This is a textual inversion version of my <a target="_blank" rel="ugc" href="https://civarchive.com/models/1543/women-standing-in-their-panties">original Dreambooth for women in their underwear</a>. It offers greater flexibility to be used with many different models based off of SD 1.5 though if you want quality at the expense of flexibility and variety, you should also try the Dreambooth.I created this textual inversion specifically due to the ubiquity of models for nudes but there were very few for sexy clothed women. As the name implies, it's designed to be used with models for nudes (like F222.)It's trained on a similar dataset as the Dreambooth using the original SD 1.5 model and on 16 vectors. I've had great results using this on <a target="_blank" rel="ugc" href="https://civarchive.com/models/1188/f222">F222</a>, <a target="_blank" rel="ugc" href="https://civarchive.com/models/1173/hassanblend-all-versions">Hassan Blend</a>, <a target="_blank" rel="ugc" href="https://civarchive.com/models/1276/dreamlike-photoreal">Dreamlike Photo Real</a>, and even <a target="_blank" rel="ugc" href="https://civarchive.com/models/1202/the-allys-mix">the Ally's Mix</a> (for anime pics.)I've moved the recommended settings to the respective model pages since they're slightly different for each.For more tips on how to create beautiful women using SD (not just with this model), I've written a guide <a target="_blank" rel="ugc" href="https://rentry.org/sd_beautiful_women">here</a>.

Keep your beautiful underwear on, lady!

The trigger word is: shirtp_stNOTE: this is not a typo, this version has one less "t" in the trigger word. This is on purpose since I want to run both TI models side by side.(Yes, the model was originally designed for women wearing sexy shirts but I had gotten better results with underwear so I pivoted towards that.)Most of my testing of this model revolves around the following settings so this would be a good place to start:Recommended SettingsModel: SD 1.5, Hassan Blend, or F222VAE: <a target="_blank" rel="ugc" href="https://huggingface.co/stabilityai/sd-vae-ft-mse-original/tree/main">vae-ft-mse-840000-ema-pruned.ckpt</a>Hypernetwork: NoneStarter Prompt: woman, perfect face, standing, t-shirt, panties, feet, shirtp_st, bedroom(Note, at a bare minimum, you'll want to mention the following 2: woman, shirtp_st)Negative prompt: small eyes, eyes closed, long legs(It seems the images I trained the model on had people with smaller than average eyes but putting these in the negative prompt seems to help a lot. And depending on your dimensions, the legs may sometimes come out unnaturally long so include "long legs" in your negative prompt if that's the case.)Height: 1024Width: 512Sampler: Euler ASteps: 20CFG: 10 (though anything above 7 seems to work fine)Restore Faces: On (makes a huge difference)Highres Fix: On, Denoising = 0.5(Note: after the recent update, use an initial resolution of 384x768, an upscaling factor of 1.33333, and an upscaler of Latent.)This one was retrained from scratch using 4000 epochs this time. The quality of the images produced seem much higher than the last one though it seems to produce more blond women (if you don't specify your preference in your prompt.) Sometimes, it may cut off the upper body or the legs/feet a little more than the last one; if it seems to happen frequently with your particular prompt, try to mention the person's face and/or feet in your prompt.

Files with SHA256: 1e370c26e9c77e3ca47a03bd1aa06e84906001e14b3af11ae441064937a27a06