LoRA for having subjects present their ass. It is recommended that you inpaint faces after generation.
Trained for realistic models, but can work with anime models as well.
Recommended weights: 0.4 - 0.7
If you like this LoRA and wish to support me, consider buying me a beer at https://www.buymeacoffee.com/WaifuWizardWong đș Your support is much appreciated!
Be sure to check out my other LoRAs!
Description
Trained on 100 images
FAQ
Comments (8)
When I load it, it says "Droptop". Is it possible that you have mistakenly uploaded another copy of the Droptop Lora instead of the Present Ass Lora?
Could you elaborate on what you mean by "when you load it"? Seems to be working okay for other users. Downloads with the correct filename for me. Would you be able to provide the steps you took when you saw the name "droptop" instead?
@WaifuWizardWong Same issue here for some odd reason. Loading it through AUTOMATIC1111/stable-diffusion-webui extra networks button below generate button. First I press blue civitai refresh to load the trigger words buttons looking like lightbulbs, then i press the lightbulb, trigger word "presentass, " added to positive, then i press the Lora that is named "presentass" and has the correct display image so that it will get added to positive prompt, in positive prompt "<lora:droptop:1>" gets added. So something is weird when it comes to the model. Might this be some mix up with registering the models on civitai and some details getting mixed up causing civitai helper extension to add wrong model info to droptop lora?
@ishimarukaito565Â There's a weird behavior with this lora where it will sometimes appear in the metadata as "droptop" instead of presentass after the initial render but it appears it still works for hiresfix or whatever.
I have some errors in the images, they look grainy, even with the negative prompt, bad resolution and a little blurry, with the lora in the folder, in extra options, I have all the parameters in 0.45.
I have found that using multiple loras tend to degrade image quality if all their weights sum up to more than 1. I try to keep my weight total less than (or close to) 1. For example, I might use weights of 0.45 if I have 2 loras, but will lower them all to 0.35 if I add a third lora or 0.25 if I add a fourth. Not sure if this would solve your problem, but I mostly encounter grainy images when my total lora weight is too high.
The only backdoor LoRA that works with my model+LoRA setup! <3
Any chance you could recreate some of your ass loras for FLUX please?
They were honestly some of the best for 1.5 so it would be great to see them for FLUX as well! :)