CivArchive
    Kirazuri (Anima) - v4.0 [anima-base-1]
    NSFW
    Preview 136180088
    Preview 136182913
    Preview 136182929
    Preview 136183551
    Preview 136182954
    Preview 136183612
    Preview 136183632
    Preview 136183753
    Preview 136183743
    Preview 136183716
    Preview 136185254
    Preview 136183727
    Preview 136183791
    Preview 136183797
    Preview 136183942
    Preview 136183960
    Preview 136184040
    Preview 136189322
    Preview 136192369
    Preview 136192974

    Kirazuri (Anima)

    Kirazuri (Anima) is a full fine-tune of the Anima Base v1.0 model by CircleStone Labs focused on several goals:

    • Learn new concepts/styles/characters past the base model dataset cutoff of 2025 September

    • Enhance the model aesthetic guided by manually applied quality, aesthetic, and style tagging

    • Improve rendering and understanding of fine-details through high-resolution training for 1024^2, 1280^2, and 1536^2 resolutions

    Version 4.0 (Latest)

    For in-depth details of training and tooling, see:

    Training Details Summary

    Trainer: diffusion-pipe

    Training device: NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition

    Total training time: ~10 days

    Total samples seen(unbatched steps): ~3,000,000

    Training resolutions:

    • 512^2

    • 768^2

    • 1024^2

    • 1536^2

    Stage 1

    • Samples seen(unbatched steps): ~2,000,000

    • Training time: ~125 hrs

    • Learning Rate: 6e-6

    • Learning Rate Scheduler: Cosine

    • LLM Adaptor Learning Rate: 8e-7

    • Precision: Mixed BF16

    • Optimizer: AdamW8bit with Kahan Summation

    • Weight Decay: 0.01

    • Timestep Sampling Strategy: Logit-Normal

    • Training Resolutions: 512^2, 768^2, 1024^2

    Stage 2

    • Samples seen(unbatched steps): ~1,000,000

    • Training time: ~84 hrs

    • Learning Rate: 2e-6

    • Learning Rate Scheduler: Cosine

    • LLM Adaptor Learning Rate: 2e-7

    • Precision: Mixed BF16

    • Optimizer: AdamW8bit with Kahan Summation

    • Weight Decay: 0.01

    • Timestep Sampling Strategy: Logit-Normal

    • Training Resolutions: 512^2, 1024^2, 1536^2

    Additional Features

    • Masked Training

    • Tag Dropout: 30% with protected first 8 tags

    • Tag Shuffle: Applied to last unprotected tags

    • Natural Language: Short and Long Caption variants

    Changes from Kirazuri (Anima) v3.0

    • Dataset includes recently curated 2,450 images increasing total size from 42,608 to 45,058 images

    • Dataset cutoff now of 29/06/2026

    • Introduced Masked Training for images with simple backgrounds

    • Updated tags+caption variants structure

    Installing and running

    Workflow:

    Reference the anima base instructions. The model is natively supported in ComfyUI. The above image contains a workflow; you can open it in ComfyUI or drag-and-drop to get the workflow.

    Note: Most preview images on the model card additionally use the custom comfyui-prompt-control node for schedule prompting syntax to mix concepts i.e. [word1|word2]
    This custom node is entirely optional but required to exactly recreate the outputs in ComfyUI.

    The model files go in their respective folders inside your model directory:

    Quantizations

    Int8 supported with latest ComfyUI:

    Distillations

    4 Step CFG 1 Turbo LoRA distilled with DP-DMD - diversity-preserved few-step distillation:

    Generation Settings

    Trained in mixed resolutions for the majority of training, and finished with dedicated high resolution training.

    Previews are generated mostly at 1280^2 e.g. 1520x1040 or 1536^2 e.g. 1248x1824 resolutions.

    30-50 steps, CFG 4-5.

    Same samplers as recommended for the base model work, I like to use:

    • er_sde: the recommended default for 30-50 steps.

    • sa_solver_pece: can converge with good detail in 15-20 steps.

    [Optional] ComfyUI-Autocomplete-Plus prompt input assistance

    An optional file danbooru_tags_kirazuri_4.txt is included to provide prompt tag autocomplete.

    This file contains metadata that is derived from public sources for prompt assistance only, and is intended to be used with the ComfyUI-Autocomplete-Plus extension.

    Rename the file to danbooru_tags_kirazuri_4.csv and place it in your ComfyUI/custom_nodes/comfyui-autocomplete-plus/data directory.

    You should see suggestions when writing comma separated tags in ComfyUI:

    Prompting

    Like the base model, this model is trained on booru-style tags, natural language captions, and combinations of tags and captions.

    Tag order

    [quality/meta/safety tags] [character] [series] [artist] [1girl/1boy/1other etc] [general tags]

    Mostly the same order as the base model, only the [1girl/1boy/other etc] groups position is towards the end in this models dataset.

    [quality/meta/safety tags] [character] [series] [artist] tag groups are also not shuffled, so their order may have some influence on generations.

    Quality and Aesthetic tags

    Human score based: masterpiece, best quality, very aesthetic, aesthetic

    The very aesthetic and aesthetic tags are where this model diverges from the base, with the intent these can be used to guide the model toward a different aesthetic - a kind of house model bias.

    Meta tags

    absurdres, official art, etc

    Styles

    painterly, chiaroscuro, ligne claire, flat color, no lineart, blending, etc

    traditional media, oil painting \(medium\), watercolor \(medium\), etc

    Known Limitations & Issues:

    Some concept bleeding and instability is noticeable when using short prompts, especially tag-only prompts.

    Longer tag strings and natural language prompts describing the image in detail should help with this.

    This reflects how the model was trained with a combination of natural language and tags.

    Recognitions

    • Thanks to CircleStone Labs for the Anima base models.

    • Thanks to tdrussell of CircleStone Labs for the diffusion-pipe trainer.

    • Thanks to bluvoll for support using their fork of diffusion-pipe.

    • Thanks to narugo1992 and the deepghs team for open-sourcing various training sets, image processing tools, and models.

    License

    This model is released under the same license as the base model.

    See the base model for details of the CircleStone Labs Non-Commercial License.

    Built on NVIDIA Cosmos

    Description

    Kirazuri (Anima) 4.0 Training Diary

    This version is trained from the Stage 1 checkpoint of version 3.0, repeating only the Stage 2 training with several differences:

    - Updated dataset with cutoff of 29/06/2026

    - Revised training hyper-parameters

    - Masked training

    These changes aim to:

    - Introduce new knowledge to the model

    - Improve stability for anatomy and complex compositions

    While maintaining the models original goals to:

    - Learn new concepts/styles/characters past the base model dataset cutoff of 2025 September

    - Enhance the model aesthetic guided by manually applied quality, aesthetic, and style tagging

    - Improve rendering and understanding of fine-details through high-resolution training up to 1536^2

    FAQ

    Comments (11)

    victorscottlandJul 9, 2026
    CivitAI

    Does the Anima Turbo LoRA works with v4.0? It doesn't seem to work with v3.0

    deitychaserJul 9, 2026· 2 reactions

    yea i noticed the same.

    motimalu
    Author
    Jul 9, 2026· 1 reaction

    Same as v3.0, the Anima Turbo LoRA is about half as effective.
    It can still be used with a CFG of 2 and 12 steps though, v4.0 example: https://civitai.red/posts/29686966

    deitychaserJul 9, 2026· 1 reaction

    @motimalu any plans on cfg distiling kirazuri for a compatible turbo lora?

    motimalu
    Author
    Jul 10, 2026· 5 reactions

    @deitychaser Will see what I can do, can't make any promises though

    motimalu
    Author
    Jul 11, 2026· 4 reactions

    A Turbo LoRA distilled from this checkpoint is available here now:
    https://civitai.red/models/2769422/kirazuri-anima-turbo

    stygianwizard42Jul 9, 2026
    CivitAI

    what's the optional txt file

    motimalu
    Author
    Jul 9, 2026

    Hello, the optional file danbooru_tags_kirazuri_4.txt is included to provide prompt tag autocomplete.

    This file contains metadata that is derived from public sources for prompt assistance only, and is intended to be used with the ComfyUI-Autocomplete-Plus extension in ComfyUI.

    To use it, rename the file to danbooru_tags_kirazuri_4.csv and place it in your ComfyUI/custom_nodes/comfyui-autocomplete-plus/data directory.

    You should see suggestions when writing comma separated tags in ComfyUI.

    Lost_HeartJul 10, 2026· 1 reaction
    CivitAI

    Best Anima finetune ever! Love this!

    1164912321Jul 14, 2026
    CivitAI

    Well, to be honest, I didn't see any significant optimization. In my opinion, the very special advantage of your model is that it has some minor style adjustments. Your fine-tuning of the style has a very good effect on the Lora effect. It makes the originally plain style slightly more delicate and gentle, and makes your mod stand out uniquely among the mods with generally plain styles in Anima. However, in the new 4.0 version, this advantage has disappeared. Although I don't want to influence your future decision-making, I sincerely hope you can maintain this feature. Thank you very much.

    motimalu
    Author
    Jul 14, 2026

    Hello, thanks for sharing that
    A difference between compatibility with LoRA trained on the Anima Base 1 model from version 3.0 to 4.0 might be down to the difference in how far the checkpoints had diverged from the base during training; version 4.0 received ~500,000 more samples than version 3.0.

    I'm not sure if I would prioritize LoRA compatibility - it is expected that they should work best with the model they are trained on - but better compatibility with the Anima Base 1 LoRA's might be achieved by using a merge of this model with it.

    Checkpoint
    Anima
    by motimalu
    Download (Beta) View on CivitAI
    digital illustration
    video games
    base model
    anima
    manga
    anime
    aesthetic

    Details

    Downloads
    3,295
    Platform
    CivitAI
    Platform Status
    Available
    Created
    7/9/2026
    Updated
    11/11/2026
    Deleted
    -
    Trigger Words:
    masterpiece
    best quality
    very aesthetic

    Files

    anima-kirazuri-v4.safetensors

    Size:
    3.89 GB
    SHA256:
    4beecddce25245d94d4ec7dbd179cf06f334d4a97fd0a6b6435fe8aab3b875af

    Mirrors

    HuggingFace (3 mirrors)
    CivitAI (1 mirrors)

    qwen_3_06b_base.safetensors

    Size:
    1.11 GB
    SHA256:
    cd2a512003e2f9f3cd3c32a9c3573f820bb28c940f73c57b1ddaa983d9223eba

    Mirrors

    HuggingFace (145 mirrors)
    CivitAI (275 mirrors)
    ModelScope (1 mirrors)
    ModelScope CN (2 mirrors)

    danbooru_tags_kirazuri_4.txt

    Size:
    531.13 KB
    SHA256:
    bff9a724d78d5db7c3ef6e2d6bd442eddd1e904792d29dd7c24b7f444ff9585a

    Mirrors

    CivitAI (1 mirrors)

    anima-kirazuri-v4-int8-convrot.safetensors

    Size:
    2.17 GB
    SHA256:
    0a1f3b75ae6e84a2c5c616f4d2051ff25bfb72c8d7d9a1c633cd56c960f4c528

    Mirrors

    HuggingFace (2 mirrors)
    CivitAI (1 mirrors)

    qwen_image_vae.safetensors

    Size:
    242.05 MB
    SHA256:
    a70580f0213e67967ee9c95f05bb400e8fb08307e017a924bf3441223e023d1f

    Mirrors

    HuggingFace (363 mirrors)
    CivitAI (310 mirrors)
    TensorFiles (1 mirrors)
    ModelScope (3 mirrors)
    ModelScope CN (7 mirrors)

    Available On (2 platforms)

    Same model published on other platforms. May have additional downloads or version variants.

    SeaArt
    Kirazuri (Anima) - v4.0 [anima-base-1]
    PixAI
    Kirazuri (Anima) v4.0 - v4.0 [Anima-Base 1]