I trained this model with art from @Deaver0501 on twitter.
Upon testing I found myself consistently going back to an earlier version that I had trained. This version, V1, does not stand alone as a style nearly as well, but seems to play with other lora nicer than V2.
---V1---
This version was the first attempt I had when training this model. It needs help from other lora, and has problems with anatomy and artifacting. However it plays nicer with other lora than V2 in my opinion and figured I should add it here.
Necessary Trigger Word: DeaverStyle,
These tags will also help, but aren't necessary: virtual youtuber, thick outline, simple background,
With some testing, I've found that every once in a while it will put animal ears on the characters without being prompted. This is probably because of the large amount of cat girls/dog girls in the training data. If this becomes a problem a simple "animal ears" or "cat ears" in the negative should solve the problem.
---V2---
V2 is a far better "stand-alone" style lora than V1. It is far improved in terms of anatomy and arifacting, and tends to keep its style far more consistently. However personally, I don't believe it interacts with other Lora as well as V1.
This model works best at 1.0
Necessary Trigger Word: DeaverStyle,
These tags will also help, but aren't necessary: virtual youtuber, simple background,
With some testing, I've found that every once in a while it will put animal ears on the characters without being prompted. This is probably because of the large amount of cat girls/dog girls in the training data. If this becomes a problem a simple "animal ears" or "cat ears" in the negative should solve the problem.