Update V2: Expanded Dataset: now with or without ball gags. Updated trigger.v2a: better tensionv2b: better variety/consistency<h1 id="this-is-a-fairly-difficult-pose-even-for-pony-xl-to-understand.-i-highly-recommend-an-image-to-image-workflow-as-described-here-here.-diyf9i7rv">This is a fairly difficult Pose even for Pony XL to understand. I highly recommend an image to image workflow as described here <a target="_blank" rel="ugc" href="https://civarchive.com/articles/4451/how-to-beat-the-odds-against-the-random-noise">here.</a></h1>Sample Outputs above are straight text to image though, so you can get results with this as well.<a target="_blank" rel="ugc" href="https://ko-fi.com/joschek">If you love my work: Consider supporting the cause ;)</a>This is a retrain of <a target="_blank" rel="ugc" href="https://civarchive.com/models/326114/joscheks-rope-bondage-hogtied-and-gagged-xl-bdsm-series">this SDXL Model</a>It is heavily biased towards realism. If you don't want that you have to heavily prompt against it and/or use style lora.<h3 id="if-you-can't-get-good-results-with-txt2img-with-this-model:-try-the-following-steps:-dkv2vp9ij">If you can't get good results with txt2img with this model: Try the following steps:</h3>-change your checkpoint. Some have an easier time with this pose than others-clean up your prompt. try getting the pose to work first and then selectively add your desired elements back in.-clean out your other lora: some of them may fight with this model and none of them will win.-as always: try higher steps, lower cfg. Upscaling will probably not save you on this one though. Don't Upscale before you have a decent base.-use some of the tags below<hr /><h3 id="taglist-ep5maqe08">Taglist</h3>To get the model to understand what's what, some pretty extensive manual tagging was necessary. Now the lora has a pretty long list of charged tags, that you may use (with varying effectiveness) in positive or negative: (don't go overboard though;)Perspective: From above | from side | from angle | facing viewer -- these may also be combined(From behind confused the model to much so it was removed from the dataset)Setting: On floor | on ground | on table | on bed ...Pose: arched back | head down | head back | feet crossed | feet together | knees together | knees apart feet together | elbows together | hands together elbows apartBondage related: (color) + gag | (color) + rope | natural rope | feet linked to hands | feet linked to shoulders | feet linked to hair | bound + (bodypart)<hr /><h3 id="good-luck-and-jqk8a61su">Good Luck and</h3><h3 id="may-the-rng-gods-bless-you-with-the-perfect-seed-3m6d9a0z9">may the RNG-Gods bless you with the perfect seed</h3>

Files with SHA256: 44432758f59e8de1848574281e2ca9a4001788a73f1fbfa0997ca72df48cf3af