v2 version has iterated on personification effects, making it more biased towards 2.6D realistic portraits, whereas v1 is more geared towards simplicity and style. Each has its unique charm! Try comparing them below is some of my scattered thoughts, feel free to skip if you want ~

lbc_Rebirth means rebirth, and for this model, I went through many... (laughs)

The most notable one was the hardware computation bottleneck (yes, I'm a poor person...), limited by a device with only a GTX 1060 Ti 16GB graphics card. The training process's computing speed was significantly restricted.

However, it is precisely in these difficulties that I treasured each failure and treated my dataset with care, because I know that the quality of a model is determined by its dataset. My model is also an inheritance of some excellent datasets from the lbc series (SD1.5 version), covering various aspects such as detail representation, artistic photography, simplification, and fantasy writing.

This comprehensive coverage makes lbc_Rebirth a multifaceted and all-encompassing model.

"Rebirth" implies that I've undergone every moment and storm in the process of creating this model. By overcoming all difficulties, I've endowed this model with new life, a more powerful and comprehensive expression.

Usage notes:

• Clipskip: 1-3, I usually use 1 or 2
• Sampling iteration steps (Samplingsteps): 30+
• You can enable high-definition repair, algorithm recommendation is: 4x-UltraSharp
• Width (Width): **0 or other values are okay
• Height (Height): 960 or other values are okay
• CFGScale: 7

If you like my model, please give me plenty of likes!

I'm grateful beyond words. If you want to train a specific style or character, feel free to leave a comment below, and I'll be happy to help you for free, sharing beautiful things together!

For conditionally-privileged friends who want to download other models that are not suitable for release in mainland China, please go to C-site support: https://civitai.com/user/fuaneng/models