Jane Doe | Zenless Zone Zero | Qwen-Image
Bring the mysterious, playful, and dangerously charming Jane Doe from Zenless Zone Zero into your Qwen-Image generations!
This character LoRA was created by Xander's Workshop, using a custom dataset of 109 manually captioned images to capture Jane's distinctive appearance while allowing plenty of flexibility with outfits, poses, and environments.
Unlike the more anime-styled SDXL versions, this Qwen-Image edition produces beautiful semi-realistic illustrations, featuring detailed skin textures, soft lighting, expressive faces, and richly rendered materials.
Whether you're creating urban fashion, casual scenes, swimwear, elegant lingerie, or mature artwork, Jane is ready for her next adventure!
Model Information
Base Model: Qwen-Image
Model Type: Character LoRA
Training: Locally trained using OneTrainer
Dataset: 109 manually captioned images
Training Duration: 20 epochs
Trigger Word:
janezzzCaption Style: Manually written visual descriptions and tags
Recommended Usage
Trigger Word: janezzz
Qwen-Image works best with natural-language prompts describing your character, pose, clothing, setting, and lighting.
Example Prompt
Create a detailed semi-realistic illustration of janezzz, Jane Doe from Zenless Zone Zero, standing in a neon-lit city street at night. She has short dark-purple hair with teal highlights, blue-green eyes, distinctive rat ears, and a long gray-purple tail ending in a pointed silver metal tip with red inlay. She wears a glossy black evening dress with high heels and looks toward the camera with a mischievous smile. Her tail curves behind her in a graceful S-shape. Use cinematic purple and pink lighting, detailed urban scenery, and a full-body three-quarter camera angle.
Tail Prompting Tips
Jane's distinctive tail is an important part of her appearance.
For better consistency, describe it clearly:
"A long gray-purple tail extending naturally from her lower back, curving behind her body, and ending in a pointed silver metal tip with red inlay."
Specifying the tail's position and direction can help produce more reliable results, especially in complicated poses.
Features & Testing
This LoRA has been tested with a variety of scenes and outfits, including:
Signature Jane Doe appearances
Casual clothing and streetwear
Elegant dresses and fashionable outfits
Full-body portraits and close-ups
Swimwear and bikini scenes
Lace lingerie and boudoir settings
Nude character studies and mature-themed artwork
Multiple camera angles and lighting conditions
The LoRA performs particularly well at maintaining Jane's facial features, hairstyle, ears, and metallic-tipped tail.
Known Limitations:
Complex poses may occasionally introduce anatomical errors or tail-placement issues. The LoRA can also carry Jane's animal features into unrelated characters, even when the trigger word is omitted.
Experimenting with LoRA strength and prompt wording may help improve consistency.
Showcase Images & ComfyUI
The showcase images were generated using ComfyUI.
Original PNG images include embedded generation metadata and workflows.
You can download an original PNG and drag it into ComfyUI to load its saved workflow, provided the workflow metadata is preserved.
Additional models or custom nodes may be required to reproduce the results.
Other Versions
Prefer a more traditional anime aesthetic?
Check out the Jane Doe Anima edition, also available from Xander's Workshop!
A Pony Diffusion XL edition is also in development.
Each edition is trained and tested separately to take advantage of its base model's strengths.
About Xander's Workshop
Created by Xander's Workshop, where we experiment with character LoRAs, anime artwork, realistic illustrations, and plenty of creative possibilities!
If you enjoy this model, please consider leaving a like, adding it to your collection, following the Workshop, or sharing your generations.
Both SFW and NSFW creations are welcome where permitted by the platform.
Have fun creating! ❤️
Description
Version 1.0 | Qwen-Image
Initial release of Jane Doe from Zenless Zone Zero, trained for Qwen-Image.
Trained locally using OneTrainer on a custom dataset of 109 manually captioned images, completing 20 epochs and 2,180 training steps.
This LoRA focuses on preserving Jane's distinctive appearance, including her purple hair, teal highlights, blue-green eyes, rat ears, and signature gray-purple tail with its pointed metallic tip.
Features:
Semi-realistic character rendering with detailed skin, hair, and materials.
Strong facial and character consistency.
Support for casual outfits, formal fashion, swimwear, lingerie, and nude artwork.
Flexible poses, camera angles, lighting, and environments.
Support for both SFW and NSFW generations.
Recommended prompting: Natural-language descriptions with the trigger word janezzz.
Known limitations: Occasional animal-feature bleed into unrelated characters and minor anatomical or tail-placement inconsistencies.
Showcase PNGs include generation metadata and embedded ComfyUI workflows where available.
Created by Xander's Workshop. ❤️












