Fan benefits: Click the avatar in the upper right corner, enter the invitation code under "Invitation Code" to receive 1000 RH coins. Log in daily to get an additional 100 coins.
Invitation code: rh v0990

The domestic site no longer supports local decoding. For local decoding, please visit the overseas site:

Registration link: Overseas version RH (Decoding test workflow, requires VPN): https://www.runninghub.ai/user center/1936823199386537986/webapp?inviteCode=rh v0990
Fan benefits: Register to receive 1000 points, log in daily to get 100 points, enjoy the power of 4090! Experience the super charm of 48G! If you don't have 1000 points, bind the activation code in the upper right corner and enter the invitation code rh v0990 to claim!

If you want to create a model but don't even have material for a single image, can the model still be made?

The answer is yes.

Ms. Aelda Wanzhou doesn’t have a single decent image available online. This model was created by extracting details from a two-second official PV video through meticulous research. It has been a long time since I spent two hours creating a model, and the quality is exceptionally high.

This model uses materials generated by the Tutu Image and Video Prompt Reverser and refines them through detailed editing. It was trained using the Tutu Trainer.

The Tutu Image and Video Prompt Reverser is an AI-powered image and video annotation tool, optimized for model training annotations. It can batch-generate high-quality training datasets with one click and quickly reverse-engineer prompts from images and videos.

The Tutu Trainer is a free LoRA training tool that supports model training, dataset management, model resource management, parameter presets, training checks, and pre-release organization. Its goal is to enable creators to train high-quality LoRA models faster and more stably without having to repeatedly tweak environments and parameters.

Official website: https://zhaotutu.xyz/
Tutu Trainer: https://zhaotutu.xyz/downloads/tututrainer/
Tutu Annotator: https://zhaotutu.xyz/downloads/tutuannotator/
Bilibili: https://space.bilibili.com/431046154
YouTube: https://www.youtube.com/@zhaotutu/videos
X: https://x.com/bigpox12
Telegram: https://t.me/zhaotutu
TikTok: https://www.tiktok.com/@zhaotututu
GitHub: https://github.com/zhaotututu
Email: bigpox12@gmail.com
QQ: 331506796

Model Introduction
This is a LoRA model for Aelda Wanzhou, a character from "Zenless Zone Zero.".

The model is primarily used for stabilizing Aelda Wanzhou’s character appearance and costume direction.

During testing, a weight of around 0.75 proved to be relatively stable.

Trigger word: tutuwanzhou Recommended weight: 0.5 - 1.0 Suggested starting weight: 0.75

For more stable reproduction of the character and costume, it is recommended to place the trigger word at the beginning of the prompt and include key costume elements.

Recommended prompts

tutuwanzhou, best quality, score_7, score_9, sensitive, very aesthetic, ultra detailed,
cleavage, blue_eyes, cape, white_hair, heterochromia, black_gloves, yellow_eyes, fingerless_gloves, hair_ornament, large_breasts, very_long_hair, thigh_strap, asymmetrical_clothes, bodysuit, strapless,
high_heels, asymmetrical_footwear, mole_on_thigh, boots, black_footwear,

Negative prompts:
worst quality, low quality, score_1, score_2, score_3, blurry, jpeg artifacts,

For more examples of generated images, please refer to the example images.