Triggers:
Characters-
-Nino
-Miku
-Ichika
-Yotsuba
-Itsuki
Styles-
-Season1
-season2
INFO-
1000~ Screenshots from the Anime
18,000 Steps/50 Epoch
Trained on NAI
Weights: I didnt really experiment, go ahead
Description
FAQ
Comments (8)
not very responsive when combined with other LORAs. tried to lower the weight as low as 0.6
Great LoRA!
To ensure that the character's traits do not blend together, what points did you pay attention to during the training process?
nothing just wd tagger on 0.35 strength but added the hair color and the ornaments myself
@wiz_ Thank you!
After tagging with Wd tagger, did you only add tags for hair color and accessories etc..? Did you perform any tag cleanup? (For example, using the Dataset tag editor extension to delete unnecessary tags.)
Did you use any scripts or did you manually take screenshots to prepare the dataset for training?
@WarriorMama777 nah idk how to do that tag cleanup shit, also no i just torrented both the seasons and took screenshots, shouldnt have done that it took like 2-3 hours to screenshot and separate them lmao
@wiz_
lol
I was curious because this LoRA was able to generate characters without mix their distinct features, resulting in clean outputs. It was helpful to know. Thanks.
@WarriorMama777 wait ur the one who made abyssorangemix?wtf
hey man, which VAE u recommend for using this LoRa?
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.