AnimaTensor Pro, is the ultimate anime-themed finetuned SDXL model. The model was trained from Animagine XL 4.0-Zero to converting the model to support V-prediction and Zero-terminal SNR. Trained on 8.4M anime-style images from danbooru with the knowledge cut-off of January 7th 2025 and finetuned for approximately 2800 GPU hours. Similar to the base model, this model was trained using tag ordering method for the identity and style training.
Key Features
V-prediction: Improved noise scheduling and sample quality.
Zero Terminal SNR: Stable, deterministic sampling endpoint.
EQ-VAE: High details in generations, better than SDXL VAE.
High-Fidelity Outputs: Extensive dataset for diverse, detailed generations.
User Guide
Prompt Tags
Quality Tags (at the end of the prompt): masterpiece, best quality, high score, great score, absurdres
Structure Example
1girl, firefly \(honkai: star rail\), honkai \(series\), honkai: star rail, safe, casual, solo, looking at viewer, outdoors, smile, reaching towards viewer, night, masterpiece, high score, great score, absurdres
Negative Prompts: lowres, bad anatomy, bad hands, text, error, missing finger, extra digits, cropped, worst quality, low quality, bad score, signature, watermark, blurry
Sampling Settings
Sampler: Euler a
Inference Steps: 28–35, 30 recommended
CFG Scale: 4–5, 4.5 recommended