MAKE SURE TO READ THE INSTRUCTIONS CAREFULLYThis is the Double Exposure model for SD 2.x specialized in clean, photorealistic content. To get the very best results it's highly recommended to follow some rules concerning composition and especially prompting:<ul><li>The model was trained on subjects of women and men combined with nature (forests, trees, leaves, mountains) and cities (buildings and roads). Different animals can be triggered but might take more effort to get good results.</li><li>To achieve the very best results the prompt should be composed in this style:</li></ul><img src="https://imagecache.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/0096f642-2b9e-4a35-683e-538a85482800/width=525/0096f642-2b9e-4a35-683e-538a85482800" />That is: token + composition + subject made out of + nature/architecture, background color.The phrase made out of is central. Look more into successful prompting through the example images posted on this page.If you want an even simpler prompt composition this sort of setup will work, but the results might be less accurate:<code>dblxp double exposure close-up profile woman forest, background color</code>InstallationPick either the .ckpt or .safetensor file version of the model and put it in the models directory of your Automatic1111 installation - models\Stable-diffusion. You trigger the model with the token dblxp.Hope you get good results and I look forward to you sharing them with everyone!

Files with SHA256: ea5da93a07a362e3aa738658fa366e626b33308d12d6fc6157c7bc39c9f30693