<h2 id="rdbt-anima">RDBT [Anima]</h2>A general finetuned model, better overall quality, better prompt adherence, less artifacts and errors.Dataset contains ~10k images, handpicked by <s>professional</s> amateur artists, captioned by latest LLM. Accurate body/hands anatomy. Complete background.Some cover images look ridiculous; they're just for demonstration purposes for corner cases.<hr />Update log and version info: <a target="_blank" rel="ugc" href="https://civarchive.com/articles/30582">link</a>.For advanced users: RDBT model is trained as LoRA natively. <a target="_blank" rel="ugc" href="https://civarchive.com/models/2364703">link</a>.<hr /><h3 id="usage:">Usage:</h3>Settings:<ul><li>Base version: has a suffix "base". Not distilled. CFG 4 and steps 24+ is recommended.</li><li>Distilled version: CFG: 1~1.5. Steps: 16+. FYI: CFG 1 = Disable CFG = run the model 2x faster. Cover images are without CFG for demonstration. "RenormCFG" node is highly recommended if CFG is enabled (CFG &gt; 1), set "renorm_cfg" value to 1.1.</li><li>Sampler: Euler (best diversity), Euler a/er_sde etc. (better stability)</li><li>Res: 1MP</li></ul>Prompt:Always specify @style in prompt, or use a style LoRA. Otherwise, you will get random/mixed style.Quality tags:Omit ALL quality tags. The fine-tuning dataset has higher quality than "masterpiece". Thus quality tags don't have effects. Omitting those redundant tokens allows LLM to pay more attention on other words.<hr /><h3 id="faq:">FAQ:</h3>Why I won't bake a default style:<ol><li>If a model has a "default" style (ignores prompt and is always active). We call it "overfitted". Technically, it's not a feature, it's a bug. This means the training dataset is not big enough, and the model has lost its ability to generate other styles. I know some default styles look very cool, but this is not my goal.</li><li>The goal of RDBT model is to "perfectly" reproduce styles. As some of you know, I use this model to aid in drawing. I need uniqueness (diversity) and consistency in style.</li><li>This model has excellent compatibility. You can stack any type of Lora on it. After all, it's you who decide what style you want, not me.</li></ol>RDBT can't remove nor resist overwhelming Al slop effectsToo long, see: <a target="_blank" rel="ugc" href="https://civarchive.com/models/2356447/rdbt-or-anima?dialog=commentThread&amp;commentId=1288701">link</a>.My homemade distillation vs Anima official turbo model:Official turbo: 8 steps. High stability, low diversity. Adds a strong 2D anime filter that will overwrite style and cause style shifting.My homemade distillation: 16 steps. High diversity, low stability. Minimal style shifting.<hr /><h3 id="sharing-merges-using-this-model-is-not-allowed.">Sharing merges using this model is not allowed.</h3>This mode is a free and it will always be free. This "restriction" won't affect anyone. It's only aimed at those who steal others' models to sell.Known model thieves: <a target="_blank" rel="ugc" href="https://tensorhub.art/u/613019935106639991">NukeA.I</a> (selling this model behind paywall on tensorart).Because he is such a jerk I wrote a <a target="_blank" rel="ugc" href="https://civarchive.com/articles/31410/dev-story-rdbt-2">story</a> about it. Also contains a guide for trainers about "how to bake special trigger word into your model".

RDBT | Anima

rdbtAnima_ym15V23Base_fp16.safetensors

rdbtAnima_ym15V23Base_int8.safetensors

ym1.5 v2.3 base

b1 v2.3

b1 v2.3 base

ym1f v2 base

b1 v2 base

b1 v2

b1 v1.2 turbo

b1 v1.2

b1 v1.1

b1 v1.0 turbo

b1 v1.0 16-step

b1 v1.0 8-step

ym1f v0.43 [int8]

ym1f v0.43

b1 v0.39.b

ym0.5 v0.39

b1 v0.37.1

b1 v0.35

b1 v0.32.b

p3 v0.32

p3 v0.29

p3 v0.27

p3 v0.25

v0.25

p3 v0.24f 4-step dmd2

p3 v0.24f dmd2 turbo

p3 v0.24f 16-step dmd2

p1 v0.20f cfg

p1 v0.19f cfg b

p1 v0.12f cfg

fp16 patch v1.2

fp16 patch

p1 fp8

Files with SHA256: f1da66880cdf640ac63d537d365ac3576625042bc6f440824f179ef7dd3fd021