<h3 id="update:-0sae9l0im"><u>UPDATE</u>:</h3><p>Added a IL/Noob version. Dataset was altered and expanded. Results are far better than expected, but keep in mind I am using <a rel="ugc" href="https://civarchive.com/models/926443/ntr-mix-or-illustrious-xl-or-noob-xl">NTR mix</a> (<a rel="ugc" href="https://civarchive.com/models/926443?modelVersionId=1159421">version XII</a>) for the testing, and example images. Due to the versatility of Noob now you can use characters and styles together with this lora. Mix and match as you see fit.</p><hr /><h3 id="a-wise-man-once-said:-&quot;a-woman-has-two-pairs-of-lips.-if-you-love-her-you-kiss-one-pair-for-good-morning-and-the-other-for-good-night.&quot;-749t0g09c"><strong>A wise man once said: "A woman has two pairs of lips. If you love her, you kiss one pair for <span style="color:#fab005">good morning</span>, and the other for <span style="color:#7950f2">good night</span>.</strong>"</h3><h3 id="the-concept-is-a-woman's-portrait-full-face-view-with-a-side-to-side-view-of-her-vulva.-ugkmvnnma"><strong><span style="color:rgb(207, 207, 207)">The concept is a woman's portrait - full face view with a side-to-side view of her vulva.</span></strong></h3><p>I'm fairly sure media with the same idea exists somewhere.</p><p>The concept touches upon the idea of portraying a woman's beauty. However, it juxtaposes the beauty of her face (the thing everyone sees on a daily basis) with a view of her vagina (the most private and vulnerable part of her and the thing very few, close people ever see).</p><p>This concept is meant to show how unique and different and beautiful a vagina can be just like the woman it belongs to.</p><p></p><h3 id="out-of-respect-i-have-opted-to-not-use-real-people's-faces.-hence-i-have-relied-on-images-crated-through-tools-such-as-artbreeder-and-this-person-does-not-exist-and-other-similar-ai-tools.-me7jr1k58">Out of respect, I have opted to not use real people's faces. Hence I have relied on images crated through tools such as <a target="_blank" rel="ugc" href="https://www.artbreeder.com/create">Artbreeder</a> and <a target="_blank" rel="ugc" href="https://this-person-does-not-exist.com/en">This Person Does Not Exist</a> and other similar AI tools.</h3><p>Making those hundreds of faces this way was not easy or quick, and definitely not fun 😕 so if you like the end result, I'd appreciate a like, a review and maybe even a <a target="_blank" rel="ugc" href="https://ko-fi.com/pulenkompot#">coffee</a> (who needs sleep when you've got <span style="color:rgb(82, 56, 56)">caffeine</span>!)</p><p></p><h2 id="useful-information:-7s7woy1y7"><u>Useful information:</u></h2><h3 id="due-to-the-format-of-the-images-landscape-orientation-is-recommended.-the-concept-can-be-produced-in-portrait-orientation-as-well-but-keep-in-mind-the-idea-is-to-create-what-is-essentially-a-combo-of-two-portraits-simultaneously-so-landscape-fits-better-(also-the-dataset-was-designed-in-landscape-orientation-so-it-makes-sense).-5tz3mlacy">Due to the format of the images, <span style="color:rgb(130, 201, 30)">landscape orientation</span> is recommended. The concept can be produced in <span style="color:rgb(250, 82, 82)">portrait orientation</span> as well, but keep in mind, the idea is to create what is essentially a combo of two portraits simultaneously so landscape fits better (also the dataset was designed in landscape orientation, so it makes sense).</h3><p>*Also keep in mind that sometimes the face might look strange, especially the eyes, so I recommend running ADetailer.</p><p>**Due to the nature of the dataset, the images this LoRA produces are aways looking quite realistic (also due to the "realistic" tag), but despite this, the LoRA works fine with anime style models (see example pictures).</p><p></p><h2 id="further-information:-2tmvrs1ou"><u>Further information:</u></h2><h2 id="several-issues-with-this-model:-t7z4w50g8">Several <span style="color:rgb(250, 82, 82)">issues </span>with this model:</h2><p><span style="color:rgb(255, 255, 255)">1) Sometimes two faces are generated, and more often two vaginas instead of a face+vulva;</span></p><p><span style="color:rgb(255, 255, 255)">2) At times the vagina may become "colored" by the color of the woman's hair (purple hair=purple labia minora and labia majora for example, don't know why and I've not figured a way to fix it yet);</span></p><p><span style="color:rgb(255, 255, 255)">3) </span><u><span style="color:rgb(255, 255, 255)">Rarely</span></u><span style="color:rgb(255, 255, 255)"> the the two different types of lips might match or switch places... a truly unique sight.</span></p><p></p><h2 id="results-as-with-any-model-may-vary.-9dqk00qqm">Results, as with any model, may vary.</h2><h3 id="i-have-tested-the-lora-but-i-obviously-cannot-produce-all-possible-environments-so-this-model-might-work-perfectly-for-your-personal-settings-negative-embeddings-and-other-resources-used-for-image-generation-or-it-might-be-terrible.3-nhd1xa68b">I have tested the LoRA, but I obviously cannot produce all possible environments, so this model might work perfectly for your personal settings, negative embeddings and other resources used for image generation <strong><em>or </em></strong>it might be terrible.<br />（づ￣3￣）づ╭❤️～</h3>

PulenKompot-Pussy Portraits [NoobAI] [PonyXL] [SD1]

<p>The Pony version. It's wild, sometimes it produces very odd and/or funny results, but when it works it works <strong><em>really </em></strong>well (<s>ain't this true for any LoRA?)</s></p><p>As mentioned in the description, making this dataset was hard. Somewhere in the dataset, I am yet to find where exactly, there is an image that often gives the portraits quite large jaws. Once I figure what's causing it, I'll fix and reupload.</p><p>The tag "pubic hair" is not mandatory to trigger the Lora, merely optional depending on your preference.</p>

Files with SHA256: 577c316d4be339c5412d70403961216e436cc8ad75ed7acb7e9d0967141cc7e5