🐰

DISCLAIMER

Please, read the description below to achieve good results with your generations and not to be disappointed!

This LoRA was trained using 111 crops from the original movie, all collected and captioned manually.

I tried really hard to train this LoRa on Illustrious XL, but the base model could not learn Judy's style and details, so I used YiffyMix v62 instead. I had already tested this model when training Judy Hopps Z1 LoRA.

I recommend to use this LoRA with YiffyMix v62 model, but it also works well with the last YiffyMix v64 version.

Always use 'jdhpps rabbit girl' in the beginning of your prompts!

If you want to generate Judy wearing clothes from the movie, you should use this tags:

For base undercover uniform:

dark blue jacket, light blue shirt, belt, dark pants

For pink 'mommy Judy' outfit:

pink sleeveless shirt, bare shoulders, wristwatch, waist bag, striped pants, pink hair band on ear, pink heart-shaped glasses

For yellow gala dress:

yellow dress, yellow flower, single bare shoulder, yellow flower crown, ears in a bun

To control her ears use:

ears up, ears down, ears in a bun

If 'ears in a bun' is not working properly, add 'ears up, ears down' to negative prompt.

To get Judy's teeth more detailed use:

rabbit teeth

Also add

rabbit tail

to positive/negative prompt for more control, as well as

glare in eyes, eye reflection

Add this in the end of your prompt to increase the quality:

BREAK, 8k, masterpiece, best quality, detailed

It is NOT really necessary to use tags like:

purple eyes, grey fur, two tone fur, animal focus, animal ears, rabbit ears, rabbit nose, whiskers, furry, anthro, Judy Hopps from Zootopia, etc.

Use this as a default negative prompt:

blurry, worst quality, bad quality, lowres, low quality, normal quality, jpeg artifacts, signature, watermark, username, artist name, malformed, deformed, mutated, disfigured, ugly, gross, bad anatomy, bad hands, bad feet, bad face, bad eyes, bad mouth, crooked nose, crooked ears, uneven eyes, missing limbs, extra limbs, 5 fingers, fewer digits, cartoon, 2d, simple, sketch, monochrome, greyscale, comic, anime, manga, render, 3d, painting, drawing, illustration, oil painting, watercolor, bad perspective, human, human face, human hands, human body, human skin, human hair, no fur, smooth skin

My personal usage recommendations

• Sampler: Euler a (euler_ancestral)

• Scheduler: Karras

• Steps: 20

• cfg: 4

• LoRA strength: 0.85

Additionally, use Hi-res fix:

• Sampler: Euler a (euler_ancestral) OR DPM++ 2M (dpmpp_2m) for better detalisation

• Scheduler: Karras

• Steps: 20

• cfg: 4

• denoise: 0.3-0.5

• LoRA strength: 0.85

Feel free to share your thoughts, results, and comments about your experience!