working with Comfy on for SDXL. ; Following the above, you can load a *. Installation in ForgeUI: 1. Please read the AnimateDiff repo README and Wiki for more information about how it works at its core. Follow these directions if you don't have Efficient Loader & Eff. Inference API The main model can be downloaded from HuggingFace and should be placed into the ComfyUI/models/instantid directory. fp16 Here's how to install and run Stable Diffusion locally using ComfyUI and SDXL. Nov 9, 2023 · Originally shared on GitHub by guoyww. Nodes/graph/flowchart interface to experiment and create complex Stable Diffusion workflows without needing to code anything. Step 3: Load the workflow. Restart ComfyUI to complete the update. Can you let me know how to fix this issue? I have the following arguments: --windows-standalone-build --disable-cuda-malloc --lowvram --fp16-vae --disable-smart-memory Created by: C. Ideal for both beginners and experts in AI image generation and manipulation. Pinto: About SDXL-Lightning is a lightning-fast text-to-image generation model. x and SD2. 7z, select Show More Options > 7-Zip > Extract Here. x, SDXL, Stable Video Diffusion, Stable Cascade, SD3 and Stable Audio. pt" Download/use any SDXL VAE, for example this one; You may also try the following alternate model files for faster loading speed/smaller file The sdxl_resolution_set. segment anything. safetensors - I use the former and rename it to diffusers_sdxl_inpaint_0. Put it in the ComfyUI > models > checkpoints folder. What is ComfyUI? ComfyUI is an easy-to-use interface builder that allows anyone to create, prototype and test web interfaces right from their browser. No-Code Workflow Nov 27, 2023 · Hey guys, I was trying SDXL 1. I tested with different SDXL models and tested without the Lora but the result is always the same. 3. Aug 8, 2023 · さてここまでできたらComfyUIを起動しましょう。ただそのままではSDXLを使えないので、SDXL用のワークフロー(※要するに処理の流れ)を読み込む必要があります。 SDXL用のワークフローは下記ページからダウンロードできます。 Yeah as predicted a while back, I don't think adoption of SDXL will be immediate or complete. Step 3: Clone ComfyUI. Models For the workflow to run you need this loras/models: ByteDance/ SDXL-Lightning 8STep Lora Juggernaut XL Detail Tweaker XL Jun 12, 2024 · LCM-LoRA download pages. About SDXL 1. FromDetailer (SDXL/pipe), BasicPipe -> DetailerPipe (SDXL), Edit DetailerPipe (SDXL) - These are pipe functions used in Detailer for utilizing the refiner model of SDXL. 12 (if in the previous step you see 3. Flux. Fully supports SD1. You switched accounts on another tab or window. Efficiency Nodes for ComfyUI Version 2. For those of you who want to get into ComfyUI's node based interface, in this video we will go over how to in @article{guo2023animatediff, title={AnimateDiff: Animate Your Personalized Text-to-Image Diffusion Models without Specific Tuning}, author={Guo, Yuwei and Yang, Ceyuan and Rao, Anyi and Liang, Zhengyang and Wang, Yaohui and Qiao, Yu and Agrawala, Maneesh and Lin, Dahua and Dai, Bo}, journal={International Conference on Learning Representations}, year={2024} } @article{guo2023sparsectrl, title Jul 29, 2023 · Stability. json file, change your ip-adapter-plus-face_sdxl_vit-h. Feb 23, 2024 · Step 1: Install HomeBrew. Anyline can also be used in SD1. In this ComfyUI Tutorial we'll install ComfyUI and show you how it works. 0 but my laptop with a RTX 3050 Laptop 4GB vRAM was not able to generate in less than 3 minutes, so I spent some time to get a good configuration in ComfyUI, now I get can generate in 55s (batch images) - 70s (new prompt detected) getting a great images after the refiner kicks in. Aug 16, 2024 · ComfyUI Impact Pack. Launch ComfyUI: python main. Download the second text encoder from here and place it in ComfyUI/models/t5 - rename it to "mT5-xl. SDXL Refiner. I’ve been loving SDXL 0. If you're running on Linux, or non-admin account on windows you'll want to ensure /ComfyUI/custom_nodes and comfyui_controlnet_aux has write permissions. safetensors or diffusion_pytorch_model. New installation. No coding required! Is there a limit to how many images I can generate? No, you can generate as many AI images as you want through our site without any limits. 0+ Derfuu_ComfyUI_ModdedNodes. py; That’s it! Share, discover, & run thousands of ComfyUI workflows. Step 3: Install ComfyUI. Sep 3, 2023 · And download diffusion_pytorch_model. A hub dedicated to development and upkeep of the Sytan SDXL workflow for ComfyUI he workflow is provided as a . Once they're installed, restart ComfyUI to enable high-quality previews. Download ComfyUI with this direct download link. c In this ComfyUI Tutorial we'll install ComfyUI and show you how it works. Real-time prompting. If you have a good GPU card, you can do interesting things locally, like real-time prompting on ComfyUI. As evident by the name, this workflow is intended for Stable Diffusion 1. You can also specify inpaint folder in your extra_model_paths. WAS Node Suite. 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. json which has since been edited to use only one image): Aug 18, 2023 · Install controlnet-openpose-sdxl-1. I then recommend enabling Extra Options -> Auto Queue in the interface. Preview of my workflow – download via the link SDXL Examples. However, I kept getting a black image. This LoRA can be used for 1, 2, 4, and 8 sampling steps. 5 (download page) Using LCM-LoRA in AUTOMATIC1111. Download the SDXL base and refiner models from the links given below: SDXL Base. The image-to-image workflow for official FLUX models can be downloaded from the Hugging Face Repository. 5 models Aug 3, 2023 · In this ComfyUI tutorial I show how to install ComfyUI and use it to generate amazing AI generated images with SDXL! ComfyUI is especially useful for SDXL as Aug 4, 2023 · Link to my workflows: https://drive. May 12, 2024 · The Hyper-SDXL team found its model quantitatively better than SDXL Lightning. safetensors files to your models/inpaint folder. bin" Download the model file from here and place it in ComfyUI/checkpoints - rename it to "HunYuanDiT. Step 2: Download SD3 model. Once you’ve downloaded these models, place them in the following directory: ComfyUI_windows_portable\ComfyUI\models\checkpoints. 0 (26 July 2023)! Time to test it out using a no-code GUI called ComfyUI!. Remember at the moment this is only for SDXL. UltimateSDUpscale. 3 GB! Place it in the ComfyUI models\unet folder. SEGS Manipulation nodes SEGSDetailer - Performs detailed work on SEGS without pasting it back onto the original image. Download a stable diffusion model. ai has released Stable Diffusion XL (SDXL) 1. 0. You signed out in another tab or window. Download a VAE: Download a Variational Autoencoder like Latent Diffusion’s v-1-4 VAE and place it in the “models/vae” folder. 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 File 也比 1. 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 generate images. Here is a workflow for using it: Save this image then load it or drag it on ComfyUI to get the workflow. It will update ComfyUI itself and all custom nodes installed. 5 Template Workflows for ComfyUI which is a multi-purpose workflow that comes with three templates. LoraInfo Aug 20, 2023 · First, download clip_vision_g. Wait while the script downloads the latest version of ComfyUI Windows Portable, along with all the latest required custom nodes and extensions OPTION 1: Once the script has finished, rename your ComfyUI/extra_model_paths. Re-start ComfyUI. yaml. 5 的大得多。 SDXL Examples. py --preset anime or python entry_with_update. 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. Then press “Queue Prompt” once and start writing your prompt. Discover More From Me:🛠️ Explore hundreds of AI Tools: https://futuretools. LoraInfo ComfyUI-Manager is an extension designed to enhance the usability of ComfyUI. This smoothens your workflow and ensures your projects and files are well-organized, enhancing your overall experience. Based on the revision-image_mixing_example. The SDXL base checkpoint can be used like any regular checkpoint in ComfyUI. 6B parameter refiner Jun 17, 2024 · Select Manager > Update ComfyUI. 5's ControlNet, although it generally performs better in the Anyline+MistoLine setup within the SDXL workflow. Step 2: Install a few required packages. bat is important, otherwise you may be using a previous version with potential bugs unfixed. The Tutorial covers:1. json') Able to apply LoRA & Control Net stacks via their lora_stack and cnet_stack inputs. Aug 29, 2024 · SDXL Examples. It offers management functions to install, remove, disable, and enable various custom nodes of ComfyUI. But at the same time, I’m obviously accepting the possibility of bugs and breakages when I download a leak. You can see all Hyper-SDXL and Hyper-SD models and the corresponding ComfyUI workflows. 0, it can add more contrast through offset-noise) Feb 23, 2024 · Step 2: Download the standalone version of ComfyUI. bat" file) or into ComfyUI root folder if you use ComfyUI Portable How To Install ComfyUI And The ComfyUI Manager. 0の特徴. Dec 19, 2023 · ComfyUI won't take as much time to set up as you might expect. Aug 27, 2023 · SDXL Prompt Styler is a node that enables you to style prompts based on predefined templates stored in multiple JSON files. Contribute to comfyorg/comfyui-ipadapter development by creating an account on GitHub. 1 GB) (12 GB VRAM) (Alternative download link) SD 3 Medium without T5XXL (5. py --preset realistic for Fooocus Anime/Realistic Edition. I am only going to list the models that I found useful below. bin: same as ip-adapter-plus_sdxl_vit-h, but use cropped face image as condition; Downloads last month 0. A collection of SDXL workflow templates for use with Comfy UI - Suzie1/Comfyroll-SDXL-Workflow-Templates Nov 13, 2023 · Searge-SDXL: EVOLVED v4. x is here. c Aug 6, 2023 · Then, download the SDXL VAE: SDXL VAE; LEGACY: If you're interested in comparing the models, you can also download the SDXL v0. ComfyUI Manager: Plugin for CompfyUI that helps detect and install missing plugins. File Name Size Update Time 说明 Download Links; control_sd15_canny. x, SDXL and Stable Video Diffusion. com/drive/folders/1C4hnb__HQB2Pkig9pH7NWxQ05LJYBd7D?usp=drive_linkStable Diffusion XL did not run quite well on my Video tutorial on how to use ComfyUI, a powerful and modular Stable Diffusion GUI and backend, is here. How to update ComfyUI. or if you use portable (run this in ComfyUI_windows_portable -folder): Feb 22, 2024 · You signed in with another tab or window. Next) root folder (where you have "webui-user. safetensors from the control-lora/revision folder and place it in the ComfyUI models\clip_vision folder. fp16. You should place diffusion_pytorch_model. Here are the Huggingface links to the LCM-LoRA weights. Feb 1, 2024 · The first one on the list is the SD1. 9 models: sd_xl_base_0. For more information check ByteDance paper: SDXL-Lightning: Progressive Adversarial Diffusion Distillation . x, 2. rgthree's ComfyUI Nodes. Share, discover, & run thousands of ComfyUI workflows. Sytan SDXL ComfyUI: Very nice workflow showing how to connect the base model with the refiner and include an upscaler. ↑ Node setup 1: Generates image and then upscales it with USDU (Save portrait to your PC and then drag and drop it into you ComfyUI interface and replace prompt with your's, press "Queue Prompt") Aug 29, 2024 · The LCM SDXL lora can be downloaded from here (opens in a new tab) Download it, rename it to: lcm_lora_sdxl. 0, this one has been fixed to work in fp16 and should fix the issue with generating black images) (optional) download SDXL Offset Noise LoRA (50 MB) and copy it into ComfyUI/models/loras (the example lora that was released alongside SDXL 1. 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: Aug 27, 2024 · First of all, to work with the respective workflow you must update your ComfyUI from the ComfyUI Manager by clicking on "Update ComfyUI". pth (for SDXL) models and place them in the models/vae_approx folder. Here is the link to download the official SDXL turbo checkpoint open in new window. Instead of creating a workflow from scratch, you can simply download a workflow optimized for SDXL v1. Follow the following update steps if you want to update ComfyUI or the custom nodes independently. Users can drag and drop nodes to design advanced AI art pipelines, and also take advantage of libraries of existing workflows. yaml and update it to point to your models Aug 12, 2023 · Last update 08-12-2023 本記事について 概要 ComfyUIはStable Diffusionモデルから画像を生成する、Webブラウザベースのツールです。最近ではSDXLモデルでの生成速度の早さ、消費VRAM量の少なさ(1304x768の生成時で6GB程度)から注目を浴びています。 本記事では手動でインストールを行い、SDXLモデルで画像 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. tinyterraNodes. Stable Diffusion v1. To enable higher-quality previews with TAESD, download the taesd_decoder. Step 1: Download SDXL Turbo checkpoint. Download the SD3 model. How to use Hyper-SDXL in ComfyUI. And we have Thibaud Zamora to thank for providing us such a trained model! Head over to HuggingFace and download OpenPoseXL2. This will avoid any errors. import torch from diffusers import FluxPipeline from huggingface_hub import hf_hub_download base_model_id = "black-forest-labs Hyper-SDXL-1step-Unet-Comfyui. safetensors from the controlnet-openpose-sdxl-1. json , the general workflow idea is as follows (I digress: yesterday this workflow was named revision-basic_example. AUTOMATIC1111 can run SDXL as long as you upgrade to the newest version. google. There is now a install. Download the SDXL Turbo model. 8 and boost 0. sdxl 1. 10 or for Python 3. ComfyUI's ControlNet Auxiliary Preprocessors. Join the largest ComfyUI community. Download Stable Diffusion models: Download the latest Stable Diffusion model checkpoints (ckpt files) and place them in the “models/checkpoints” folder. 0の概要 (1) sdxl 1. Simply download the . Mar 18, 2024 · The SDXL Prompt Styler is a versatile custom node within Comfy UI that streamlines the prompt styling process. 9 on ClipDrop, and this will be even better with img2img and ControlNet. Reload to refresh your session. Note that running update. Feb 24, 2024 · ComfyUI is a node-based interface to use Stable Diffusion which was created by comfyanonymous in 2023. Choose the version that aligns with the version your desired model was based on. - shingo1228/ComfyUI-SDXL-EmptyLatentImage Dec 5, 2023 · The workflow uses SVD + SDXL model combined with LCM LoRA which you can download (Latent Consistency Model (LCM) SDXL and LCM LoRAs) and use it to create animated GIFs or Video outputs. 2. (cache settings found in config file 'node_settings. 12) and put into the stable-diffusion-webui (A1111 or SD. 0 Model Files We would like to show you a description here but the site won’t allow us. After digging that huggingface blog I only found one bigger file in Files And Vesions tab called pytorch_lora_weights. 6 boost 0. io/ Created by: OpenArt: What this workflow does This basic workflow runs the base SDXL model with some optimization for SDXL. ComfyUI ControlNet aux: Plugin with preprocessors for ControlNet, so you can generate images directly from ComfyUI. ThinkDiffusion - SDXL_Default. 9; sd_xl_refiner_0. This can be useful for systems with limited resources as the refiner takes another 6GB or ram. bat to run. Step 3: Update ComfyUI Step 4: Launch ComfyUI and enable Auto Queue (Under Extra Options) Step 5: Drag and drog and sample image into ConfyUI Step 6: The FUN begins! If queue didn't start automatically, press Queue Prompt Anyline, in combination with the Mistoline ControlNet model, forms a complete SDXL workflow, maximizing precise control and harnessing the generative capabilities of the SDXL model. Jul 6, 2024 · The simplest way to update ComfyUI is to click the Update All button in ComfyUI manager. 5B parameter base model and a 6. safetensors and put it in your ComfyUI/models/loras directory. Masquerade Nodes. Download the SDXL Turbo Model. Comfyroll Studio. Even with great fine tunes, control net, and other tools, the sheer computational power required will price many out of the market, and even with top hardware, the 3x compute time will frustrate the rest sufficiently that they'll have to strike a personal balance between wait time and quality. Apr 11, 2024 · segmentation_mask_brushnet_ckpt and random_mask_brushnet_ckpt contains BrushNet for SD 1. bat you can run to install to portable if detected. Anyline, in combination with the Mistoline ControlNet model, forms a complete SDXL workflow, maximizing precise control and harnessing the generative capabilities of the SDXL model. Download the workflow JSON file below and drop it in inpaunt工作流. A good place to start if you have no idea how any of this works is the: ComfyUI Basic Tutorial VN: All the art is made with ComfyUI. ComfyUI provides a powerful yet intuitive way to harness Stable Diffusion through a flowchart interface. You also needs a controlnet, place it in the ComfyUI controlnet directory. 2. json file during node initialization, allowing you to save custom resolution settings in a separate file. . ComfyUI https://github. Learn about how to run this model to create animated images on GitHub. Here is the link to download the official SDXL turbo checkpoint. Nov 30, 2023 · Step 3. To update ComfyUI: Click Manager (instead of using the VAE that's embedded in SDXL 1. ComfyUI . pth (for SD1. Step 2: Download this sample Image. x) and taesdxl_decoder. Install ForgeUI if you have not yet. The SDXL base checkpoint can be used like any regular checkpoint in ComfyUI (opens in a new tab). 11) or for Python 3. But you can use the LCM-LoRA speed up in a limited way. ControlNet-LLLite-ComfyUI. Asynchronous Queue system. There have been several new things added to it, and I am still rigorously testing, but I did receive direct permission from Joe Penna himself to go ahead and release information. The Face Detailer and Object Swapper functions are now reconfigured to use the new SDXL I am trying out using SDXL in ComfyUI. LCM-LoRA for SDXL (download page) LCM-LoRA for Stable Diffision v1. 5 models while segmentation_mask_brushnet_ckpt_sdxl_v0 and random_mask_brushnet_ckpt_sdxl_v0 for SDXL. Hyper-SDXL 1-step LoRA. 14 GB compared to the latter, which is 10. We provide unlimited free generation. 3 in SDXL and 0. When it is done, right-click on the file ComfyUI_windows_portable_nvidia_cu118_or_cpu. 0 repository, under Files and versions; Place the file in the ComfyUI folder models\controlnet. x, SD2. The default installation includes a fast latent preview method that's low-resolution. 5 models and is a very beginner-friendly workflow allowing anyone to use it easily. pth: 5. Content. 5 This detailed guide provides step-by-step instructions on how to download and import models for ComfyUI, a powerful tool for AI image generation. In essence, choosing RunComfy for running ComfyUI equates to opting for speed, convenience, and efficiency. Explain the Ba Contribute to cubiq/ComfyUI_IPAdapter_plus development by creating an account on GitHub. Dec 5, 2023 · The workflow uses SVD + SDXL model combined with LCM LoRA which you can download (Latent Consistency Model (LCM) SDXL and LCM LoRAs) and use it to create animated GIFs or Video outputs. Understand the differences between various versions of Stable Diffusion and learn how to choose the right model for your needs. 71 GB: February 2023 Download prebuilt Insightface package for Python 3. 11 (if in the previous step you see 3. safetensors, because it is 5. Jul 30, 2023 · 1.sdxl 1. 5 Improved AnimateDiff integration for ComfyUI, as well as advanced sampling options dubbed Evolved Sampling usable outside of AnimateDiff. pt" Download/use any SDXL VAE, for example this one; You may also try the following alternate model files for faster loading speed/smaller file An extension node for ComfyUI that allows you to select a resolution from the pre-defined json files and output a Latent Image. by @robinjhuang in #4621; Cleanup empty dir if frontend zip download failed by @huchenlei in #4574; Support weight padding on diff weight patch by @huchenlei in #4576; fix: useless loop & potential undefined variable by @ltdrdata The first time you run Fooocus, it will automatically download the Stable Diffusion SDXL models and will take a significant amount of time, depending on your internet connection. For example: 896x1152 or 1536x640 are good resolutions. Per the announcement, SDXL 1. 0 for ComfyUI.
xaupu
bjip
qoho
sqvchr
nmukygx
wvfkswj
xdqw
mqf
touciah
zozggisw