Stable Diffusion SDXL Wildcards and ComfyUI

<p>I like kicking off large random wildcard runs with Stable Diffusion &mdash; this consists of the prompt for the image generation being comprised of random text pulled out of specific files. I&rsquo;ve written about this several times so far:</p> <h2><a href="https://ericri.medium.com/making-wild-images-with-stable-diffusion-f87605b4aa99?source=post_page-----e8483e00e1c--------------------------------" rel="noopener follow" target="_blank">Making Wild Images With Stable Diffusion</a></h2> <h3><a href="https://ericri.medium.com/making-wild-images-with-stable-diffusion-f87605b4aa99?source=post_page-----e8483e00e1c--------------------------------" rel="noopener follow" target="_blank">Since my last post, I&rsquo;ve been indulging in generating thousands and thousands of Stable Diffusion images with&hellip;</a></h3> <p><a href="https://ericri.medium.com/making-wild-images-with-stable-diffusion-f87605b4aa99?source=post_page-----e8483e00e1c--------------------------------" rel="noopener follow" target="_blank">ericri.medium.com</a></p> <h2><a href="https://ericri.medium.com/the-x-y-plot-to-eliminate-some-stable-diffusion-ai-art-models-50d04500fecd?source=post_page-----e8483e00e1c--------------------------------" rel="noopener follow" target="_blank">The X/Y Plot to Eliminate Some Stable Diffusion AI Art Models</a></h2> <h3><a href="https://ericri.medium.com/the-x-y-plot-to-eliminate-some-stable-diffusion-ai-art-models-50d04500fecd?source=post_page-----e8483e00e1c--------------------------------" rel="noopener follow" target="_blank">I discuss the X/Y Plot script for a practical task: I analyze the 50+ models I have in my Stable Diffusion and promote&hellip;</a></h3> <p><a href="https://ericri.medium.com/the-x-y-plot-to-eliminate-some-stable-diffusion-ai-art-models-50d04500fecd?source=post_page-----e8483e00e1c--------------------------------" rel="noopener follow" target="_blank">ericri.medium.com</a></p> <h2><a href="https://ericri.medium.com/smile-get-wild-with-stable-diffusion-wildcards-and-facial-expressions-83a52e4ca9fd?source=post_page-----e8483e00e1c--------------------------------" rel="noopener follow" target="_blank">Smile! Get Wild With Stable Diffusion &mdash; Wildcards and Facial Expressions</a></h2> <h3><a href="https://ericri.medium.com/smile-get-wild-with-stable-diffusion-wildcards-and-facial-expressions-83a52e4ca9fd?source=post_page-----e8483e00e1c--------------------------------" rel="noopener follow" target="_blank">Quick share of two techniques: (1) using wildcards for variety in your prompt and (2) having specific expressions in&hellip;</a></h3> <p><a href="https://ericri.medium.com/smile-get-wild-with-stable-diffusion-wildcards-and-facial-expressions-83a52e4ca9fd?source=post_page-----e8483e00e1c--------------------------------" rel="noopener follow" target="_blank">ericri.medium.com</a></p> <p>I&rsquo;m going to cover how I set up my random prompt generator in ComfyUI, and then I&rsquo;m going to discuss how I screwed something up and needed to fix it. Don&rsquo;t try to implement what I did without reading all the way through.</p> <h1>Hello to SDXL and Goodbye to Automatic1111</h1> <p>Once SDXL was released I of course wanted to experiment with it. It seems just as disruptive as SD 1.5 was. But I can&rsquo;t use Automatic1111 anymore with my 8GB graphics card just because of how resources and overhead currently are. I can, however, use the lighter weight ComfyUI.</p> <h1>I Want My Random Back!</h1> <p>Automatic comes with well-thought out features and exceptionally well-integrated extensions. I used the wildcard extension and the X/Y/Z plot to kick off runs for hundreds of images. The first thing I want back: something to replace my wildcard random prompt generation.</p> <p><a href="https://ericri.medium.com/stable-diffusion-sdxl-wildcards-and-comfyui-e8483e00e1c"><strong>Visit Now</strong></a></p>