Hello welcome!
Here I leave a small introduction of the LoRA:
I had already uploaded a version in fp32 but I retrained it with other parameters :3
IL | V1
I was reviewing my folders and found this dataset (technically it doesn't give the same result because the sd version was trained with Meina and this one was trained in Illustrious 0.1, although I was hesitating whether to train it again in IL 2.0)
SD1.5 | V1
After several tests I can say that something complex came out.
While it's understandable that hands are a big part of AI mistakes, this is one of the worst I've ever had (without help).
I sincerely recommend using a LORA for hands or simply use it as a tensioner adetailer to generate faces, this is the only thing I see useful. Of course, if you manage to master it, it will give you good results, but... did I win another Christmas star? :/
Description
I like how it looks with my checkpoint, but I'm generating a dataset so it looks the same in other models. However, if you can't wait, use my executor style (no triggers needed). This will make some models perform like mine in coloring. (v1.5)
FAQ
Comments (2)
for at least get the same result as you, your Checpoiny is mandatory. IF only you could make a lora style of this image, that we can use.
It's true, I just want them to use my checkpoint because I'm a bad person... Jokes aside, although I trained it in Illustrious 0.1, I liked the result generated with the checkpoint, but if you use the style executor, it should work just as well (I tested it with wai nsfw 14).
For the next version, I'll use those results to train it in Illustrious 2.0 or 0.1 (I think 2.0 is better). Here are the examples: https://civitai.com/posts/17561279