RDBT [Anima]
This is a general finetuned + distilled model.
Better quality, better prompt adherence.
Dataset contains ~10k handpicked images with accurate NL captions from LLM, body/hands anatomy. Does not contain any shiny plastic glossy AI image.
Some cover images look ridiculous; they're just for demonstration purposes for corner cases.
I use this model as a clean starting point to stack more LoRAs.
See this page for update log and version info.
For advanced users: RDBT model is trained as LoRA natively. See this page for original LoRA.
For those who don't want to stack style LoRAs and is looking for a out-of-the-box ckpt: RDBT | Anime, which is based on RDBT finetuned base, but optimized for high quality digital art.
Links to base model:
prefix with ym: AnimaYume (hf link) (civitai link).
prefix with b (base), p (preview): Anima pretrained (hf link)
Usage:
Settings:
CFG: 1~3. This model has been distilled. You can disable CFG (CFG 1) and run the model 2x faster. Cover images are without CFG for demonstration. "RenormCFG" node is highly recommended if CFG is enabled (CFG > 1), set "renorm_cfg" value to 1.1.
Steps: 16+
Sampler: Euler (best diversity), Euler a/er_sde etc. (better stability)
Res: 1MP
Prompt:
Always specify style in prompt, or use a style LoRA. Otherwise, you will get random/mixed style. This is a feature, not a bug. This model does NOT have overfitted default style (which ignores prompt and is always active).
Quality tags:
Omit ALL quality tags. You don't need those. 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.
Sharing merges using this model is not allowed.
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: NukeA.I (selling this model behind paywall on tensorart).
I wrote a story about it. Also contains a guide for trainers about "how to bake special trigger word into your model".
Description
FAQ
Comments (18)
Can you please put your fp16 patch back for download? They might have implemented fp16 support to ComfyUI, but it's almost 3 times slower than your patch on my hardware.
Thank god I saved the patch on my own, here: anina_fp16_patch.py · RicemanT/Loras_Collection at main
Does my patch still work? I didn't test.
I deleted my patch because I thought my patch will mess up their implement.
@reakaakasky it still work yeah, i havent test if comfy is slower or not but i'm literally genning rn so your patch seems safer lol
I don't think my patch is needed. I checked and tested comfyui's patch.
Although I'm on 4xxx, so idk.
"3 times slower" sounds like fp32
@reakaakasky On RTX 2070 with your patch is 1.5 s/it. With your patch removed (ComfyUI's patch) it's 3.2 s/it. (made sure it's fp16 computedtype). More than 2x slower without your patch. Significant difference.
@gannibal do you have the --fast arg in your comfy startup command? My patch also enabled fp16_accumlation.
@reakaakasky My bad, i also had xformers turned on, it did not play well with ComfyUI's patch. With pytorch cross attention it's about as fast as your patch was.
I don't know how to do this stuff correctly, with the fp8 version on 4080S I get the same gen times as the bf16.
Does Forge Neo supports it? I'm getting an error : NotImplementedError: "LayerNormKernelImpl" not implemented for 'Float8_e4m3fn'. What do I do wrong?
use default on dtype
@Meowzilla What is this option? On main page or in settings?
Forge Neo supports Anima model now (not sure about quants tho)
@MarkinZzZ My bad, I missed the forge part. My thing is for comfy.
@orhay1, Yeah, it seems that Anima_preview works, but not fp8. Anyway OG anima's images in forge neo (at least what I tried to generate) isn't so good though lol. Maybe we need to wait for a proper release, 'cause I just don't like Comfy for its overcomplication and uneven generation
NotImplementedError: "rms_norm" not implemented for 'Float8_e4m3fn'
Looks like we don't have an active mirror for this file right now.
CivArchive is a community-maintained index — we catalog mirrors that volunteers upload to HuggingFace, torrents, and other public hosts. Looks like no one has uploaded a copy of this file yet.
Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.