HS-Anima-1.0

HS-Anima-1.0 is an enhanced fine-tuned version based on Anima-Base, designed to significantly expand style compatibility while preserving the excellent prompt understanding and character generation capabilities of the original model.

Unlike the original Anima model, HS-Anima-1.0 has been optimized to perform consistently across multiple visual styles, including:

• Anime Illustration
• Painterly Artwork
• Semi-Realistic Anime
• Stylized 3D Rendering
• Fantasy Concept Art
• Character Portraits
• Cinematic Environment Design

The model responds exceptionally well to artist tags, style tags, atmospheric prompts, and natural language descriptions, making it suitable for both simple tag-based workflows and advanced creative prompting.

Key Features

• Enhanced aesthetic quality and composition

• Stronger lighting and atmosphere rendering

• Improved compatibility with artist styles

• Better color harmony and visual storytelling

• Supports anime, painterly, semi-realistic, and stylized 3D looks

• Excellent character consistency

• Strong environmental and cinematic scene generation

• Designed for both character-focused artwork and large-scale fantasy illustrations

Recommended Settings

Resolution:
768×1344, 832×1216, 1024×1024

Steps:
30–50

CFG:
4–5

Recommended Samplers:

• er_sde

• dpmpp_2m_sde_gpu

• euler_a

Recommended Prompt Structure

quality tags → character → artist tags → scene description → style description

Example:

masterpiece, best quality, score_9, newest, highres, safe, 1girl, Fern, @Yoneyama_Mai, flower field, cinematic lighting, atmospheric depth, painterly illustration

Positive Prefix

masterpiece, best quality, score_9, score_8, newest, highres, safe

Negative Prompt

worst quality, low quality, score_1, score_2, score_3, artist name

Prompting Tips

HS-Anima-1.0 performs particularly well with:

• Artist-driven prompts

• Fantasy landscapes

• Atmospheric compositions

• Emotional storytelling scenes

• Anime key visuals

• Stylized 3D aesthetics

• Painterly illustrations

• Character-focused portraits

For the best results, combine strong environmental descriptions, lighting keywords, and style-oriented prompts.

Installing and running

Get the text encoder and VAE from the HuggingFage page.

The model is natively supported in ComfyUI. The model files go in their respective folders inside your model directory:

• HS-Anima-v1.0.safetensors goes in ComfyUI/models/diffusion_models

• qwen_3_06b_base.safetensors goes in ComfyUI/models/text_encoders

• qwen_image_vae.safetensors goes in ComfyUI/models/vae (this is the Qwen-Image VAE, you might already have it)