Name: wyrde's ComfyUI Workflows
Rating: 5 (116 reviews)
Author: wyrde

<h1>wyrde's workflows for various things</h1>More examples and help documents on github: <a target="_blank" rel="ugc" href="https://github.com/wyrde/wyrde-comfyui-workflows">https://github.com/wyrde/wyrde-comfyui-workflows</a>The recent changes to civit's UI make sharing these on civit a painful process.Expand the About this Version box to the right → to see more.

download the zip for the <a target="_blank" rel="ugc" href="http://readme.md">readme.md</a> or goto <a target="_blank" rel="ugc" href="https://github.com/wyrde/wyrde-comfyui-workflows/tree/main/compare/prompts-x4">https://github.com/wyrde/wyrde-comfyui-workflows/tree/main/compare/prompts-x4</a>

Compare 4 Prompts v0.8

wyrde img2img-hr-fix-sharpen-x2-x2Uses WAS nodes. <a target="_blank" rel="ugc" href="https://github.com/WASasquatch/was-node-suite-comfyui">https://github.com/WASasquatch/was-node-suite-comfyui</a>Takes an image and<ul><li>resizes it to 512x768</li><li>uses bicubic to scale it x2</li><li>samples it for a hi-rez fix</li><li>upscales with a model by x4 (now x8 original size) then scales it down again for a final result of 2048x3072</li></ul>Values can be changes easily.Model<ul><li><a target="_blank" rel="ugc" href="https://civitai.com/models/8281/perfect-world">https://civitai.com/models/8281/perfect-world</a></li></ul>Uspcale<ul><li>Lollypop <a target="_blank" rel="ugc" href="https://drive.google.com/u/1/uc?id=10h8YXKKOQ61ANnwLjjHqXJdn4SbBuUku&amp;export=download">https://drive.google.com/u/1/uc?id=10h8YXKKOQ61ANnwLjjHqXJdn4SbBuUku&amp;export=download</a></li></ul>Embeds<ul><li>EasyNegative <a target="_blank" rel="ugc" href="https://civitai.com/models/7808/easynegative">https://civitai.com/models/7808/easynegative</a></li><li>bad-hands-5 <a target="_blank" rel="ugc" href="https://huggingface.co/yesyeahvh/bad-hands-5/tree/main">https://huggingface.co/yesyeahvh/bad-hands-5/tree/main</a></li></ul>

HRF #1

wyrde branch sharpen HR-Fix model upscaling x4 (many images)Takes an image from text prompt (or image) and<ul><li>samples it</li><li>uses an upscale model on it</li><li>reduces it again and sends to a pair of samplers</li><li>they upscale and reduce</li><li>then send results to another pair of samplers</li></ul>Not quite a true double-pass sharpen, but the results are decent. Lots of variation.Unhook noodles or mute nodes to reduce undesired images and paths. Move noodles to switch VAE and img2img.For more information and examples, see <a target="_blank" rel="ugc" href="https://github.com/wyrde/wyrde-comfyui-workflows/blob/main/hr-fix-upscale/branch-sharpen-x6/README.md">the readme on github.</a>

HRF #2

wyrde multi sample sharpening with model upscaleThere are <a target="_blank" rel="ugc" href="https://github.com/wyrde/wyrde-comfyui-workflows/tree/main/hr-fix-upscale/multi-sample-sharpening">several versions of this workflow on github!</a>Takes an image from text prompt and<ul><li>samples it</li><li>latent upscales</li><li>resamples it</li><li>repeats</li><li>uses a 4x upscale model to increase the size again</li><li>saves it at a total of x8 size</li></ul>Basically a series of hi-res fixes then an model upscaleMoving noodles switches between model vae and vae loader. Mute nodes or unhook noodles to decrease generated images. If you see red boxes, you probably need WAS nodes.

HRF #3

multi sample pixel space sharpeningThere are <a rel="ugc" href="https://github.com/wyrde/wyrde-comfyui-workflows/tree/main/hr-fix-upscale/multi-sample-pixel-space-sharpening">multiple versions of this workflow on github</a>! (Along with links to resources and models used.)Takes an image from text prompt and<ul><li>samples it</li><li>uses a model to upscale</li><li>resamples it</li><li>repeats</li><li>at the end, uses uses a 4x upscale model to increase the size again</li><li>saves it at a total of x8 size</li></ul>Basically a series of hi-res fixes then an model upscaleMoving noodles switches between model vae and vae loader.

HRF #4

wyrde HR-Fix model upscaling x2 x2 (x4 final)Includes WAS nodes. <a target="_blank" rel="ugc" href="https://github.com/WASasquatch/was-node-suite-comfyui">https://github.com/WASasquatch/was-node-suite-comfyui</a>Takes an image from text prompt (or image) and<ul><li>samples it</li><li>uses an upscale model to double the size</li><li>hi-rez fixes the image through a sampler</li><li>uses a 4x upscale model to increase the size again</li><li>saves it at a total of x4 size</li></ul>Includes a couple lora. There's also a places where moving noodles changes the generation.resourcesModel<ul><li><a target="_blank" rel="ugc" href="https://civitai.com/models/8281/perfect-world">https://civitai.com/models/8281/perfect-world</a></li></ul>Lora<ul><li><a target="_blank" rel="ugc" href="https://civitai.com/models/17892/cat-lingerie">https://civitai.com/models/17892/cat-lingerie</a></li><li><a target="_blank" rel="ugc" href="https://civitai.com/models/16687/butterfly-wings">https://civitai.com/models/16687/butterfly-wings</a></li></ul>Uspcale<ul><li>PSNR <a target="_blank" rel="ugc" href="https://drive.google.com/drive/folders/1ldwajXL50uC7PCS63B4Wato6Dnk-svNL">https://drive.google.com/drive/folders/1ldwajXL50uC7PCS63B4Wato6Dnk-svNL</a></li><li>Lollypop <a target="_blank" rel="ugc" href="https://drive.google.com/u/1/uc?id=10h8YXKKOQ61ANnwLjjHqXJdn4SbBuUku&amp;export=download">https://drive.google.com/u/1/uc?id=10h8YXKKOQ61ANnwLjjHqXJdn4SbBuUku&amp;export=download</a></li></ul>Embeds<ul><li>EasyNegative <a target="_blank" rel="ugc" href="https://civitai.com/models/7808/easynegative">https://civitai.com/models/7808/easynegative</a></li><li>bad-hands-5 <a target="_blank" rel="ugc" href="https://huggingface.co/yesyeahvh/bad-hands-5/tree/main">https://huggingface.co/yesyeahvh/bad-hands-5/tree/main</a></li></ul>