If you like the model, please add image to the model page + Add Post and leave a review, it really helps. The buzz makes it easier for me to make improvements/train new models.<hr />ZIT:Prompting: gagged with a ball gagPONY V3:Prompting: large ball gag, (color) ball gag, gaggedMuch better version.I noticed most of my generation buzz comes from the pony version of this model, which has allowed me to train more models. This is a better version as thanks for using my models. should be able to control the color more consistently, plays better with other loras and gives cleaner outputs.<hr />PONY V2:Prompting: b4llg4g, large ball gag, gagged,This is a much cleaner version of the pony model that won't introduce blurriness like V1<hr />FLUX:Prompting: (Subject) gagged with large (color) b4llg4g in mouth with rivets and leather straps<ul><li>Capable of decent nudity without support from other LoRas</li><li>Can do medium/normal size by not mentioning "large" and saliva with "saliva. (Better at full strength)</li></ul>If trying to do non-realistic style and having issues, either increase strength of other lora or lower this lora's strength to 0.85<hr />

Large Ball Gag (ZIT |FL | PD)

Pony version, I don't have much experience training pony models, sometimes the results come out blurry, especially if it's not a close-up. If anyone is experienced with pony and doesn't mind sharing good settings for the on-site trainer, I would really appreciate it and will retrain the model. I don't feel like the dataset is the issue, I have &gt;50 high quality images in different contexts and angles. With this model, I can take the best images between both versions and increase the data set to 100 images.

Files with SHA256: 36b8977e337412c4e69395c915a5f4392e45ef5d91a8d4c6215b226fed9ac800