LATEST: X-RAY 2.5
if you wanna enter my small discord server to discuss about my LoRA's: https://discord.gg/jqmqnkdNBZ
I am sorry for my trouble caused, after hours of troubleshooting with the model, the model won't upload and I did a mess... I am really sorry. I clicked save for every change I made. I will finish this page later.
For my old user's: I know the 2.3 and 2.4 model are gone, but I will try to upload them later, and also, this model, outperforms the older models by 60% if you follow my steps: try to generate an image as you normally would with xray and stuff, then save that image and interroage it in deepdanbooru, then copy the generated prompt in text2image and try to generate, it will be much better. I will update this model page later, and again, I am sorry. Also, I can't upload more showcase because of an error when trying to Publish the LoRA's. Sorry.
Edit: Added them back.
It's recommended to iterrogate an anime image before with DeepBooru.
Recommeneded to add the related prompts about x-ray in general ( x-ray view, cross-section, vaginal, uterus, cervix, penis hitting uterus, internal cumshot, etc,...) and Use either Euler at Euler A at around 20 or 50 steps
Have fun!
PS: Might release a improved version in the future if it's requested enough + fertilization/impregnation model
Done!
If you would like to buy me a small coffee to keep me up at night, it will be much appreciated!!
Description
FAQ
Comments (15)
With this new 2.0 version, what if we don't use Automatic's suit? Outta luck?
What do you mean? You wanna use it outside Automatic1111? Like Invoke? If so... well, yeah, no luck.
Invoke, indeed.
Ahh thats a real shame, no chance for a regular version as previous releases, even if at a later date? I hope going forward they aren't all like this :')
My old method which I used Dreambooth doesn't work anymore, I tried Stable Tuner, cloud Dreambooth, everything, but only LoRA training worked. So I might release models like this until there is gonna be an alternative. I hope someone will make an extension in which I can transform LoRA's models into regular sized models compatible with Invoke or any webui
I see, thanks for your efforts regardless, I'll keep watching and waiting <3
Please provide a normal, full-blown, non-Lora checkpoint, thanks
At the moment I can't train with dreambooth to generate full .ckpt models, I wish I could, but my dreambooth has problems, so I uploaded this type of model since there was no way around. I am sorry. If there is an method in which I can convert LoRAs to ckpt I will convert all the models into .ckpt in the future.
Amazing, cool to see an application of Lora and looking forward to trying that out thanks for sharing!!
Can this model do Deepthroat?
Well.. my other model has deepthroat, check it
You can change your model type from checkpoint to lora. To make it more clear.
how?
Editing model --> Second field "Type". Change it from checkpoint to LORA.
I don't know, I trained this model using a LoRA script, but I think there was a choice to make it into a .pt LoRA file, I tried but it doesn't work, also.. I don't know to what are you referring to since I am not that good in stable diffusion, could you please clarify more?
Your downloadable X-Ray Hentai 2.0.safetensors is a LoRA model. You also descripe in your description, how to use this LoRA. Your model guides a main model with its trainig to something, like your xray vision with a addjustible strength (those sliders, next were you select your model). Because your model is a LoRA, i suggested to change your civit site type from checkpoint to LORA.