HAV0X1014/Z-Image-Turbo-KF-Bat-Eared-Fox-LoRA

Bat Eared Fox (Kemono Friends)

Prompt

anime image of bat_eared_fox_kemono_friends sitting in a cafe with a steaming cup of coffee on the table in front of her. she is looking to the side at the viewer with a light smile.

Negative Prompt

blurry ugly bad

Prompt

a real photo of a woman cosplaying bat_eared_fox_kemono_friends walking through a datacenter holding a tray of cheese

Negative Prompt

blurry ugly bad

Prompt

anime drawing of bat_eared_fox_kemono_friends in a mall excitedly holding a blue elephant plush that she won from a crane game that says "PHP to win!" on top of it that is in the background. she is standing on one leg and winking.

Negative Prompt

blurry ugly bad

Model description

A LoRA I trained on 64 images of Bat Eared Fox using ai-toolkit. I trained it to 3000 steps and used the default config settings - except I used adafactor as the optimizer because I trained this on my 7900xtx. Learning rate was 1e-4.

trigger word is "bat_eared_fox_kemono_friends". Images were captioned by me by hand with a mix of danbooru tags and natural language. The dataset is reused from the one I used on the Chroma version of this LoRA. It seems to only want to do anime images even if you specify "photo" or "cosplay."

I just wanted to see if it was possible to do this on my GPU and apparently it is. It took FOREVER to do though. For some reason, it used up all 24gb of my vram and ran at 5s/it, which is SUPER slow. It worked though, in spite of all of this.

its ok i guess

Trigger words

You should use bat_eared_fox_kemono_friends to trigger the image generation.

Download model

Download them in the Files & versions tab.