This is a 二次元-style boy model featuring Chinese traditional clothing, or hanfu. Originally created for personal use, the model is now being made available to others. As many existing open source models focus on female characters, we wanted to offer a male alternative that can be used for a variety of creative projects. This model offers a range of possibilities for exploring new expressions of gender and identity, and we hope it will be a valuable resource for anyone interested in the intersections of technology, art, and culture.
The introduction about this model is generated by chatgpt.
Description
FAQ
Comments (6)
Thank you for making this. I'm happy for anything focusing on male characters. ❤️ It's a nice style. I may extract it to a Lora and try using it with a different checkpoint.
The problem is what's the base model of this checkpoint. I don't think it's trained from official SDv1.5.
@Picaluo Right, that makes sense. When I tried extracting it, I got a LoRA that sort of works, but also doesn't quite. It's vey picky about the strength, and it really only looked okay at a very narrow range, which varied from one checkpoint to another. That's why I haven't ever shared it.
@dita I tried to extract it base on novelai model, dimension 128. The clothes are gorgeous but this lora would slightly change the character's face. The eyes will be larger and face maybe shorter.
@Picaluo You were able to do better than I was, then. I don't remember which base model I used when extracting it. I don't think it's a bad thing that it changes the characters' faces. Can you not turn the strength down a bit and still get the clothes without the face issue? Another idea is to use the Lora Block Weight (LBW) extension and try to identify the blocks that will give you the style/clothing without changing the face. If you want to share your extracted LoRA, I'd like to play with it to see if I can get it to cooperate.
great model :DDD are you planning on releasing a safetensors version?
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.