(POV + Sideways) deepthroat (FLUX + SDXL)
Details
Download Files
About this version
Model description
Please Donate Buzz for FLUX Lora Training !
FLUX Sideways Deepthroat :
Works well in a 50/50 mix with sideways deepthroat sevenof9 (/model/656365?modelVersionId=745970).
Trained on FLUX.1 [dev] with 190 photos of Concept of Deepthroating (skewed towards Black Skin) with emphasis on balls-deep. Tested on FLUX 1.D (full) and FLUX fp8 and FLUX nf4 ! Use around strength 0.7-1.2. Use with keywords "ballsxdeep", "deepthroat","tongue out", "performing oral sex" and "from the side"! Also make sure to keep to the structure you see in the examples or below Distilled CFG around 1-4 and CFG 1.0 (without negative prompt). Clipskip 1. Can be used for example as follows:
Positive : (Optional : ballsxdeep, deepthroat, tongue out), {Character and style description} performing oral sex/deepthroating on a {recipient description}'s {skin color (works best for dark skin)} penis, lora:Sideways\_Deepthroat\_FLUX\_concept\_v1\_r1\_merger\_13\_34\_52\_03\_02\_05:1.0
SDXL POV deepthroat:
Lora trained on 88 images on the concept "POV deepthroat" on the SDXL 1.0 base model. Works fine for normal blowjob/deepthroat but isn't good at producing ballsdeep deepthroating even though it has the additional keyword "blsdp". The main keyword is "pov_deepthrt" and action words " deepthroating a man's penis". The keyword can be used to invoke the Lora even when giving a more complicated scenario as a prompt e.g. "in a forest with a crown on her head with green eyes, camera, Nikon Z9, high quality, etc.".The Lora strength should be held at 0.85-1.1.The Lora is a first version and can definitely be improved on.
SDXL Sideways deepthroat
Lora trained on 145 images on the concept "Sideways deepthroat" on the SDXL 1.0 base model. Normal keyword is "sdwys_deepthrt",should produce ballsdeep deepthroating with keyword "blsdp_sdwys_deepthrt". The main concept keyword is "deepthroating a man's penis". The keyword can be used to invoke the Lora even when giving a more complicated scenario as a prompt e.g. "in a forest with a crown on her head with green eyes, camera, Nikon Z9, high quality, etc.". The Lora strength should be held at 0.7-0.95.The Lora is a first version and can definitely be improved on







