Nova Anime AM
Nova Anime Anima is Nova Anime: Anime/2.5D/3D checkpoint model for Anima
All images on example were created with diffusers with custom png tags
Rules
You cannot use the generated images for commercial use if it's not edited (or just turning it to black and white)
You can share images without any restriction if you don't monetize it
Advertising this model to outside is always welcome
Recommend Settings
Sampler: Euler a Normal
Steps: 20~30
Clip Skip: 1-2
Denoising Strength: 0.65 - 0.8
CFG Scale: 4~6
Prompt: masterpiece, best quality, score_9, score_8, score_7, year 2025, newest, highres, absurdres, very aesthetic, scenery, {Prompt}
Negative Prompts: worst quality, low quality, early, old, score_1, score_2, score_3, cartoon, graphic, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured, long body, bad anatomy, bad hands, missing fingers, extra fingers, extra digits, fewer digits, cropped, very displeasing, artist name, blurry, jpeg artifacts, lowres, censor
Description
Changes
Much stable posing / structures
More details
Stable coloring
Changed base into Anima base v1.0
Anima base v1.0 based
FAQ
Comments (20)
I've been trying to get this model to work in my local gen but I've had no luck. All I keep getting are grey images. anyone got any tips?
What software did you used?
If you're using Comfy, try this:
Nova Anime AM.safetensors to ComfyUI/models/diffusion_models
qwen_3_06b_base.safetensors to ComfyUI/models/text_encoders
qwen_image_vae.safetensors to ComfyUI/models/vae
Both qwen3_06b_base and qwen_image_vae can be downloaded on official Anima's page under Required Components
https://civitai.com/models/2458426/anima?modelVersionId=2945208
@Crody Ah, I see. that might be the issue. My pc can't handle Comfyui (not enough space) I use auto1111 for local when I'm not using the generator on this site.
@yeguangenjoyer You'll have to use Forge Neo to load Anima models, idk if you're using base A1111 or sum, but for Anima models, Forge Neo and Comfy are the way to go
Clip skip on qwen text encoder?
Anima is Diffusion Transformer not Unet architecture, Clip Skip is not advised and won't have any effect either.
@Jedziah yeah that's my point
Are you using the same dataset from Illustrious? it is killing the art style knowledge from the base....
I didn't use / train the model using Illustrious versions but I felt the same way for the style knowledge as well
Regarding this, I made v2.0 and will release it on the first day of the next month
So for every one who interested: maximal size for Hires.fix/image2img is 1920px per side (may be different in Comfy)
adherence is average, this one produce very good backgrounds
Brilliant as always! Waiting for Nova 3DCG AM sooo baaad
This model is perfect!
Looking forward for v2
After benchmarking over 500+ generated images, based on Anima2B Basic, the overall screen saturation is way too high. When paired with the intense contrast that accentuates its geometric rendering, it causes serious eye strain over long sessions. Without specific tags, the model tends to significantly "lolify" / baby-face characters compared to NovaAnimeXL. However, injecting tags like "adult woman/mature woman" violently forces the character into full adulthood, completely missing that "high school / late teen" middle-ground age range.
I've noticed that adding it after "@" has some effect. It might be a symbol referring to the overall art style rather than the artist's name.
It's beautiful, but its versatility is reduced.
This applies to Anima-group issues in general, but prompts related to the image style, such as "anime screencap," hardly work.
este modelo se puede usar en confyui?
Si se puede, tienes problemas en hacerlo?
@sovietlat se puede usar en confyui?
@deivytejada0124 yes
Sí, debes cargarlo con un clip para el prompt por lo general Qwen 3 de 0.6b y el VAE lo consigues en HugginFace - ambos en la carpeta text_encoders y VAE. Los pasas a tu instalación de comfy https://huggingface.co/circlestone-labs/Anima/tree/main/split_files