Instructions to use circlestone-labs/Anima with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusion Single File
How to use circlestone-labs/Anima with Diffusion Single File:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
Regarding artist mixing through the cross-attention layer of hook Anima
Anima has caught my attention since its preview 1 release. It is an open-source anime model with a brand-new architecture, but its artist mixing capability is somewhat lacking. Over the past few days, I have pondered many solutions. Ultimately, by hooking into Anima's cross-attention layer, I have achieved controllable mixing of conditioning from multiple artists during the cross-attention stage, solving the problem of LLM TE struggling to correctly understand artist combinations when directly stacking multiple artist tags in the prompt. After testing with my collaborative developers, the results seem promising. I will further optimize this plugin and officially release it in the next few days!(By the way, I haven't used GitHub yet and might need to learn it first, as this is my first attempt to develop a plugin with the assistance of AI)
Very interesting! I’m also working on a similar custom node. It seems to work to some extent, but when the number of tags increases — around 10 or more — they don’t blend well and the result starts to look flat. Have you experienced a similar issue? https://github.com/toyxyz/ComfyUI_toyxyz_test_nodes
Dose your node has github link?
Dose your node has github link?
I am still continuing to optimize this node. Once the optimization is completed, I will upload it to GitHub
Very interesting! I’m also working on a similar custom node. It seems to work to some extent, but when the number of tags increases — around 10 or more — they don’t blend well and the result starts to look flat. Have you experienced a similar issue? https://github.com/toyxyz/ComfyUI_toyxyz_test_nodes
I haven't tested with so many artist tags yet, but the fundamental reason should still be a mathematical issue. Anima cannot achieve the kind of artist mixing like SDXL
Very interesting! I’m also working on a similar custom node. It seems to work to some extent, but when the number of tags increases — around 10 or more — they don’t blend well and the result starts to look flat. Have you experienced a similar issue? https://github.com/toyxyz/ComfyUI_toyxyz_test_nodes
I haven't tested with so many artist tags yet, but the fundamental reason should still be a mathematical issue. Anima cannot achieve the kind of artist mixing like SDXL
By the way, it seems that directly intervening in the text encoder part is not feasible. I have been trying this approach before intervening in the attention layer
Very interesting! I’m also working on a similar custom node. It seems to work to some extent, but when the number of tags increases — around 10 or more — they don’t blend well and the result starts to look flat. Have you experienced a similar issue? https://github.com/toyxyz/ComfyUI_toyxyz_test_nodes
Hello! May I ask which artist tags did you use in these examples?
Edit: Forgot to ask, these ones too - https://huggingface.co/circlestone-labs/Anima/discussions/168#6a12f44034d09655ba70626b
Very interesting! I’m also working on a similar custom node. It seems to work to some extent, but when the number of tags increases — around 10 or more — they don’t blend well and the result starts to look flat. Have you experienced a similar issue? https://github.com/toyxyz/ComfyUI_toyxyz_test_nodes
Hello! May I ask which artist tags did you use in these examples?
Edit: Forgot to ask, these ones too - https://huggingface.co/circlestone-labs/Anima/discussions/168#6a12f44034d09655ba70626b
Hi @baycon , I saw you asked about the artist tags in these examples about 24 days ago. I’m actually trying to track down the same thing — specifically the Figure B (second from the left) in the grid from discussion 168: https://huggingface.co/circlestone-labs/Anima/discussions/168#6a12f44034d09655ba70626b
Just wondering — have you had any luck figuring out the tags since then? Any leads would be super helpful. Thanks!
Very interesting! I’m also working on a similar custom node. It seems to work to some extent, but when the number of tags increases — around 10 or more — they don’t blend well and the result starts to look flat. Have you experienced a similar issue? https://github.com/toyxyz/ComfyUI_toyxyz_test_nodes
Hello! May I ask which artist tags did you use in these examples?
Edit: Forgot to ask, these ones too - https://huggingface.co/circlestone-labs/Anima/discussions/168#6a12f44034d09655ba70626b
Hi @baycon , I saw you asked about the artist tags in these examples about 24 days ago. I’m actually trying to track down the same thing — specifically the Figure B (second from the left) in the grid from discussion 168: https://huggingface.co/circlestone-labs/Anima/discussions/168#6a12f44034d09655ba70626b
Just wondering — have you had any luck figuring out the tags since then? Any leads would be super helpful. Thanks!
Hello @ke6N5 🙂 I actually asked toyxyz at https://github.com/toyxyz/ComfyUI_toyxyz_test_nodes/discussions/64, but they can't remember the artist tag.
It looks similar to ixy's style (Figure B in this thread), but I can't make it look similar to toyxyz's result (Figure B in the other thread) at all.
I tried manually searching for it using style explorers, and out of the 1,000 artist tags I checked, poranka_(porankaran) is the closest one but still doesn't produce the same results.
Here are the style explorers I used:
https://animadex.net/?mode=artists
https://anima.mooshieblob.com/
https://thetacursed.github.io/Anima-Style-Explorer/index.html
I even asked ChatGPT, Claude, Grok, and Gemini, but they were unable to identify the artists.
I'd like to try asking in Danbooru/Gelbooru forums but I'm not sure if they're fine with AI-generated images though, so I haven't.
Would love to hear back from you if you make some progress regarding this.
Very interesting! I’m also working on a similar custom node. It seems to work to some extent, but when the number of tags increases — around 10 or more — they don’t blend well and the result starts to look flat. Have you experienced a similar issue? https://github.com/toyxyz/ComfyUI_toyxyz_test_nodes
Hello! May I ask which artist tags did you use in these examples?
Edit: Forgot to ask, these ones too - https://huggingface.co/circlestone-labs/Anima/discussions/168#6a12f44034d09655ba70626b
Hi @baycon , I saw you asked about the artist tags in these examples about 24 days ago. I’m actually trying to track down the same thing — specifically the Figure B (second from the left) in the grid from discussion 168: https://huggingface.co/circlestone-labs/Anima/discussions/168#6a12f44034d09655ba70626b
Just wondering — have you had any luck figuring out the tags since then? Any leads would be super helpful. Thanks!
Hello @ke6N5 🙂 I actually asked toyxyz at https://github.com/toyxyz/ComfyUI_toyxyz_test_nodes/discussions/64, but they can't remember the artist tag.
It looks similar to ixy's style (Figure B in this thread), but I can't make it look similar to toyxyz's result (Figure B in the other thread) at all.
I tried manually searching for it using style explorers, and out of the 1,000 artist tags I checked, poranka_(porankaran) is the closest one but still doesn't produce the same results.
Here are the style explorers I used:
https://animadex.net/?mode=artists
https://anima.mooshieblob.com/
https://thetacursed.github.io/Anima-Style-Explorer/index.htmlI even asked ChatGPT, Claude, Grok, and Gemini, but they were unable to identify the artists.
I'd like to try asking in Danbooru/Gelbooru forums but I'm not sure if they're fine with AI-generated images though, so I haven't.
Would love to hear back from you if you make some progress regarding this.
Hi @baycon , thanks for the detailed reply — really appreciate you sharing what you've tried so far.
I've also been digging into this for a while, and here's where I'm at:
Closest match so far: Himura Kiseki (比村奇石)
· Colors/tonality are very close.
· But the details don't match: eyes, eyelashes, the hatching on the wings, and the hard-edge shading in the original image — I just can't replicate those with Himura Kiseki alone.
What I've tried:
· Added various style modifiers like lineart, sketch, hard shading, cel shading, flat color, clear lines, sharp edges, cross-hatching, hatching, ink drawing, manga style, strong outlines, rough sketch, no gradient shading — alone or combined, with different weights.
· Still can't get close to the original Figure B.
How far I've searched:
· I spent several days going through the Anima-Style-Explorer, sorted by sample count.
· I went all the way down to the 66-sample tier — by which point I had already checked over 30,000 artists.
· I also used https://animaartistlookup.github.io/Anima-Artist-Lookup/ to check whether a given artist exists in the Anima model and to see their rank among the 59,676 total tags (it doesn't show sample counts, only the rank).
At this point, Himura Kiseki is the closest I've found, but I'm no longer confident it's the correct one.
Also, I asked on another community and someone suggested 1ssakawaguchi (川口イッサ) — the facial expression and nose rendering are somewhat similar, but the rest doesn't quite line up either.
if the original unedited PNG of Figure B still exists, the prompt and workflow might still be embedded in the metadata. You can either use a metadata viewer, or simply drag the image into ComfyUI — it will load the prompt and workflow used to generate it.
If you make any progress, please let me know — and I'll do the same if I find anything new. Thanks again for reaching out!
Update: I've gone through all 42,509 artists with images in Anima-Style-Explorer. Didn't find the match. There are about 17,000 more artists listed only as names in a txt file with no images — I can't visually identify those. Also, it's possible that I missed it because the example images in the explorer might not reflect the style of Figure B consistently, so the artist could still be among those I already checked but I didn't recognize it.
Good day @ke6N5 , thanks for sharing your results.
I haven't tried Himura Kiseki's tag before, so I generated some images with it. And I agree, it's the closest to the style we're looking for.
It's still not quite the same though, so in order to check if I can replicate toyxyz's results, I used the exact parameters their post here (except for the test node - so, I'm basically using the default Anima template bundled with ComfyUI):
I got very similar results (second set of images) to toyxyz's images (first set) even though I didn't have their test node. I still got a similar style to Figure C even though I used an entirely different workflow and parameters for the third set of images.
But when I tested Himura Kiseki's tag on the default Anima template, the results in the second set of images still didn't get close enough to the one we're looking for. The style stays consistent in the third set of images which used a different workflow, so I'm not really sure if Himura Kiseki is actually the exact tag - but it's still the closest I've seen.
I'd like to test all the artists in https://github.com/ThetaCursed/Anima-Style-Explorer/blob/main/Anima2B_Artist_Index_59k.txt but that would take more than two months even with a turbo LoRA at 12 steps on my PC.
Btw, this is a little bit off topic, but I'd like to ask anyway - hope that's fine. haha
Which ones are your favorites among the artist tags in Anima? Here's mine in no particular order:
- piromizu
- poranka (porankaran)
- stanley lau
- sasaoka gungu
- yaegashi nan
- yamashita shun'ya
- tsunako
- tony taka
- amano yoshitaka
- araki hirohiko






