Comfyui workflow download sdxl

Comfyui workflow download sdxl. Merging 2 Images together. It now includes: SDXL 1. But it separates LORA to another workflow (and it's not based on SDXL either). I've been doing some tests in A1111 using the Ultimate Upscaler script together with Controlnet Tile and it works wonderful it doesn't matter what tile size or image resolution I throw at it but in ComfyUI I get this error: --------------. Select an upscaler and click Queue Prompt to generate an upscaled image. It is based on the SDXL 0. パラーメータ Oct 10, 2023 · Same for anyone with ComfyUI vs Automatic1111. I recommend you do not use the same text encoders as 1. Real-time prompting Dec 19, 2023 · One of the best parts about ComfyUI is how easy it is to download and swap between workflows. Add LoRAs or set each LoRA to Off and None. json file which is easily loadable into the ComfyUI environment. ComfyUI provides a powerful yet intuitive way to harness Stable Diffusion through a flowchart interface. 🥈84 12:00. I have attached a TXT2VID and VID2VID workflow that works with my 12GB VRAM card. ComfyUIでStable Diffusion XLを動かす方法. You switched accounts on another tab or window. json: High-res fix workflow to upscale SDXL Turbo images; app. If the term "workflow" is something that has only been used exclusively to describe ComfyUI's node graphs, I suggest just calling them "node graphs" or just "nodes". Workflow features: RealVisXL V3. SDXL Prompt Styler is a node that enables you to style prompts based on predefined templates stored in multiple JSON files. Custom nodes used in this workflow. Only the LCM Sampler extension is needed, as shown in this video. Using LoRA's (A workflow to use LoRA's in your generations) View Now. PeterL1n Update ComfyUI workflow. First, you need to download the SDXL model: SDXL 1. Consider using the 2-step model for much better quality. 0 Base SDXL 1. 1. SD and SDXL and Loras models are supported. go to img2img, choose batch, dropdown refiner, use the folder in 1 as input and the folder in 2 as output. 9 facedetailer workflow by FitCorder, but rearranged and spaced out more, with some additions such as Lora Loaders, VAE loader, 1:1 previews, Super upscale with Remacri to over 10,000x6000 in just 20 seconds with Torch2 & SDP. New installation. Welcome to the unofficial ComfyUI subreddit. Users can drag and drop nodes to design advanced AI art pipelines, and also take advantage of libraries of existing workflows. Hidden Faces (A workflow to create hidden faces and text) View Now. they are also recommended for users coming from Auto1111. LEv145_. 6 and 1. ComfyUI workflow to play with this, embedded here: Oct 9, 2023 · Crisp and beautiful images with relatively short creation time, easy to use. safetensors, and save it to comfyui/controlnet. x, SDXL, LoRA, and upscaling makes ComfyUI flexible. Set the CFG scale between 0. That's the one I'm referring to. safetensors; Another LoRA, download additionals LoRA from CivitAI or select "None" to not use it. This workflow perfectly works with 1660 Super 6Gb VRAM. Starting workflow. diffusers/stable-diffusion-xl-1. Generate unique and creative images from text with OpenArt, the powerful AI image creation tool. The image should have been upscaled 4x by the AI upscaler. If you installed via git clone before. ago. The workflow first generates an image from your given prompts and then uses that image to create a video. SDXL-Lightning / comfyui / sdxl_lightning_workflow_full. they will also be more stable with changes deployed less often. json') Able to apply LoRA & Control Net stacks via their lora_stack and cnet_stack inputs. I'm looking for comfyUI workflow (s) that do image generation with sdxl and loras. 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 Feb 22, 2024 · Use my workflow and it have the model listed you only will have to download it, the workflow is embed in the last image, Im using 8 steps but 4 and two give good results Beta Was this translation helpful? Oct 14, 2023 · This is a ComfyUI workflow to swap faces from an image. Download models from lllyasviel/fooocus_inpaint to ComfyUI/models/inpaint. Nov 25, 2023 · Inpainting workflow (A great starting point for using Inpainting) View Now. 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. Set the aspect ratio of the video for output. 0 model, the highly anticipated addition was the support for the ControlNet function. VAE selector, (needs a VAE file, download SDXL BF16 VAE from here, and VAE file for SD 1. pth (for SDXL) models and place them in the models/vae_approx folder. json file（点击load加载）. Cog packages machine learning models as standard containers. Now with controlnet, hires fix and a switchable face detailer. Enter your main image's positive/negative prompt and any styling. SDXLモデルのダウンロード. Then in Part 3, we will implement the SDXL refiner. Oct 12, 2023 · A and B Template Versions. 0 most robust ComfyUI workflow. safetensors and put it in your ComfyUI/models/loras directory. Set the filename_prefix in Save Checkpoint. OpenPose SDXL: OpenPose ControlNet for SDXL. json file. Author. youtube. There some Custom Nodes utilised so if you get an error, just install the Custom Nodes using ComfyUI Manager. Upscaling ComfyUI workflow. 以前、ノードベースのWeb UIである「 ComfyUI 」という画像生成ツールをご紹介しました。. ComfyUI ControlNet aux: Plugin with preprocessors for ControlNet, so you can generate images directly from ComfyUI. 24 KB. Workflow. Jan 17, 2024 · 6. Instantly replace your image's background. Buy Me A Coffee. Step 3. Prerequisites Before you can use this workflow, you need to have ComfyUI installed. Aug 6, 2023 — 8 min read. Perfect for artists, designers, and anyone who wants to create stunning visuals without any design experience. There is an Article here explaining how to install Aug 4, 2023 · Find the Workflow here: https://civitai. 基本的な手順は以下4つです。. Is it safe? 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. To load a workflow, simply click the Load button on the right sidebar, and select the workflow . Edit: I realized that the workflow loads just fine, but the prompts are sometimes not as expected. The default installation includes a fast latent preview method that's low-resolution. this will be the prefix for the output model. Reload to refresh your session. Only Youtube videos that the owners allow to be ebedded in other websites Aug 19, 2023 · Select the models and VAE. Enable Real-Time Prompting. . If you want to play with parameters, I advice you to take a look on the following from the Face Detailer as they are those that do the best for my generations : SDXL cliptext node used on left, but default on right sdxl-clip vs default clip. x) and taesdxl_decoder. 2. For now, I have to manually copy the right prompts. Open a command line window in the custom_nodes directory. Brace yourself as we delve deep into a treasure trove of fea Aug 2, 2023 · This is my current SDXL 1. Mar 20, 2024 · To use this upscaler workflow, you must download an upscaler model from the Upscaler Wiki, and put it in the folder models > upscale_models. Provide a source picture and a face and the workflow will do the rest. 1) This is a basic outpainting workflow that incorporates ideas from the following videos: ComfyUI x Fooocus Inpainting & Outpainting (SDXL) by Data Leveling. 1. You can use the popular Sytan SDXL workflow or any other existing ComfyUI workflow with SDXL. Nov 30, 2023 · Drag and drop the workflow image file to ComfyUI to load the workflow. The workflow posted here relies heavily on useless third-party nodes from unknown extensions. 我也在多日測試後，決定暫時轉投 ComfyUI。. Select an upscale model. Finally made a workflow for ComfyUI to do img2img with SDXL Workflow Included Share Add a Comment. In contrast, the SDXL-clip driven image on the left, has much greater complexity of composition. Select the downloaded JSON file to import the workflow. The default value is 1 person. x and SD2. To enable higher-quality previews with TAESD, download the taesd_decoder. Replace your image's background with the newly generated backgrounds and composite the primary subject/object onto your images. We don't know if ComfyUI will be the tool moving forwards but what Apr 22, 2024 · The examples directory has workflow examples. Interactive SAM Detector (Clipspace) - When you right-click on a node that has 'MASK' and 'IMAGE' outputs, a context menu will open. 在 ComfyUI 上使用 SDXL 1. Once they're installed, restart ComfyUI to enable high-quality previews. 5 for final work. The workflow is provided as a . Jul 29, 2023 · About this version. Unpack the SeargeSDXL folder from the latest release into ComfyUI/custom_nodes, overwrite existing files. Download the SDXL Turbo Model. When using the SDXL base model I find the refiner helps improve images, but I don't run it for anywhere close to the number of steps that the official workflow does. You signed in with another tab or window. You can run it as an img2img batch in Auto1111: generate a bunch of txt2img using base. 5 aspect ratio Node. EDIT: For example this workflow shows the use of the other prompt windows. 5 Lora with SDXL, Upscaling Future tutorials planned: Prompting practices, post processing images, batch trickery, networking comfyUI in your home network, Masking and clipseg awesomeness, many more. 0_0. On the left-hand side of the newly added sampler, we left-click on the model slot and drag it on the canvas. safetensors. the templates produce good results quite easily. This workflow use the Impact-Pack and the Reactor-Node. Navigate to your ComfyUI/custom_nodes/ directory. I mean, the image on the right looks "nice" and all. Loader SDXL. A hub dedicated to development and upkeep of the Sytan SDXL workflow for ComfyUI he workflow is provided as a . ComfyUIのインストール. 5 and 2. Los modelos los tienes que descargar y añadir tú por tu cuenta. In Part 1, we implement the SDXL base in the simplest way possible. Render. B-templates. This makes it super easy to download a clip of a few seconds from anywhere in a youtube video for local use and processing. 今回も Stable Diffusion XL（SDXL） に関する話題で、タイトルの通り. Please share your tips, tricks, and workflows for using this software to create your AI art. Aspect Ratio and Resolution These are the standard image ratios recalculated to pixels. Think Diffusion's Stable Diffusion ComfyUI Top 10 Cool Workflows. A-templates. SD. Go to this link and download the JSON file by clicking the button labeled ComfyUI . Aug 12, 2023 · After the Stability AI team officially released the SDXL 1. 3. Making Videos with AnimateDiff-XL. It provides workflow for SDXL (base + refiner). Here is the workflow with full SDXL: Start off with the usual SDXL workflow SDXL-controlnet: OpenPose (v2) Comfy Workflow (Image is from ComfyUI, you can drag and drop in Comfy to use it as workflow) License: refers to the OpenPose's one. Opening the image in stable-diffusion-webui's PNG-info I can see that there are indeed two different sets of prompts in that file and for some reason the wrong one is being chosen. Sort by: Experimental LCM Workflow "The Ravens" for Würstchen v3 aka Stable Cascade is up and ready for download. The full article on CosXL can be found here. But for a base to start at it'll work. In researching InPainting using SDXL 1. ワークフローの読み込み. ControlNet Depth ComfyUI workflow. What it's great for: This Warning. SDXL 1. Update your ComfyUI to the latest version. 5 and SDXL models. A method of Out Painting In ComfyUI by Rob Adams. Click on the "Queue Prompt" button to run the workflow. This is an implementation of the ComfyUI text2img workflow as a Cog model. Try DPM++ 2S a Karras, DPM++ SDE Karras, DPM++ 2M Karras, Euler a and DPM adaptive. pth (for SD1. The SDXL base checkpoint can be used like any regular checkpoint in ComfyUI. Feb 9, 2024 · Alternatively, you can also find the SDXL workflow for ComfyUI here. com/WASasquatch/was-node-suit Feb 19, 2024 · However, CosXL models require a ComfyUI workflow-based user interface to function. • 9 mo. 0. Jul 14, 2023 · ComfyUIのSDXLのフルポテンシャルを引き出す日本語バージョンのワークフローです。 これはComfyUI SDXL ワークフローで、ComfyUIユーザーにとってより使いやすいようにできるだけシンプル且つ全てのポテンシャルを活かせるように設計しました。 Downloading SDXL pics posted here on reddit and dropping them into comfyUI doesn't work either so I guess will need a direct download link comments sorted by Best Top New Controversial Q&A Add a Comment text_to_image. Dec 19, 2023 · Place the models you downloaded in the previous step in the folder: ComfyUI_windows_portable\ComfyUI\models\checkpoints; If you downloaded the upscaler, place it in the folder: ComfyUI_windows_portable\ComfyUI\models\upscale_models; Step 3: Download Sytan's SDXL Workflow. Set the filename_prefix in Save Image to your preferred sub-folder. ControlNet Workflow. If the term "workflow" has been used to describe node graphs for a long time then that's unfortunate because now it has become entrenched. Table of contents. Click on the "Load" button. 便利な ComfyUI SDXL simple workflow released. 46 GB) Verified: 3 months but you can achieve similar results with this ComfyUI Workflow: Workflow for this one is a bit more complicated than Share. Next is as simple as clicking the model in the "Reference models" folder of the networks section and waiting for the download to complete. these templates are the easiest to use and are recommended for new users of SDXL and ComfyUI. . This is the kind of thing that is pretty easy to figure in auto1111, but nigh impossible for the average person to figure out in comfy. I think it was 3DS Max. 5 without lora, takes ~450-500 seconds with 200 steps with no upscale resolution (see workflow screenshot from ver 1. Go to Civitai. In the added loader, select sd_xl_refiner_1. Here's a list of example workflows in the official ComfyUI repo. While the normal text encoders are not "bad", you can get better results if using the special encoders Feb 22, 2024 · The SDXL workflow includes wildcards, base+refiner stages, Ultimate SD Upscaler (using a 1. It's important to get all the steps and noise settings right: Aug 14, 2023 · Today, we embark on an enlightening journey to master the SDXL 1. Step 5. 5 + SDXL Base+Refiner - using SDXL Base with Refiner as composition generation and SD 1. Scan this QR code to download the app now. Entra en ComfyUI Manager y selecciona "Import Missing Nodes" y dentro los seleccionas todos y los instalas. 9vae. com/watch?v=zyvPtZdS4tIEmbark on an exciting journey with me as I unravel th SDXL Controlnet Tiling Workflow. In the top Preview Bridge, right click and mask the area you want to inpaint. Just my two cents. AUTOMATIC1111 can run SDXL as long as you upgrade to the newest version. json: Image-to-image workflow for SDXL Turbo; high_res_fix. 0 Inpainting model: SDXL model that gives the best results in my testing. First, download the pre-trained weights: Our robust file management capabilities enable easy upload and download of ComfyUI models, nodes, and output results. This probably isn't the completely recommended workflow though as it has POS_G and NEG_G prompt windows but none for POS_L, POS_R, NEG_L, and NEG_R which is part of SDXL's trained prompting format. Honestly you can probably just swap out the model and put in the turbo scheduler, i don't think loras are working properly yet but you can feed the images into a proper sdxl model to touch up during generation (slower and tbh doesn't save time over just using a normal SDXL model to begin with), or generate a large amount of stuff to pick and Nov 14, 2023 · I am using vanilla ComfyUI, with SDXL and LCM LoRA for SDXL. SD1. Set control_after_generate in the Seed node to LCM loras are loras that can be used to convert a regular model to a LCM model. All that is needed is to download QR monster diffusion_pytorch_model. SDXL Turbo synthesizes image outputs in a single step and generates real-time text-to-image outputs. So in this workflow each of them will run on your input image and you Aug 22, 2023 · The Easiest ComfyUI Workflow With Efficiency Nodes. SDXL has 2 text encoders on its base, and a specialty text encoder on its refiner. Explore new ways of using Würstchen v3 architecture and gain a unique experience that sets it apart from SDXL and SD1. I have updated the workflow submitted last week, cleaning up a bit the layout and adding many functions I wanted to learn better. safetensors, rename it e. Ok, so this is a bunch of tutorials I made centered on updating the same workflow step by step to look better. Follow these directions if you don't have Nov 13, 2023 · beta_schedule: Change to the AnimateDiff-SDXL schedule. Aug 20, 2023 · Step 4: Download and Use SDXL Workflow. Step 4. There is also a controlnet pre Efficient Loader & Eff. All legacy workflows was compatible. ThinkDiffusion - SDXL_Default. 1 at main (huggingface. I want a ComfyUI workflow that's compatible with SDXL with base model, refiner model, hi-res fix, and one LORA all in one go. This smoothens your workflow and ensures your projects and files are well-organized, enhancing your overall experience. The example workflow utilizes SDXL-Turbo and ControlNet-LoRA Depth models, resulting in an extremely fast generation time. The node specifically replaces a {prompt} placeholder in the 'prompt' field of each template with provided positive text. g. Choose to connect the SDXL or SD1. Preview of my workflow – download via the link SDXL setup, basic to advanced workflow. Add a Comment. To enable real-time prompting in ComfyUI, click on the Extra Options checkbox and then enable the Auto Queue checkbox. Jul 22, 2023 · if you need a beginner guide from 0 to 100 watch this video: https://www. Aug 20, 2023 · Now let’s load the SDXL refiner checkpoint. 1-Step The 1-step model is only experimental and the quality is much less stable. Reply reply Download the LoRA checkpoint (sdxl_lightning_Nstep_lora. ComfyUI: Node based workflow manager that can be used with Stable Diffusion. 0 in ComfyUI I've come across three different methods that seem to be commonly used: Base Model with Latent Noise Mask, Base Model using InPaint VAE Encode and using the UNET "diffusion_pytorch" InPaint specific model from Hugging Face. 5 refined model) and a switchable face detailer. json: Text-to-image workflow for SDXL Turbo; image_to_image. 在 Stable Diffusion SDXL 1. The ComfyUI Manager Add-On allows for the installation of custom nodes, enhancing the capabilities and functionalities of ComfyUI. 5 + SDXL Base shows already good results. There are two points to note here: SDXL models come in pairs, so you need to This uses more steps, has less coherence, and also skips several important factors in-between. 20 hours ago · Download the ComfyUI workflow below. Run git pull. safetensors) to /ComfyUI/models/loras; Download our ComfyUI LoRA workflow. 9; sd_xl_refiner_0. 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. The benefit of ComfyUI is that you can build custom workflow for the images you want to create and process, upscale, stylize etc or you can discover and download pre-built workflows in form of PNG files (which contain the workflow) or . df6d2ec 3 months ago. Enter the inpainting prompt (what you want to paint in the mask) on the right prompt and any Feb 1, 2024 · 12. Create animations with AnimateDiff. 0 版本推出以來，受到大家熱烈喜愛。. ComfyUIをインストール後、SDXLモデルを指定のフォルダに移動し、ワークフローを読み込むだけで簡単に使えます。. Then you can load this image in ComfyUI to get the workflow that shows how to use the LCM SDXL lora with the SDXL Download the provided JSON file from this repository. co) Thanks for sharing this setup. DrStalker. to control_v1p_sdxl_qrcode_monster. rhet0ric. • 8 mo. They can be used with any SDLX checkpoint model. If you want to use Stable Video Diffusion in ComfyUI, you should check out this txt2video workflow that lets you create a video from text. Instead of creating a workflow from scratch, you can download a workflow optimised for SDXL v1. 0-inpainting-0. JSON files which is the data of the workflow. 0 Refiner (opens in a new tab): Also place it in the models/checkpoints folder 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. com/models/117505?modelVersionId=132150 Custom Nodes Required: https://github. 2. json. Set the start and end time in seconds. 0 Base (opens in a new tab): Put it into the models/checkpoints folder in ComfyUI. Generate an image. x, 2. I have had to adjust the resolution of the Vid2Vid a bit to make it fit within those constraints. 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. 5 model. Img2Img ComfyUI workflow. To enhance results, incorporating a face restoration model and an upscale model for those seeking higher quality outcomes. Download it, rename it to: lcm_lora_sdxl. 5 from here. For those of you who are into using ComfyUI, these efficiency nodes will make it a little bit easier to g LCM-Turbo, download from here for SDXL based models, and then put into put into \ComfyUI\models\loras\SDXL\LCMTurboMix_Euler_A_fix. ComfyUI SDXL Turbo Workflow. Run ComfyUI and follow these steps: Click on the "Clear" button to reset the workflow. 9 models: sd_xl_base_0. This workflow can maintain 1 to 2 subjects on the screen, or multiple subjects. The Manager Add-On expands the functionality of ComfyUI by enabling the installation of custom nodes. This new cap Si te salen nodos en rojo y errores al cargar el workflow, es normal, quizá no tengas todos los nodos necesarios. Put it in the ComfyUI > models > checkpoints folder. Download the SDXL Turbo model. In auto1111 I can add one or more lora's just by selecting which ones I want to add. The LCM SDXL lora can be downloaded from here. Restart ComfyUI. 原因如下：. We will release Part 2 soon, where we add conditioning parameters and discuss those in detail. Click Queue Prompt to generate an image. If you installed from a zip file. SDXL Default ComfyUI workflow. And then, select CheckpointLoaderSimple. Production value zero, rambling? a little anyway, some nice tips and tricks, shows you in a basic way how to build this workflow and why things in that workflow are done the way they are. Ending Workflow. The quality of SDXL Turbo is relatively good, though it may not always be stable. In this case he also uses the ModelSamplingDiscrete node from the WAS node suite, supposedly for chained loras, however in my tests that node made no difference whatsoever so it can be ignored as well. Download. Enter the right KSample parameters. but it has the complexity of an SD1. 5. 0 Workflow. txt: Required Python packages Aug 8, 2023 · 画像生成AI. From this menu, you can either open a dialog to create a SAM Mask using 'Open in SAM Detector', or copy the content (likely mask data) using 'Copy (Clipspace)' and generate a mask using 'Impact SAM Detector' from the clipspace menu, and then paste it using 'Paste Okay, back to the main topic. Enhanced control and workflow with ComfyUI Manager Add-On. NOTE: You will need to use autoselect or linear (HotshotXL/default) beta_schedule, the sweetspot for context_length or total frames (when not using context) is 8 frames, and you will need to use an SDXL checkpoint. You signed out in another tab or window. This enables real-time prompting in ComfyUI allowing you to SDXL Turbo as intended. を丁寧に解説するという内容になっています。. com and select the models Jan 29, 2024 · 先日からstable diffusionの「ComifyUI」を使い始めて、少し慣れ始めたところで、今回は「SDXL」モデルを使って画像生成を行いたいと思います。 「SDXL」モデルをまだ持っていないという方は以下の記事を参考に「SDXL」モデルをダウンロードしてお試しください。 「SDXL」モデルはあるけれど Description. 5 + SDXL Base - using SDXL as composition generation and SD 1. Aug 17, 2023 · Download the included zip file（下载压缩文件）. 9; Install/Upgrade AUTOMATIC1111. With SDXL I often have most accurate results with ancestral samplers. In essence, choosing RunComfy for running ComfyUI equates to opting for speed, convenience, and efficiency. ComfyUI Txt2Video with Stable Video Diffusion. For ComfyUI you just need to put the models in the right folders and add the necessary custom nodes (can't recall the names, but Google is your friend). The usage description is inside the workflow. Step 1: Download a Hyper-SDXL model. Download the full checkpoint Jan 4, 2024 · ComfyUIでSDXLを使う方法. (cache settings found in config file 'node_settings. Extract the zip file (解压) Launch ComfyUI (运行comfyui) Click "Load" on ComfyUI and select the workflow . Upcoming tutorial - SDXL Lora + using 1. From there, we will add LoRAs, upscalers, and other workflows. 5 + SDXL Base+Refiner is for experiment only. Feb 22, 2024 · Download (6. py: Gradio app for simplified SDXL Turbo UI; requirements. Alternatively, set up ComfyUI to use AUTOMATIC1111’s model files. ComfyUI Manager: Plugin for CompfyUI that helps detect and install missing plugins. Support for SD 1. For example: 896x1152 or 1536x640 are good resolutions. do not try mixing SD1. I typically use 20 steps of the base model and 5 steps of the refiner, using ddim. context_length: Change to 16 as that is what this motion module was trained on. Nodes that can load & cache Checkpoint, VAE, & LoRA type models. But comfy is pretty esoteric. Please keep posted images SFW. You can directly load these images as workflow into ComfyUI for use. Dec 17, 2023 · ComfyUI-Background-Replacement. Aug 30, 2023 · Choose base model / dimensions and left side KSample parameters. A standard SDXL model is usually trained for 1024×1024 pixels, and performance in various image ratios can vary. Simply type in your desired image and OpenArt will use artificial intelligence to generate it for you. Reply. Not entirely sure everything is correct. make a folder in img2img. ComfyUI 可以一次過設定整個流程，對 SDXL 先要用 base model 再用 refiner model 的流程節省很多設定時間 With a higher config it seems to have decent results. Put into HotshotXL support (an SDXL motion module arch), hsxl_temporal_layers. ku vn va tp cy an dy kd bw rz