概要 / Model Summary
PrismPalette ANIMAは、Anima系モデルをベースとしたマージモデルです。美しいライティングと豊かな質感表現、そしてクリアな空気感が特徴です。
PrismPalette ANIMA is a merged model based on the Anima model series. It features beautiful lighting, rich texture expression, and a clear, atmospheric depth.
📜 ライセンス / License
CircleStone Labs Non-Commercial License v1.2
📂 ファイルの配置 / File Placement (ComfyUI)
本モデルのご利用には、以下のコンポーネントをダウンロードし、それぞれのディレクトリに配置してください。
To use this model, please download the following components and place them into their respective directories:
Model
File: PrismPalette_ANIMA_v1.safetensors
Path: ComfyUI/models/diffusion_models/Text Encoder
Repository: Hugging Face（circlestone-labs/Anima）
File: qwen_3_06b_base.safetensors
Path: ComfyUI/models/text_encoders/VAE
Repository: Hugging Face（circlestone-labs/Anima）
File: qwen_image_vae.safetensors
Path: ComfyUI/models/vae/
📃 サンプルワークフロー / Sample Workflow (ComfyUI)
Repository: Hugging Face（Noginowa/merge-experiment）
File: comfyui-anima-workflow_basic.json
⚠ ご利用上の注意 / Usage Notes
❗ご利用開始前に必ずお読みください： 最適な結果を得るため、以下の「おすすめ設定」と「おすすめプロンプト」を必ず確認してください。
❗MUST READ Before Use: To achieve optimal results, please ensure you check the "Recommended Settings" and "Recommended Prompts" sections below.
✨ おすすめ設定 / Recommended Settings
CFG scale: 4-5 (おすすめ/Recommended 4.5)
Steps: 20-40(おすすめ/Recommended 30)
Sampler: ER SDE BETA / Euler A Normal
💬 おすすめプロンプト / Recommended Prompts
Positive:
masterpiece, best quality, score_9, score_8, score_7, year 2025, newest, highres, absurdres, very aesthetic, sceneryNegative:
nsfw, lowres, bad quality, worst quality, score_1, score_2, score_3, poorly drawn, flat color, sketch, jpeg artifacts,
mutated, ugly, disfigured, long body, bad anatomy, bad hands, missing fingers, extra fingers, extra digits, fewer digits,
simple background, transparent background, blurry, furry🖼 サンプル画像 / Sample Images
サンプル画像はPrismPalette ANIMA v1で生成し、アップスケールをしています。
Sample images were generated with PrismPalette ANIMA v1 and upscale.
🔗リンク / Link
Hugging Face: https://huggingface.co/Noginowa
Bluesky: https://bsky.app/profile/noginowa-ailab.bsky.social
Description
Merge models:
Thank you to all the original model creators.
Merge Tools:
Thank you to all the tool creators.
References:
Crody's Model Merge Guide v3.0 // Team-C - Section 2: Block Weight Ratio
FAQ
Comments (5)
I said i would test it, and must come back to say it's really good ! If someone hesitate pls don't ! Give it a try you will not regret it ! :D
Thank you so much! I'm thrilled to hear you like it so much. And wow, your image is amazing! 🥰
The style of the output image is somehow determined by the sampler settings. Using Turbo LoRA with Euler A shows very little difference when compared to WAI-anima. To achieve the advertised artstyle, ER SDE shall be used.
On my side Euler A is working well, but i don't use turbo lora on top of the checkpoint. You tried without ? i'll give a try to Er-sde also, thanks for the info.
perfect!