Sudachi Anima
A style model with a bright anime style
Join my Discord for everything related to anime models and AI art.
If you’d like to support my work, you can do so on SubscribeStar
Every bit helps me continue making AI models and even an AI VTuber, which is currently in the works.
How to Use
You need to use ComfyUI or Forge Neo.
20-50 steps, CFG 3-6, (i like to use 25 stepts with cfg of 3.5 and er-sde as sampler).
- Put @ in front of artist tags
- Use lowercase for tags and spaces instead of underscores.
- Recommended positive: "masterpiece, best quality, score_7", but since masterpiece is altered, in some cases masterpiece is enough
- Recommended negative: worst quality, low quality, score_1, score_2, score_3, blurry, jpeg artifacts"
- When using a tag that is different between Danbooru and Gelbooru, prefer the Gelbooru version.
Supporters
Special thanks to my supporters:
Estryark, SETI, Kodokuna and dataset captioners mariotheknight and tamashiicolle
Recommended Resolutions for Anima
Here are some resolution suggestions for Anima:
1536 × 640
1344 × 768
1216 × 832
1152 × 896
1024 × 1024
896 × 1152
832 × 1216 — most recommended
768 × 1344
640 × 1536
Larger resolutions are possible.
Note: It is a Diffusion Transformer (DiT) model, it is less restrictive on sizes in compare to a Unet like Illustrious, you can use bigger resolutions.
Version and License Info
V1:
Initial model release.
License: Anima
Description
Initial V1 release:
The style isn’t exactly Sudachi. It’s more of a combination with Hassaku as the base and Sudachi on top. It has a more standard anime style with large, mostly round eyes.
FAQ
Comments (2)
A mxfp8 (int8) file of it is up. Since ComfyUI v0.27.0 official supports int8 convrot now. It offers a smaller file size (less vram) and is faster with a minimal quality tradeoff.
Hi, thank you so much for releasing Sudachi Anima —
it’s one of the clearest and most unique anime-style checkpoints I’ve found so far,
and I would really like to use it as a base model for Anima LoRA training.
I tried training an Anima LoRA with Sudachi using anima_train_network.py (sd-scripts / Anima TrainFlow),
but the training fails with a RuntimeError: Missing keys in checkpoint for several Anima / LLM adapter fields
(for example: x_embedder.proj.1.weight, t_embedder.1.linear_1.weight, llm_adapter.embed.weight, etc.).
From what I can tell, Sudachi works perfectly for inference,
but the current safetensors format seems to be missing a few keys that the official Anima LoRA training pipeline expects.
If it’s at all possible (and if it fits your design goals),
would you consider releasing a “LoRA-training-ready” Sudachi checkpoint that:
keeps the full Anima DiT / LLM adapter structure expected by anima_train_network.py, and
can be used directly as a base model for Anima LoRA training (not just inference)?
I understand this might take extra work and may not be your priority,
but having a LoRA-training-compatible version of Sudachi would be extremely helpful
for users who want to build custom LoRAs on top of your anime style.
Even a short note like whether Sudachi is / is not intended to be compatible with Anima LoRA training,
and whether a training-ready version is planned, would already be very helpful.
Thank you again for all your work on Anima models —
I really appreciate the effort you put into each update.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.