Stable diffusion upscale models

Stable diffusion upscale models. A notebook that demonstrates the original implementation can be found here: Stable Diffusion Upscaler Demo. We use Stable Diffusion Automatic1111 to upscale Latent upscaler. There are a few ways. We'll also do a latent upscale to give details to a washed out photo and go from washed There are couple good ones that 1) don't oversharpen the edges and 2) don't smudge the details. Read part 2: Prompt building. I checked my directories and the SwinIR and LDSR models are in there. ) support for stable-diffusion-2-1-unclip checkpoints that are used for generating image variations. In addition to choosing right Upscale model, it is very important to choose right model in Stable Diffusion img2img itself. 可以使用相同的模型和提示生成新的放大细节。. Select your preferred model and click on the Generate button to initiate the upscaling process. lsdr is also dont use tiles, so my attemt to get 4k image failed because of "CUDA out of memory" Update the RealESRGAN AnimeVideo-v3 model. I don't see anything else. Stable Diffusion v1 refers to a specific configuration of the model architecture that uses a downsampling-factor 8 autoencoder with an 860M UNet and CLIP ViT-L/14 text encoder for the diffusion model. A humble non-scientific test of comparing 4 popular upscalers in upscaling images to 2x via SD Upscale script. 0 Author: Philip Hofmann Network: ATD Scale: 4 Release Date: 22. Apr 2, 2024 · Step 1: Upload Your Image. - SwinIR has a painterly style and is less photorealistic. So I'm happy to announce today: my tutorial and workflow are available. 0. - WDN 4xV3 produces more detail than 4xV3, looks less cartoony. Unfortunately, I didn’t find many references about this latent upscaler trained by Katherine Crowson in collaboration with Stability AI. Fix deprecated float16/fp16 variant loading through new `version` API. The following resources can be helpful if you're looking for more information in Aug 16, 2023 · How to upscale any image with AI?In this tutorial, I'll guide you through the process of upscaling images using advanced AI-driven methods. May 10, 2023 · My extras tab only has 2 upscale models. stable-diffusion-webui\extensions\sd-webui-controlnet\models. The image still gets upscaled, but it's just stretched 4 times. May 12, 2023 · 9. pth inside the folder: "\ YOUR~STABLE~DIFFUSION~FOLDER\models\ESRGAN\") Restart you Stable Diffusion. To g The Stable Diffusion x4 Upscaler (SD x4 Upscaler) is a Latent Diffusion model used for upscaling images by a factor of 4. Add RealESRGAN_x4plus_anime_6B. io AI Image Enhancer & Upscaler Use a focused purpose-trained ESRGAN model like the one you (incorrectly) suggested was somehow outdated in the thread body post to upscale your initial output, and THEN run a secondary detailing pass on it with Stable Diffusion. It should appear in the Extras tab. Jan 25, 2023 · Today, we announce a new feature that lets you upscale images (resize images without losing quality) with Stable Diffusion models in JumpStart. 4x NMKD Siax - also very good. me/win10tweaker Discord https://discord. Updated 25 days ago • 469. • 2 yr. krnl/stable-diffusion-x4-upscaler. 10. to/xpuct Telegram https://t. In this example, the skin of girls is better on 3rd image , because of different model used while doing img2img Ultimate SD Upscale. Feb 18, 2024 · Applying Styles in Stable Diffusion WebUI. Super-resolution. The data set for testing consist of 5 images that have different graphic styles and different detailed. It is used to enhance the output image resolution by a factor of 2 (see this demo notebook for a demonstration of the original implementation). This is a base version of the model that was trained on LAION-5B. In this case I just did 4x upscales at 0. This section will show the five best ways to upscale images from stable diffusion. Stable Diffusion web UI では Extras タブ内に複数のアップスケーラーが搭載されています。簡単に検索してみると R-ESRGAN 4x+ (Anime6B) などがよく利用されているようです。一方、今回ご Text-to-Image with Stable Diffusion. Less is more approach. Viewing them on a HD or 4k screen they will look blurry or pixelated. Add the ncnn implementation Real-ESRGAN-ncnn-vulkan. Aug 23, 2023 · In this ComfyUI tutorial we look at my favorite upscaler, the Ultimate SD Upscaler and it doesn't seem to get as much attention as it deserves. 解析度拉越高，所需算圖時間越久，VRAM 也需要更多、甚至會爆顯存，因此提高的解析度有上限. pth, which is optimized for anime images May 5, 2023 · https://mega. The high-res fix is for fixing the generation of high-res (>512) images. After the last block of code finishes, you'll be given a gradio app link. Read part 3: Inpainting. This new site has a tag and search system, which will make finding the right models for you much easier! If you have any questions, ask here: https://discord. These AI stable diffusion upscalers use models that have been deeply trained with large data to upgrade your small-sized, low-res photos effectively. For example, see over a hundred styles achieved using prompts with the Aug 24, 2023 · Model card: Stable Diffusion x4 Upscaler Model Card; Pipeline: Stable Diffusion Latent Upscale Unraveling the Theory. Feb 11, 2024 · You signed in with another tab or window. s. Dec 21, 2023 · Stable Diffusion is accessible and easy to use. I share many results and many ask to share. Stable Diffusion本身也有提高分辨率的能力，在使用Stable Diffusion生成图像后，再用img2img功能提高分辨率 这种方法的好处是. Nodes/graph/flowchart interface to experiment and create complex Stable Diffusion workflows without needing to code anything. The extension helps you to generate or upscale large images (≥2K) with limited VRAM (≤6GB) via the following techniques: Reproduced SOTA Tiled Diffusion methods Mixture of Diffusers; MultiDiffusion; Demofusion; Our original Tiled VAE method; My original Tiled Noise Inversion method May 26, 2023 · About the testing. Result will be affected by your choice relative to the amount of denoise parameter. A number for scaling the image. Ultimate SD upscale padding: 128. 5 are trained primarily on smaller images, so choosing higher resolutions creates a lot of absurdities. Put the model file(s) in the ControlNet extension’s model directory. The upscale model loader throws a UnsupportedModel exception. The image quality falls short of greatness. We build on top of the fine-tuning script provided by Hugging Face here. It will download models the first time you run. 1. ↑ 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") ↑ Node setup 2: Upscales any custom image Latent upscaler. pth But automatic1111 version proactively refuses to use them: and gives this error: raise Exception("The file is a RealESRGAN model, it can't be used as a ESRGAN model. Set an URL to get a POST API call once the image generation is complete. However, moving the Automatic1111 workflow to ComfyUI has not been successful because the The hlky SD development repo has RealESRGAN and Latent Diffusion upscalers built in, with quite a lot of functionality. If the model isn't listed, download it and rename the file to model. Double-click run. If upscaling models do include fine detail it's usually just random noise than doesn't look like it would be there in the ground truth. Dec 5, 2023 · Stable Diffusion is a text-to-image model powered by AI that can create images from text and in this guide, I'll cover all the basics. Using prompts alone can achieve amazing styles, even using a base model like Stable Diffusion v1. Stable Diffusion is an AI-powered tool that enables users to transform plain text into images. Check the superclass documentation for the generic methods the library implements for all the pipelines (such as downloading or saving, running on a particular device, etc. Jul 3, 2023 · 方法 2：SD Upscale. Stable Diffusion is available in more ways; Midjourney is still only available through Discord. 4 denoise. I downloaded one that sounded fun to try out, it's a . finally, the tiles are almost invisible 👏😊. Best Anime Model: Anything v5. com). weight'. Mar 28, 2024 · Clarity AI | AI Image Upscaler & Enhancer - free and open-source Magnific Alternative - philz1337x/clarity-upscaler Whenever I try to use it, the console throws below error: Unable to load ESRGAN model C:\XXX\Stable-Diffusion\stable-diffusion-webui\models\ESRGAN\4xFaceUpDAT. An image that is low resolution, blurry, and pixelated can be converted into a high-resolution image that appears smoother, clearer, and more detailed. 3. Best Overall Model: SDXL. It didn't work out. In this post, we explore the following pre-trained Stable Diffusion models by Stability AI from the Hugging Face model hub. 3 usually gives you the best results. It is a node Real ESRGAN upscale and Esrgan Upscale (compatible with any pth file) Outpainting; Inpainting; Stable Diffusion Upscale; Attention, specify parts of text that the model should pay more attention to a man in a ((tuxedo)) - will pay more attention to tuxedo; a man in a (tuxedo:1. Plain vanilla ESRGAN 4x - surprisingly good for textures. gg/cpAUpDK. Gigapixel does a good job to the faces and skin, but nothing significant compared to open source models. Superscale is the other general upscaler I use a lot. As is to be expected, when I upscale, my people turn into plastic. We'll explore two Boosty (эксклюзив) https://boosty. This is the approved revision of this page, as well as being the most recent. Therefore, this article intends to talk about how to upscale an image in Stable Diffusion , so that we can catch up with frontier technology. google. Put the zip file to the folder you want to install Fooocus. We’re on a journey to advance and democratize artificial intelligence through open source and open science. We have moved to https://openmodeldb. Open Stable Diffusion WebUI and navigate to the “ Extras ” tab, where you’ll find the upscaling tools. To do upscaling you need to use one of the upscaling options. I've been messing with this for the last few days and cannot for the life of me get the Detailer panel to work. It sounds not easy to operate AI for average people, although we have never been closer to advanced technology. ckpt and upload it to your google drive (drive. Learn A111 and ComfyUI step-by-step. x, SD2. Apr 26, 2023 · It uses the Stable Diffusion x4 upscaler model and can quadruple the resolution of an image in somewhere between 20 and 40 seconds. Cupscale, which will soon be integrated with NMKD's next update. Problem solved. Jan 16, 2024 · Still, there's a lot more to explore, so don't call this one for Stable Diffusion just yet. 3-1. You should see the ControlNet section on the txt2img page. Upscale Type: Chess. bat to start Fooocus. g. However, moving the Automatic1111 workflow to ComfyUI has not been successful because the For a dozen days, I've been working on a simple but efficient workflow for upscale. Most Stable Diffusion interfaces provide some form of upscaling. Now in " Extra Tab " you got the new upscaler. Jul 25, 2023 · HOW TO INSTALL: Rename the file from: 4x-UltraSharp. - Both 4xV3 and WDN 4xV3 are softer than x4plus. May 2, 2024 · Initially, I was inspired by LDWorks' work; however, due to the limitations of their initial workflow, I decided to create my own, thus Ultra Upscale was born. Should just be able to drop the pth file into your stable-diffusion-webui\models\ESRGAN folder, and then relaunch the WebUI window, should show up in your upscalers list from there. scale. Best SDXL Model: Juggernaut XL. I highly recommend it, you can push images directly from txt2img or img2img to upscale, Gobig, lots of stuff to play with. This is the model that will be used by the AI upscaler. (YOUR ROOT)\models\ESRGAN\. Oct 31, 2023 · AVIIAX/upscale. I didn't create this upscaler, I simply downloaded it from a random link on reddit and uploaded here as I couldn't find it anywhere else. Furthermore, it gives you control over various parameters to produce excellent results. Controlnet + SD Upscale came out quite nice. Right-click on the zip file and select Extract All… to extract the files. But as you can see, still seams! May 18, 2023 · 因為是透過 Stable Diffusion Model 算圖，除了放大解析度外，還能增加細部細節！. Restart AUTOMATIC1111. A 4x model for 4x photo upscaler, handles jpg compression . ago. For teaching purposes, let’s go through upscaling with. Click it, and away you go, have fun! Find a chainner on github, it lets you use various upscale py models. More details are in anime video models. You switched accounts on another tab or window. Apr 29, 2023 · After specifying the desired upscaling value, the next step is to select the appropriate model for Stable Diffusion to use. I've downloaded "Anime sharp" and theres "4x-AnimeSharp. 4 for denoise for the original SD Upscale. 消除 AI upscaler引入的伪影。. gg/xpuct00:00 Начало08:35 Держите это Pipeline to upscale the resolution of Stable Diffusion output images by a factor of 2. The Stable Diffusion upscaler diffusion model was created by the researchers and engineers from CompVis, Stability AI, and LAION. Note the high tile width and height, padding and mask blur - all attempts to mitigate the issue, and it has definitely helped a lot from the base settings (especially changing from linear to chess actually). The higher the denoise number the more things it tries to change. Add small models for anime videos. Model Description: This is a model that can be used to generate and modify images based on text prompts. In the comfy UI manager select install model and the scroll down to see the control net models download the 2nd control net tile model(it specifically says in the description that you need this for tile upscale). AUTOMATIC1111 / stable-diffusion-webui Public My standard upscale method with that model is to use SD Upscale (not Ultimate) with Controlnet tile to do two 2x upscales, giving me the opportunity to fix things or use Lora after the first upscale as trying to fix a >4K image is painful. Upscaling-focused models seem much less "creative" compared to diffusion models, so you have to combine them if you want feasible fine detail. That's it. Download the zip file on this page. It's one of the most widely used text-to-image AI models, and it offers many great benefits. Copy the file 4x-UltraSharp. This process, called upscaling, can be applied to Video tutorial on how to use ComfyUI, a powerful and modular Stable Diffusion GUI and backend, is here. I've generated images using JauggernautXL in a 1360x768 resolution. Resources for more information: GitHub Repository. Many of them doesnt't work, it must be an ESRGAN model, and you should place it in. - Running ESRGAN 2x+ twice produces softer/less realistic fine detail than running ESRGAN 4x+ once. In addition to all upscalers you have available on extras tab, there is an option to upscale a latent space image, which is what stable diffusion works with internally - for a 3x512x512 RGB image, its latent space representation would be 4x64x64. I have switched over to the Ultimate SD Upscale as well and it works the same for the most part, only with better results. E. You signed out in another tab or window. g8nz/stable-diffusion-x4-upscaler. Currently, my focus has shifted more to tools like Clarity Upscaler and ultimately Magnific AI. Stable Diffusion comes with a lot of default models, I will list the best options for your use case here: Realism: R-ESRGAN-4x+ or LDSR (Slower) Paintings: ESRGAN_4x; Anime: R-ESRGAN 4x+ Anime6B; I will select R-ESRGAN-4x+ since my video is May 26, 2023 · Types of Stable Diffusion models. Media. Restart WebUI. The two things are not related really, don't be mislead by them both being "AI". Updated Jan 11 • 92. At the time of publishing, the latest version is called SDXL 1. Read part 1: Absolute beginner’s guide. For illustration/anime models you will want something smoother that PR, ( more info. Use small steps for upscale, 1. Models based on SDXL are better at creating higher Apr 18, 2024 · Follow these steps to install Fooocus on Windows. The company says that it provides more detailed results and a It seems that Upscayl only uses a upscaling model, so there is no difussion involved and the result will depend only on the upscaling model. Jul 12, 2023 · When using the extras tab in A1111 for upscaling, the results are acceptable but far from ideal. A latent diffusion upscaler for the Stable Diffusion autoencoder. Reply. The process is faster, because is less complex. model_id. It seems a proccess similar to the one we can find in the EXTRAS Menu in Automatic1111 or the upscaling nodes in ComfyUI. When searching for ways to preserve skin textures, in guides Feb 22, 2024 · The Stable Diffusion 3 suite of models currently ranges from 800M to 8B parameters. I think sd passes is necessary to add detalization. 21) - alternative syntax The Stable Diffusion Latent Upscaler model was created by Katherine Crowson in collaboration with Stability AI. 03. Please see anime video models and comparisons for more details. They are not Stable Diffusion models but neural networks trained for enlarging images. Fully supports SD1. We assume that you have a high-level understanding of the Stable Diffusion model. What I'm calling Stable Diffusion is actually a number of open source models. wiki/wiki/Model_Database Jul 27, 2023 · I try to upscale SDXL output images and want to use the stable-diffusion-x4-upscaler. 缺點：. Best Realistic Model: Realistic Vision. Nov 22, 2023 · For "Upscaler 1" we have a few different options. 4x NMKD Superscale - my current favorite. This might be options built into the txt2img pipeline (upscaling after image generation), or it may be a feature which allows any image to be upscaled. May 1, 2024 · Initially, I was inspired by LDWorks' work; however, due to the limitations of their initial workflow, I decided to create my own, thus Ultra Upscale was born. Unable to determine this model's library. What are the Best Upscale Models in Stable Diffusion? I don't think there really is an ideal model right now. Nov 23, 2023 · Stable Diffusion models based on SD 1. I was always told to use cfg:10 and between 0. For this guide, we will be using 4x-UltraSharp. Nov 24, 2022 · The Stable Diffusion 2. Prompts. pth RealESRGAN_x4plus_anime_6B. ago • Edited 2 yr. It is used to enhance the resolution of input images by a factor of 4. Random notes: - x4plus and 4x+ appear identical. Stable Diffusion 3 combines a diffusion transformer architecture and flow matching. 如果您刚刚在txt2img页面上生成图像，请单击 Mar 19, 2024 · We will introduce what models are, some popular ones, and how to install, use, and merge them. 4x_foolhardy_Remacri is now available in the Extras tab and for the SD Upscale script. I could not find an example how to use stable-diffusion-x4-upscaler with ComfyUI. Model description. pth. Dec 28, 2022 · This tutorial shows how to fine-tune a Stable Diffusion model on a custom dataset of {image, caption} pairs. The Stable Diffusion latent upscaler model was created by Katherine Crowson in collaboration with Stability AI. However, fear no I was actually using for every SD version the Real-ESRGAN models which are under the filenames: RealESRGAN_x4plus. Check the docs . i dont have valar but raw lsdr makes something like this on face when i try to upscale first image. If it's the best way to install control net because when I tried manually doing it . 0 release includes robust text-to-image models trained using a brand new text encoder (OpenCLIP), developed by LAION with support from Stability AI, which greatly improves the quality of the generated images compared to earlier V1 releases. (#21) 10 months ago Upscaling inside a Local Stable Diffusion Interface. 2024 Iterations: 240'000 epoch: 152 batch_size: 6, 3 HR_size: 128, 192 Dataset: nomos8k Number of train images: 8492 OTF Training: Yes Pretrained_Model_G: 003_ATD_SRx4_finetune Description: 4x photo Model Database. webhook. This model inherits from DiffusionPipeline. face_enhance. 0. URL of the image that you want in super resolution. Feb 12, 2024 · With extensive testing, I’ve compiled this list of the best checkpoint models for Stable Diffusion to cater to various image styles and categories. Anyone can use the model to generate smoother, sharper, and finer images with details. 5 or SDXL. However, looking at the tedious Stable Diffusion upscale process, beginners may switch to an alternative. x, SDXL, Stable Video Diffusion and Stable Cascade; Asynchronous Queue system; Many optimizations: Only re-executes the parts of the workflow that changes between executions. The right upscaler will always depend on the model and style of image you are generating; Ultrasharp works well for a lot of things, but sometimes has artifacts for me with very photographic or very stylized anime models. I wanted a very simple but efficient & flexible workflow. ChaiNNer offers support for PyTorch (with select model architectures), NCNN, and Do an ULTIMATE Upscale to get Ultra 8k quality, 8192x8192 resolution. How do I setup this Upscaler? :/. Sometimes models appear twice, for example “4xESRGAN” used by chaiNNer and “4x_ESRGAN” used by Automatic1111. If you’ve just created an image you want to upscale, simply click “ Send to Extras ,” and it will take you to the upscaling section with your image ready. 5x. 調整 prompt 和 denoising strength，在此階段同時對圖片作更進一步的優化. Download this first, put it into the folder inside conmfy ui called custom nodes, after that restar comfy ui, then u should see a new button on the left tab the last one, click that, then click missing custom nodes, and just install the one, after you have installed it once more restart comfy ui and it ahould work. The text-to-image models in this release can generate images with default The hlky SD development repo has RealESRGAN and Latent Diffusion upscalers built in, with quite a lot of functionality. What I've been doing to make it work is using IMG to IMG and using the Ultimate SD Upscale script in combination with ControlNet tiling and an SD model such as Jauggernaut Aftermath, RevAnimated, ChilloutMix or similar. The model was pretrained on 256x256 images and then finetuned on 512x512 images. An AI upscaler is an AI model for enlarging images while filling in details. You can access this via the button in the upper-right-hand corner. Stable Diffusion is a latent diffusion model conditioned on the (non-pooled) text embeddings of a CLIP ViT-L/14 text encoder. Which one is better will depend on Images generated with Stable Diffusion most often are sized around 512x512 pixels. pth" in the folder. Striking-Long-2960. stable-diffusion-2-1-base. Best Fantasy Model: DreamShaper. Don't know about your particular upscale, but I've pulled several upscalers from this wiki, all added to ESRGAN folder and worked fine. This approach aims to align with our core values and democratize access, providing users with a variety of options for scalability and quality to best meet their creative needs. Use this model to generate images based on a text prompt. an AI upscaler; Hi-res fix; Ultimate Upscale; AI upscale. nz/folder/qZRBmaIY#nIG8KyWFcGNTuMX_XNbJ_g Ultra Sharp x4 Upscaler (Found In The Upscaler Database Here: https://upscale. Then I pressed Fetch updates and Update ComfyUI and the line got up as it should and those two items disappeared. You can change image size and add details (via denoise parameter) to skin, texture, etc Just enter new width & height and Denoising value. If you need to look at the old Model Stable Diffusion Upscale; Attention, specify parts of text that the model should pay more attention to a man in a ((tuxedo)) - will pay more attention to tuxedo; Sep 26, 2023 · Stable Diffusion, for example, is an artificial intelligence model that creates images. Mar 20, 2024 · There are several ways to upscale in Stable Diffusion. We provide a reference script for sampling, but there also exists a diffusers integration, which we expect to see more active community development. pt to: 4x-UltraSharp. A boolean flag ( true/false) for face enhancement feature. It is a Latent Diffusion Model that uses a fixed, pretrained text encoder ( OpenCLIP-ViT/H ). A common question is applying a style to the AI-generated images in Stable Diffusion WebUI. This is part 4 of the beginner’s guide series. So I'm kinda new to all this but so far standard upscale tests have proven to me that topaz is way better than stable diffusion upscaled images. generate your 2048x2048 image using the high-res fix, then send to extras, then upscale to 8k using any of the available options. If the model you want is listed, skip to step 4. 2 and 0. 5 Best Methods to Upscale Images from Stable Diffusion. Image-to-Image • Updated Oct 31, 2023 • 252. Reload to refresh your session. Dec 3, 2023 · Learn about ControlNet, ESRGAN, Superscale, "Hires Fix", the scripts "SD Upscale" and "Ultimate SD Upscale". pth file. Note: Stable Diffusion v1 is a general text-to-image diffusion For upscaling I mainly used the chaiNNer application with models from the Upscale Wiki Model Database but I also used the fast stable diffuison automatic1111 google colab and also the replicate website super resolution collection. info/. 0 depth model, in that you run it from the img2img tab, it extracts information from the input image (in this case, CLIP or OpenCLIP embeddings), and feeds those into For example, this is all you need to do in order to perform an upscale: Before you get to this point though, you'll need to install one of the neural network frameworks from the dependency manager. I now need to get them upscaled. The tile model should be available for selection in the Model dropdown menu. In any case, I think it’s a safe bet to assume it was trained in a similar way to how . This test was made in the most lazy method as you can imagine, without any enhancing images via inpainting. py file into your scripts directory \stable-diffusion-webui\scripts\ Example Generation A-Zovya Photoreal [7d3bdbad51] - Stable Diffusion Model Jan 10, 2023 · yes its better. May 6, 2023 · Install Path: You should load as an extension with the github url, but you can also copy the . upscale model to use, default is realesr-general-x4v3. Nothing fancy. 以 google Copy it to: \stable-diffusion-webui\models\ESRGAN. It can be used on top of any StableDiffusionUpscalePipeline checkpoint to enhance its output image resolution by a factor of 2. Aug 6, 2023 · Stable Diffusionで高画質の画像を生成したいなら、アップスケールは必要不可欠です。 しかし、アップスケールしていると次のような経験をする人も多いはず。 アップスケールしたら全然雰囲気の違う画像になってしまった… 思ったほど綺麗 Feb 19, 2023 · 動作環境は google colab T4 インスタンスの Stable Diffusion web UI です。 LDSR & SD Upscale. Or use script SD Upscale. Below we’ll look at using some of the most common SD interfaces to upscale; Part 2. It works in the same way as the current support for the SD2. ") I also had this problem in the beginning. It works by denoising the latent image noise directly conditioned on the low-resolution image. I wholeheartedly recommend Remacri as your general go-to upscaler, it's easily the best results you'll get outside of the paid upscalers. Link to Github Release Name: 4xNomos8k_atd_jpg License: CC BY 4. ) Do you have ComfyUI manager. Structured Stable Diffusion courses. May 16, 2023 · A dropdown allows you to to select the kind of upscaler to use for resizing the image. p. pth: 'trunk_conv. bh jl cj op ic mu uy bd vk sv