Share your images here and leave a like if you enjoyed it! Get exclusive NSFW models at DeviantArt or Booth.pm.
Tags: top-down bottom-up, from behind, ass, doggystyle
Recommended quality tags: score_9, score_8_up, score_7_up, score_6_up. Optional: source_anime
Recommended negative tags: score_4, score_5. Optional: 3d, lips
Recommended checkpoint: AutismMix_confetti
Description
LoRA trained to generate a doggystyle position where the receiving partner has their face down and ass up. I couldn't find a name for this pose so don't mind the random trigger I used. Note that this is a complex pose and doesn't solve issues inherent to SD. So it may take you more work to get a good image. It took quite a few iterations before I could get to a point that was stable enough.
Please rate and leave a like if you enjoyed it! And be sure to follow to check out the upcoming models.
Note: this model only works with a square aspect ratio (e.g. 512x512). You can increase the width a bit, not the height. It was trained only with images where the male is out of the frame, so if won't be able to generate a portrait aspect ratio. I recommend using hires fix as well.
The main tag is the bare minimum that must be in the prompt: top-down bottom-up
Useful tag to help the effect: piledriver doggystyle (yeah, I know it's a bit random)
Tags related to the view from the back, especially if it's not achieved with just the main tag: from behind, looking back, ass.
Tags related to the sexual activity, especially if it's not achieved with just the main tag: sex, sex from behind, penis, testicles, anal, etc.
FAQ
Comments (9)
Is it possible to have this but without the male/sex involved? Just the ass raised?
yes, it should be. just use the main tag "top-down bottoms-up", and in the negative prompt add any tag related to the male sex, if visible, e.g.: 1boy, penis, sex, etc.
@psoft Great! Thank you!
Was this trained on only anal pics?
not only, but it does seem to be easier to generate
😍
No matter what I do, I cant' get this to work.
Any helpful things to try?
I agree that this is a difficult lora to use. I may give it another try later on.
you must have all of these conditions: you must use 512x512 resolution, include all of the recommended tags, and use an anime base model. try it first without other loras.
some base models recommend negative embeddings, which tend to mess things up. so avoid those as well.
yeah, this is a difficult one. I may give it another try later on.
you must have the following: 512x512 resolution, include the recommended tags, use an anime model, preferably one that's not an insane remix of merges that requires negative embeddings to hold it together. those embeddings tend to mess things up. and still it won't always work
Details
Files
doggystyle_piledriver_v0.3.safetensors
Mirrors
doggystyle_piledriver_v0.3.safetensors
doggystyle_piledriver_v0.3.safetensors
doggystyle_piledriver_v0.3.safetensors
doggystyle_piledriver_v0.3.safetensors
doggystyle_piledriver_v0.3.safetensors
doggystyle_piledriver_v0.3.safetensors
doggystyle_piledriver_v0.3.safetensors
doggystyle_piledriver_v0.3.safetensors
doggystyle_piledriver_v0.3.safetensors
doggystyle_piledriver_v0.3.safetensors
doggystyle_piledriver_v0.3.safetensors
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.








