This is a fun model to play with, as the generation is so quick although the quality on the 512 version (the only one which is released yet) is not awesome. Learn more about the model here: https://github.com/IDKiro/sdxs. Setup is a little involved, and the below instructions are also in a note inside the workflow. Some other tips are in this ComfyUI issue: https://github.com/comfyanonymous/ComfyUI/issues/3147 SETUP ===== 1. Download this file [1] and place into your models/unet folder, rename to "sdxs-512-dreamshaper.safetensors" 2. Download this file [2] and place into your models/CLIP folder, rename to "clip_l.safetensors" 3. VAE should work with any SD1.5 compatible model. If the Load VAE node does not have any available, download [3] and place into models/VAE folder 4. Install any missing custom nodes, and restart if needed [1]: https://huggingface.co/IDKiro/sdxs-512-dreamshaper/blob/main/unet/diffusion_pytorch_model.safetensors [2]: https://huggingface.co/IDKiro/sdxs-512-dreamshaper/blob/main/text_encoder/model.safetensors [3]: https://huggingface.co/stabilityai/sd-vae-ft-mse-original/blob/main/vae-ft-mse-840000-ema-pruned.safetensors RUN === Use the Fast Groups Muter below to select between OBP and Wildcard mode (which also lets you type any text prompt). Auto Queue is fun to watch for a minute. Saving images is disabled by default. Why? Because when you can generate many images a second, auto queue could easily consume all of your disk space, that's why.