Warning: This is an EA version, small data set, so it has low chance of success images, only about 15%. Do make your expections lower...

Usage: doggy style sex front view, <any woman, any role, any place, any role>

Tigger words: doggy style sex front view,

Example:

doggy style sex front view
Negative prompt: illustration, 3d, sepia, painting, cartoons, sketch, (worst quality:2), (low quality:2), (normal quality:2), lowres, bad anatomy, bad hands, normal quality, ((monochrome)), ((grayscale:1.2)), futanari, full-package_futanari,, newhalf, collapsed eyeshadow, multiple eyebrows, vaginas in breasts,holes on breasts, fleckles, stretched nipples, nipples on buttocks, analog, analogphoto, signatre, logo,2 faces

Steps: 30, Sampler: DPM++ SDE Karras, CFG scale: 7, Seed: 3363772375, Face restoration: CodeFormer, Size: 512x768, Model hash: fc2511737a, Model: chilloutmix_NiPrunedFp32Fix, Clip skip: 2, AddNet Enabled: True, AddNet Module 1: LoRA, AddNet Model 1: doggy_style_sex_front_view.v1.0-sd(ffb3b1675ba4), AddNet Weight A 1: 1, AddNet Weight B 1: 1

Recommend weights: 1

Description

100 epochs, 57 files, 512x768

FAQ

Comments (7)

cerinthusMay 6, 2023· 3 reactions

CivitAI

I think you're selling yourself short with expectations. I had a lot of luck with my pictures, maybe 1 or 2 out of 20 turned out bad. Cheers for making this.

XX007May 6, 2023

I totally agree, the performance is very acceptable. The key is a checkpoint that is good with porn stuff. The flexibility isn't great since it's clearly overtrained on Blacked stuff, but the performance is very good. Creating good photorealistic sex LoRa isn't easy at all and this one works.

XX007May 6, 2023

CivitAI

I let a positive review since it's a good prototype.

bomber055May 7, 2023· 20 reactions

CivitAI

Crank the LoRA weight up to 1.4 and you can clearly see the Blacked logo on every picture. You really need to get rid of the watermarks on your training data.

XX007May 9, 2023· 4 reactions

Buzz Wrong diagnostics! It means it's been trained on a too small and undiversified dataset. You're focusing on a superficial consequence, not on the deep root.

disapproving_daddyMay 18, 2023· 5 reactions

Both are really true.

This LoRA is a good start, but it should be trained on a larger dataset, AND the dataset should be cleaned of logos, watermarks, etc.

zoesixxxMay 5, 2024· 9 reactions

CivitAI

I used "doggy style sex front view" and ended up with literal dogs in the scene (not participating, thank god). Probably dependent on the checkpoint, but def not what I was looking for.

LORA

SD 1.5

by dalixyz

Download (Beta) View on CivitAI

porn

action

doggystle