<h2 id="rissa-may-m0f21oxl7"><strong>Rissa May</strong></h2><p></p><h2 id="image-generation-suggestions-ep4q1z6mr"><u>Image Generation Suggestions</u></h2><p>I generate these images via Mcmonkeyprojects' fantastic branch of the SwarmUI (a powerful but easy to use front end for ComfyUI and other backends). I find presently other UIs producing lacklustre results. YMMV.</p><h3 id="core-parameters-p5i5gfqdl"><br /><br /><strong>Core Parameters</strong></h3><ul><li><p>Model: Flux1-dev with clip_l, t5xxl_fp16.</p></li><li><p>Steps: 30</p></li><li><p>CFG Scale: 1</p></li></ul><p></p><h3 id="sampling-b93tnvpxq"><strong>Sampling</strong></h3><p>I've had success with a good amount of samplers and schedulers, but at present am using the following,</p><ul><li><p>Sampler: DEIS</p></li><li><p>Scheduler: Beta</p></li><li><p>Flux Guidance Scale: 3.5</p></li></ul><p></p><h3 id="prompting-z6dhpracq"><strong>Prompting</strong></h3><p>With SwarmUI use the <em>&lt;Segment:&gt;</em> feature to inpaint particular portions of the output. This can be nearly anything the segment model used was trained on, <em>face, eyes, feet, top</em>, etc.</p><h3 id="segmentation-(spot-in-painting)-btbs40g6b">Segmentation (spot in-painting)</h3><p>A general statement I tag to most of these prompts is:</p><ul><li><p><em><span style="color:color(srgb 0.902745 0.864706 0.902745)">&lt;segment:object,0.7&gt;prompt</span></em></p></li></ul><p>Where <em>object </em>is the particular object in the image you'd like to inpaint and the number is the strength. SwarmUI explains additional variables. Use like a superior aDetailer. Yolo models can also be added to SwarmUI and used as well.</p><p></p><h3 id="refineupscale-(hi-res-fix)-c1jkpwru9"><strong>Refine/Upscale (Hi-Res Fix)</strong></h3><p>These settings can be used to get higher resolution outputs without the banding it produces. Starting with a higher resolution and using this as a Refiner mechanism is often the way to maintain detail fidelity. I generally stay between 1 and 1.25 upscales for this reason.</p><ul><li><p>Refiner Control Percentage: 0.3 - 0.7</p></li><li><p>Refiner Method: Post-Apply</p></li><li><p>Refiner upscale: 1 - 1.5</p></li><li><p>Refiner Upscale Method: Latent: Bislerp</p></li><li><p>Refiner Do Tiling: Checked <em>(Important!)</em></p></li></ul><p></p><h2 id="posted-for-civitai.com-and-tensor.art-only-kdlz93p7y"><u>Posted for </u><a target="_blank" rel="ugc" href="http://civarchive.com"><u>civitai.com</u></a><u> and </u><a target="_blank" rel="ugc" href="http://tensor.art"><u>tensor.art</u></a><u> only</u></h2><p><em>Each of these LoRAs take me dozens of hours and GPU usage out of pocket. Every review is appreciated and encourages me to do more of these.</em></p><h2 id="moc-9gksjbhbt">-MoC</h2>

Files with SHA256: cdf920968813d5ce49c2d7d0815ada3f86c5b71816f7ad291fcca9693546200f