Grace Howard from Zenless Zone Zero, brought to life as a character LoRA by Xander's Workshop!
This LoRA is designed to capture Grace's recognizable appearance while allowing plenty of flexibility with outfits, poses, expressions, and environments.
Whether she's working on machinery at Belobog Heavy Industries, enjoying a relaxing day off, dressing up for an evening out, or posing for more intimate artwork, this model aims to keep her recognizable across a wide variety of scenes.
Available Versions
Anima: The first release, trained for 13 epochs (1,950 steps), with an emphasis on anime-style visuals, character consistency, and outfit flexibility.
Illustrious XL: A separate version will also be available, giving creators another option for generating Grace with their preferred checkpoint.
Recommended Trigger
For the most consistent results, use the following activation tags:
grace_howard, black hair, messy twin ponytails, orange eyes, dark goggles on head, black choker
The grace_howard trigger activates the character LoRA, while the additional appearance tags help maintain her signature hairstyle, eye color, and accessories.
Prompting Tips
Use Danbooru-style tags for outfits, poses, expressions, and backgrounds.
Include the recommended appearance tags for better character consistency.
Experiment with different outfits, from her signature engineer clothing to casual fashion, evening wear, lingerie, and nude artwork.
The LoRA supports both SFW and NSFW character generation.
Try different camera angles, environments, and compositions to explore the model's flexibility.
Created by Xander's Workshop.
If you enjoy the model, consider leaving a like, sharing your generations, or following the Workshop for more character LoRAs!
Description
v1.0 | Initial Anima Release
The first release of Grace Howard for Anima!
Training details:
13 epochs, approximately 1,950 training steps
Text encoder training enabled
10% full caption dropout
Keep Tag Count: 1
Tag shuffling disabled
Danbooru-style training captions
Features:
Strong character reproduction with the recommended activation tags
Flexible outfit generation
Supports casual, workwear, glamour, lingerie, and nude scenes
Consistent anime-style rendering
Tested across portraits, full-body images, and multiple poses
Known limitation:
Using only grace_howard may produce variations in hairstyle, eye color, or accessories. For the best results, include the full recommended activation phrase in your prompts.
This version was selected after testing multiple training checkpoints, with the final 1,950-step model producing the preferred overall visual results.
An Illustrious XL version is also planned for release.













