Катерина в лисьей форме! В целом, работает нормально, и вблизи, и на расстоянии, но если обрезает голову, прописывайте (cropped head):1.2 в отрицательном поле. Просьба не судить строго, первый раз НиРА обучаю! Надеюсь, Вам понравится!Catherine's fox form! In general, it works fine, both up close and at a distance, but if it crops the head, write (cropped head):1.2 in the negative field. Please do not judge harshly, it's the first time I'm training a LoRA! I hope you like it!Btw, thanks to Tobias for the images, they made for a great dataset!

Catherine Sue (Delusional Monthly Magazine)

Now it includes Ed! Finally Catherine and Ed can be together!! &lt;3&lt;3&lt;3Note, Pony, the underlying model, doesn't know what a monocle is, and it won't always work, so use it at a high value, like (monocle):2.0Use "size difference" to make sure she's bigger than him!Example of how to generate them both (use the Regional Prompter extension):score_9, score_8_up, score_7_up, best quality, anime, kitchen background, red wallpaper, thick lines, ghibli, sink, side view, (duo)BREAK&lt;lora:caterina6-v1.2:0.5&gt;, edward-chi, smiling, blue eyes, monocle, from side, blue blazer, looking right, white turtleneck, earringsBREAK&lt;lora:caterina6-v1.2:0.5&gt;, tall athletic kemono female anthro fox sitting, grey fur, extremely fluffy, bracelet, side view, caterina-sue, detailed purple eyes with yellow sclera, orange t-shirt, upper body, close up, speaking, talking, open mouth, fox snoutNote that you have to decrease each individual Lora such that they sum to 1, otherwise it will be overexposed! Maybe even 0.4! Try it yourself!

caterina6-v1.3.safetensors

caterina6 v1.3

NoobAI-eps-v1.1 - v1.0

caterina6 v1.2

caterina6 v1.0

Катерина в лисьей форме! В целом, работает нормально, и вблизи, и на расстоянии, но если обрезает голову, прописывайте (cropped head):1.2 в отрицательном поле. Просьба не судить строго, первый раз НиРА обучаю! Надеюсь, Вам понравится!Catherine's fox form! In general, it works fine, both up close and at a distance, but if it crops the head, write (cropped head):1.2 in the negative field. Please do not judge harshly, it's the first time I'm training a LoRA! I hope you like it!Btw, thanks to Tobias for the images, they made for a great dataset!

Now it includes Ed! Finally Catherine and Ed can be together!! <3<3<3

Note, Pony, the underlying model, doesn't know what a monocle is, and it won't always work, so use it at a high value, like (monocle):2.0

Use "size difference" to make sure she's bigger than him!

Example of how to generate them both (use the Regional Prompter extension):

score_9, score_8_up, score_7_up, best quality, anime, kitchen background, red wallpaper, thick lines, ghibli, sink, side view, (duo)

BREAK

<lora:caterina6-v1.2:0.5>, edward-chi, smiling, blue eyes, monocle, from side, blue blazer, looking right, white turtleneck, earrings

BREAK

<lora:caterina6-v1.2:0.5>, tall athletic kemono female anthro fox sitting, grey fur, extremely fluffy, bracelet, side view, caterina-sue, detailed purple eyes with yellow sclera, orange t-shirt, upper body, close up, speaking, talking, open mouth, fox snout

Note that you have to decrease each individual Lora such that they sum to 1, otherwise it will be overexposed! Maybe even 0.4! Try it yourself!

Files with SHA256: cd7370863eb47946860c96864a121a72aabc52eb8f109b820fb8e118f1247599