11/30 Update
Universal Show Pro Pose Control Version V3 Released

Experience link: https://www.runninghub.cn/post/1994708121975201794/?inviteCode=kol01 rh009

New features in V3 version:

1. Brand new model, supports direct 2K resolution output!

2. Generation speed increased by 50%

3. Restrictions on lingerie, special clothing such as sexy outfits lifted

4. Adapted to flat images, mannequin images, and real person wearing images

11/25 Update
Universal Show Pro2 The Official Sequel to Universal Show (Non-Pose Control) Alternative to Banana Pro
Experience link: https://www.runninghub.cn/post/1992860351335297025/?inviteCode=kol01 rh009

Features:
1. Restrictions on lingerie, special clothing such as sexy outfits lifted
2. Direct 2K output, stable quality
3. Adapted to flat images, mannequin images, and real person wearing images
4. Fast speed, continuous generation of 2K takes only 30–40 seconds

10/28 Update

Universal Show Pro Version Released, Model Capacity Doubled, Compatible with Most Clothing Types, More Realistic Textures, Higher Restoring Accuracy

https://trydiffusion.com/ /toolsRet?toolId=2422474498359813

09/25 Update

Qianwen Shoe Model Master Released, Specially Optimized for Shoes, Adapted to E-commerce Scenarios, Performance Surpasses Banana

https://www.runninghub.cn/post/1970926881845481473/?inviteCode=kol01 rh009

New features in V2 version:

1. Upload a reference pose to ensure the model aligns with the reference pose during generation, enabling precise pose control

2. Some detail optimizations

Features in V1 version:

1. Perfect alternative to Banana's clothing fitting function, including single-image fitting and dual-image (specified appearance) fitting, no longer reliant on Google API

2. Adapted to Qwen LoRA, can generate different styles based on personal aesthetics and needs, better aligning with Chinese aesthetics

3. Natural, can directly generate in Chinese, and allows precise control over output resolution and quantity, with maximum resolution exceeding 1.5K

4. Adapted to acceleration model, single-image generation takes less than one minute

Contact the author vx: xdtr6688

mail:15058643272@163.com