[SD 1.5][Checkpoint]Refined Refined_v10
<p>Merged Model with slight fine-tuning focused on human anatomy and skin texture, can also be used as a general purpose model. v10 is a foundation that will be further built upon with more mergers + fine-tuning.</p><p></p><p><strong><u>List of models used in merge can be found in the version description.</u></strong></p><p></p><p><strong><u>Limitations</u></strong><em> (as of now)</em></p><ul><li><p>Photos of Men</p></li><li><p>Anime </p></li><li><p>SFW generations without specific tags</p></li></ul><p></p><p><strong><u>Trigger Words</u></strong></p><p>Trigger words are not required.</p><p>However, the following tags can be used to alter the generation.</p><p><em>analog style, modelshoot style, wa-vy style, synthwave, sci-fi, cyborgdiffusion, erect penis, flaccid penis, postapocalypse, 3d, render, nsfw, sfw</em></p><p>For tags like "nsfw", "sfw", "3d", and "render", you can swap these between your positive and negative prompts to get more specific outputs.</p><p>For example, putting <strong>"sfw"</strong> in your <strong>positive prompt</strong>, and <strong>"nsfw"</strong> in your <strong>negative prompte</strong> will push the model in the direction of generating sfw outputs. You can also add/decrease attention using <code>()</code>and <code>[]</code></p><p></p><p><strong><u>Textual Inversion (TI)</u></strong></p><p>Many of sample images include a negative textual inversion that helps fix things like hands, ect.. This isn't a requirement, I've tested the model without this TI and didn't have any issues. I do recommend using it though, as it does help.</p><p></p><p>Bad Prompt TI - <a target="_blank" rel="ugc" href="https://huggingface.co/datasets/Nerfgun3/bad_prompt">https://huggingface.co/datasets/Nerfgun3/bad_prompt</a></p><p>I recommend using Version 2 and setting it's weight to 0.8 in the negative prompt; <code>(bad_prompt_version2:0.8)</code></p><p></p><p>For webui, <code>bad_prompt_version2.pt</code> goes in:</p><p>sd/stable-diffusion-webui/embeddings</p><p><strong><u>Hires Fix</u></strong></p><p>I used Hires Fix for many of the sample images. This is not a requirement but it does seem to improve results. I mainly use latent(bicubic antialiased) with Denoising strength set to 0.65.</p><p><strong><u>Links to my other models</u></strong></p><ul><li><p>Experience - <a target="_blank" rel="ugc" href="https://civitai.com/models/5952/experience">https://civitai.com/models/5952/experience</a></p></li><li><p>Elegance - <a target="_blank" rel="ugc" href="https://civitai.com/models/5564/elegance">https://civitai.com/models/5564/elegance</a></p></li></ul><ul><li><p>Clarity - <a target="_blank" rel="ugc" href="https://civitai.com/models/5062/clarity">https://civitai.com/models/5062/clarity</a></p></li><li><p>VisionGen - Realism -<a target="_blank" rel="ugc" href="https://civitai.com/models/4834/visiongen-realism">https://civitai.com/models/4834/visiongen-realism</a></p></li></ul>
页:
[1]