Animagine 4 Prompt Tutorial for AI Anime Art
Introduction
Welcome to Animagine 4! This guide will help you get started with generating high-quality anime-style images using the Animagine4 model. Follow these tips to ensure your images are both aesthetically pleasing and technically sound.
Prompting
Positive Prompt
Use a structured prompt format: '1girl/1boy/1other, character name, from which series, rating, additional details, quality enhancement'. For example: 'cafe, 1girl, blonde, blue eyes, elegant, bunny cosplay, waitress'.
Negative Prompt
Use a negative prompt to avoid undesirable elements. Example: 'lowres, bad anatomy, bad hands, text, error, missing finger, extra digits, fewer digits, cropped, worst quality, low quality, low score, bad score, average score, signature, watermark, username, blurry'.
Tags
Quality Tags
Control overall image quality and detail level. Available tags: 'masterpiece', 'best quality', 'low quality', 'worst quality'.
Score Tags
Provide nuanced control over image quality. Available tags: 'high score', 'great score', 'good score', 'average score', 'bad score', 'low score'.
Temporal Tags
Influence artistic style based on specific time periods. Supported tags: 'year 2005', 'year {n}', 'year 2025'.
Rating Tags
Control content safety level. Supported tags: 'safe', 'sensitive', 'nsfw', 'explicit'.
Generation Parameters
CFG Scale
Set CFG scale between 4-7 for optimal results. Recommended: 5.
Steps
Use 25-28 sampling steps. Recommended: 28.
Sampler
Preferred sampler: Euler Ancestral (Euler a).
Resolution
Choose from recommended resolutions: 1024x1024 (1:1), 1152x896 (9:7), 1216x832 (3:2), 1344x768 (7:4), 1536x640 (12:5), 896x1152 (7:9), 832x1216 (2:3), 768x1344 (4:7), 640x1536 (5:12).
Advanced Settings
Guidance Rescale
Set to 0.5 to avoid overexposed images.
PAG Scale
Set to 3 for better results.
Prepend Preprompt
Enabled by default. Adds quality enhancement tags to your prompt.
Seed
Set to -1 for a random seed or specify a seed for reproducibility.
Limitations
Prompt Format
Limited to tag-based text prompts; natural language input may not be effective. Prompts have to be in English.
Anatomy
May struggle with complex anatomical details, particularly hand poses and finger counting.
Text Generation
Text rendering in images is currently not supported and not recommended.
New Characters
Recent characters may have lower accuracy due to limited training data availability.
Multiple Characters
Scenes with multiple characters may require careful prompt engineering.
Resolution
Higher resolutions (e.g., 1536x1536) may show degradation as training used original SDXL resolution.
Style Consistency
May require specific style tags as training focused more on identity preservation than style consistency.
Conclusion
By following these guidelines, you can effectively use Animagine4 to create stunning anime-style images. Experiment with different prompts and settings to find what works best for you. Happy creating!