RDBT Anime
RDBT finetuned + some style/quality LoRAs. Optimized for digital art.
Built-in style trigger words: @digital art
Built-in style isn't very strong. Additional style LoRA is still recommended.
v2 base
Based on RDBT v2.0 base (without distillation).
Settings: steps: 24+, cfg: 4+.
v2 preview
Based on RDBT v2.0 (preview). Compared to v1. It has a clean and consistent default effect. And much better details and more diverse due to new base.
Settings: steps: 16+, cfg: 1.
v1
RDBT v1.3 finetuned base, and some private LoRAs.
Settings: steps: 16+, cfg: 1.
Sharing merges using this model is not allowed. This is a free model and it will always free. This "restriction" won't affect anyone, it's only aimed at those who steal others' models, merge them, wipe all merge metadata and credits, and sell it as their own model.
FYI: all my models have special trigger words and latent watermark.
Description
FAQ
Comments (5)
Anime v2 is a preview of next rdbt finetuned base v2 (not released yet), which is trained from scratch. New dataset, new distillation settings. More details, natural, diverse, better contrast, sharpness, less sloppiness
We might need to wait awhile for the next base version as they only just released the turbo model (not a fan of it) and aesthetic fine-tune. To be honest they're isn't much they can do as it has a great world and character understanding already. Maybe the text encoder could be better like Qwen3.5 2b or 0.8b Instead of Qwen3 0.6b.
The official anima turbo model is pretty poor they need better distillation techniques or something. Your DMD2 LORAs and the official turbo Lora and even the Cosmos one is so much better.
I don't bother. Unlike all previous base model trainers, the trainer of anima doesn't have the passion at all. Anima is purely motivated by profit.
I was also surprised how bad the final released turbo is. You can copypasta a dmd2 script in a day, but you can't find the right settings unless you keep trying.
That why I'm doing my final fine-tuning on anima, motivated by passion. I'm not happy about my v1 release. Then move on to krea2.
@reakaakasky I really like all your anima releases literally every single one I've been using back on preview 1 and I can tell you and other great creators like duongve13112002 really have alot of passion for this model. Well done for all your hard work. Take care man 🙏
Where is RDBT finetuned base v2?
Not released yet. The goal is maximizing diversity and minimizing AI sloppiness. However this also means low stability. The v2 base has quite high error rate, as expected. I'm still testing it.