Hey All, presenting my Daphne Blake LoRA from Scooby-Doo! Return to Zombie Island.

Scooby-Doo! Return to Zombie Island – Wikipedia

UPDATE: MOSS VERSION

Hey all! This is my first attempt at creating a MOSS LoRA, so I'd love to see your generations and hear your feedback!

Recommended Strength

0.9

Trigger Word

daphnertzi

Default Outfit

green scarf around neck, purple dress, pink pantyhose, purple high heels

You can refer to my example images for reference. I've used a pruned model with the settings below due to my system's hardware limitations.

However, if your system allows it, I'd recommend trying the following generation settings with the full model:

Recommended Generation Settings

• Sampler: er_sde

• Scheduler: beta

• CFG: 3.0

• Steps: 50

My Settings

• Sampler: er_sde

• Scheduler: beta

• CFG: 3.0

• Steps: 25 (to save time)

I haven't tested this LoRA with the Turbo model yet, so I'd be interested to hear how it performs if anyone gives it a try. Feedback and generation examples are always welcome!

_____________________________________________________________________________________________

Base Models

Anima 1.0 — Base checkpoint (anima-base-v1.0). Not the Aesthetic or Turbo variants.

Illustrious — Trained on WAI-illustrious-SDXL_v11

Recommended Weight

Anima

0.9

Illustrious

0.8

Trigger Word

daphnertzi

Style (Anima and Illustrious)

2d animation, cel shading, flat colors

Appearance (Anima and Illustrious)

orange hair, long hair, purple headband, red lips, purple eyes

Default Outfit

Anima

green scarf around neck, purple dress, pink pantyhose

Illustrious

green scarf, purple dress, pink pantyhose

For full-body generations, you can add:

purple footwear

The appearance and default outfit tags can be used separately depending on what you want to generate. The LoRA should also work with different outfits while retaining Daphne's identity.

Suggested Generation Settings

My testing was done using:

ANIMA Settings with Turbo LoRA

• LoRA weight: 0.9

• Sampler: Euler a

• Scheduler: Simple

• Steps: 12

• CFG: 2.5

Illustrious Settings

Use appropriate quality and negative tags mentioned in the WAI_11 documentation, or you can use positive and negative embeddings. I generated my images using Smooth Positive and Negative embeddings along with Face Detailer for final face refinement.

• LoRA weight: 0.8

• Sampler: Euler a

• Scheduler: Simple

• Steps: 27

• CFG: 4.5

You can refer to my images for more examples. I have tried to tag the training images with accurate metadata and seeds.