<h2>EndlessMix is a merge of multiple models</h2>For sampling methods, use either Euler a or DPM++ 2M Karras<h1>DOWNLOAD/INSTALLATION</h1>Step 1: download SAFETENSORS file.Step 2: put the SAFETENSORS file under "stable-diffusion-webui\models\Stable-diffusion"Step 3: Done! enjoy the model<h1>USAGE</h1>-use a minimal negative prompt for best results-use Euler A or DPM++ 2M Karras and 20-25 steps for best results-use <a target="_blank" rel="ugc" href="https://danbooru.donmai.us/">danbooru </a>tags-I used a CLIP blending (optional) Link to script: <a target="_blank" rel="ugc" href="https://gist.github.com/catboxanon/bdcf23a2699c920da0d8177dbfa34e15">https://gist.github.com/catboxanon/bdcf23a2699c920da0d8177dbfa34e15</a><a target="_blank" rel="ugc" href="https://github.com/AUTOMATIC1111/stable-diffusion-webui/discussions/8266">https://github.com/AUTOMATIC1111/stable-diffusion-webui/discussions/8266</a>-I use the upscaler Latent (nearest) with a highres step of 0-20 and denoise of 0.6 for when there's a detailed background (optional)-I use the upscaler Latent (nearest-exact) with a highres step of 0-20 and denoise of 0.6 for close up of characters (optional)EXAMPLE PROMPTmasterpiece, best quality, ultra-detailed, beautiful lighting, 1girl, short hair, (multicolored hair:1.2), green eyes, cowboy shot, turtleneck, turtleneck sweater negative prompt: (low quality, worst quality:1.4)<h3>VAE</h3>No VAE is needed, loras should work as intended. Using a VAE could make your images look "burnt" or "fried".<h2>CREDIT TO THE PEOPLE WHO MADE THE MODELS <a rel="ugc" href="https://civarchive.com/models/17427/ether-blu-mix-ebm-v2">Ether Blu Mix</a>, <a rel="ugc" href="https://civarchive.com/models/5706/anyhentai">AnyHentai</a> and <a rel="ugc" href="https://civarchive.com/models/16164/icomix">iCoMix</a></h2>

Files with SHA256: 9da9c17d65173eccca79fbe5599a883901ae76a43195217433509d938df9f4cb