PLEASE UPLOAD YOUR IMAGES. I'M CURIOUS HOW WELL THIS'LL DO!! LET'S GET MORE BARA CONTENT HERE FOR Z-IMAGE TURBO/Anima!

Anima Description:

Anima is a lot quicker and excellent at following prompts! May need to turn strength down unless you really want the giant size. Does drawing, comic, anime, digital styles best. A more realistic Anima checkpoint might be needed if that's what you want? It was trained mainly on original drawings, so might not do photorealism well?

Z-Image Turbo Description:

I've been actually looking into training a Remerting-style LoRA for Flux earlier this year, but Z-Image Turbo has shown a lot more potential. This first version on here is technically the 3rd iteration I've trained for this style, making slight improvements each time. The goal has been to try achieving the artist's hyper-muscular style mixed in with photorealism that Z-Image Turbo does well. Some of the original drawings from the author were used in the dataset; however, most of the images were created through:

HomoVeritas NAI Checkpoint

Illustrious Remerting Style LoRA created by Detoxic (he took it down a while ago, but his other LoRAs are awesome!)

V3 (Versions 1 and 2 are okay, but not planning to upload):

Optional Trigger: r3m3rt ("hyper-muscular male" works usually since it was used in most captions)

Positive: hyper-muscular male, photorealistic, [any other details such as the size of muscles like the pectorals might be bigger here if you mention it], other desired content such as clothing, lighting, art style, facial expression, etc.

Negative that I normally use: old, ugly, liver spots, female, girl, child, children, woman, women, extra limbs, low detail, watermark, blur, bad anatomy, lowres, deformed, blurry

Take a look at the image grids for a better idea on strength of the LoRA. 0.5 seems to be the sweet spot without affecting the art style too much. Lots of the natural details ZiT does really well seem to be removed when the LoRA is stronger. This may be because the HomoVeritas Checkpoint will make the human body have little to no skin imperfections unless prompted....

Z-Image Turbo has trouble with male anatomy. There were nude images used in the training, so you can try to prompt it, but don't expect it to be perfect! You may need to use another LoRA that focuses specifically on the penis. It can generate large bulges depending on what size you want. ;)

I have NO IDEA how this will do with generating women since they are not part of the artist's style! (Also, there's a ton of female content on this site here already.....)

Please support Remerting: Twitter | Bluesky | Patreon | Furaffinity

I will remove this LoRA if it is requested by the original author. Do NOT sell this content!