This is a LoRA for Kitsuneno Ne キツネノ ネ(@kitsuneno_ne)'s style. He is the auther of the doujin circle Rush Rise Line.
I will remove this if requested by the original artist.
Info:
Trigger word: @k1tsun3n0n3
Note:
This LoRA produces very unstable outputs (especially for the private parts). Online generation is not recommended.
This LoRA may output monochrome images. Please input prompts related to color to avoid this.
Description
Trained on Anima Base v1.0.
FAQ
Comments (7)
wow! :O
Dude I literally thought about training a style lora of this artist yesterday and now I see this. You read my mind
Ahhh, man, I wish I could've seen the one you made. Honestly, this LoRA turned out pretty bad...
@toghashie441 Maybe I'll give it a shot. I didn't even gather a dataset yet tho so it might be a while till then
@toghashie441 Now that I took a look at it, I don't think I'll train one because there barely are any neutral images. I actually didn't know the artist before, but instead I used your Illustrious version and I absolutely loved it. Never actually took the time to check out the artist until now
@Lopodelas Ah, I see... So you liked the LoRA I made of him without actually knowing the artist himself. I'm sorry about that, as I think some of his artwork includes pretty extreme fetishes. And I guess now you understand why my LoRA turned out so terrible. It's really a mystery to me why training the Illustrious version went so much better compared to the Anima version.
@toghashie441 Nothing for you to be sorry about.
I feel training on Anima is very hard if the dataset is not the best. I'm struggling with training a style lora of ZONE's flash games with a similar issue regarding the dataset (that being not too many different images and many same-ish images) but it's so hard to get it acceptable. It's getting to a point where I'm seriously considering adding AI generated images to the dataset but I really want to avoid it to ensure the quality stays the best it can be. :/
Meanwhile on Illustrious it always turns good or acceptable. I feel like Illustrious is just so much more reliable to train on as well as a lot more forgiving in regards to dataset size and variety. I always try to get the best quality possible but Anima is just a pain in the ass to train for some reason.