It is forbidden to sell the images generated by this model for profit, and it is forbidden to sell this model. Thank you
禁止将此模型生成出的图片进行出售盈利行为，禁止出售此模型。谢谢
Please read about this version
The model generalizes and transfers very well, and you can use it on almost any submodel (except SDXL).
If you want to restore the original painting style, you need a lot of prompt words to correct, the model's data set contains all nekopara images, the game illustrator artbook and fanclub cover, theoretically it can restore all images
Recommended weight 0.5-1.1
prompt
heterochromia,nekoyezi,1girl, cat girl, bell, solo, long hair, jingle bell, animal ears, neck bell, <lora:super-yezi-SUrh-v9:0.85>
Description
Please reply to the image, because my ability to write prompt words is very poor, I did not test this model. It is based on the previous version, expanded the dataset, added regularization to repair the eye. That's the theory, but I didn't test it
FAQ
Comments (5)
Hello again!
Sorry for the newbie question. I seem to do everything like you, but the result I get is worse than the preview in terms of quality. Please, if you don't mind, tell me what I'm doing wrong. Here are my settings:
Model: Anything V5;
Sampling method: Euler a;
Steps: 20;
Resolution: 512X768;
CFG: 7;
Using Hires,fix
Upscale by: 2;
Hires steps: 0;
Denoising strenght: 0.7;
Upscaler: Latent
This is not only with this Lora, but with all your others. I would be very grateful if you answer
Use the adetailer plugin. If the seed is locked and the image is not the same, see if the model and clip and tag are complete. My negative tip word "图片修复" is easynegative emb model, badhand is also emb model and these are all searchable on this site
@Rnglg2 My results are much better now, thank you!
By the way, what are your near future plans for new models? Damn, I think I've gotten a little too obsessed with it lately, I didn't think the AI was that good.
@infodech Currently I am iterating on the coconut lora model, after training this I will start training the maple lora model