I can guide you with the best practices, tips, and tricks to achieve your goal in your Stable Diffusion training. Take a look at my Patreon, subscriptions for just $5. Exclusive content. <a target="_blank" rel="ugc" href="https://www.patreon.com/sdisn">Patreon</a>. :)<h3 id="about:-4s6juybzk">About:</h3>The purpose of this LoRA is to <s>(try to)</s> replicate a sideview of a Blowjob. Be careful, may generate body horror, disembodied penises and weird things.It was trained in multires (512x768, 576x768, 640x768, 704x768 and 768x768), using simple captions and with 106 high-resolution images.The demo images of the model aren't edited, but they are cherry-picked and generated using hires-fix (0.15 ~ 0.45 denoising).<h3 id="suggested-settings-for-inference:-tfwsd1yee">Suggested Settings for Inference:</h3>Model: <a target="_blank" rel="ugc" href="https://civarchive.com/models/8030/analog-madness-realistic-model">analogMadness_v70</a> (but works with 97,2% of models)Positive: a woman is sucking a cock, cute asian, short hair, wide open mouth, full facial shot &lt;lora:bj_from-side_v1_by_IsnAI:1&gt;Negative: morbide-body-parts, wrong eyeballs, low res, lower qualitySampler: DPM++ 2M KarrasSteps: 32CFG: 6Resolution: 512x768, 576x768, 640x768, 704x768 or 768x768Note¹: You can get great image variations if you tweak resolution a bit.Note²: In this version, it's not necessary to reduce weight, it's perfectly possible to generate the images using weight 1 in LoRA (all demos use weight 1), but maybe someone with more patience can discover some interesting things that I didn't have time to test (I'm also not very creative with prompts).<h3 id="triggers:-e0f8n5kcj">Triggers:</h3>Main: a woman is sucking a cock, wide open mouth, open mouthAngle: head and chest only, full scene shot, full face shotTags Used on Training: eye contact, deepthroat, throatfuck, kneeling, bbc, busty, teeth, indoor scene, outdoor scene, shot from below, open wide eyesNote³: <code>wide open mouth</code> results in bigger cocks and <code>throatfuck</code> doesn't seem to work well.

Files with SHA256: 4f4aa2387bde0b998893698e888a54586c989d8ddf969f8ee9c7d4746d008c1d