Download
1 variant available
Overview
This is a fine-tuned SD 1.5 checkpoint specifically crafted for high-quality anime generation. By merging AnyLoRa and NothingV4, this model achieves a unique balance: it retains the vast style compatibility and LoRA-friendliness of AnyLoRa while adopting the clean lines, vibrant colors, and modern "digital art" feel characteristic of NothingV4.
Whether you are looking for flat 2D illustrations, detailed "Makoto Shinkai" style backgrounds, or character-focused portraits, this mix provides a stable and versatile base.
Key Features
High LoRA Compatibility: Designed to be a "neutral" base that takes LoRAs exceptionally well without color distortion.
Clean Linework: Inherits NothingV4’s ability to render clean, sharp edges and professional-grade shading.
Versatile Stylization: Capable of handling everything from retro 90s anime to modern high-budget cinematic styles.
Anatomical Stability: Improved consistency in character proportions compared to standard base models.
Recommended Settings
To get the best results from this checkpoint, I recommend the following:
VAE: [Recommended: vae-ft-mse-840000-ema-pruned or AnimeVAE]
Sampler: DPM++ 2M Karras or Euler a
Steps: 20 - 30
CFG Scale: 5 to 7.5
Clip Skip: 2
Resolution: 512x768 (Portrait) or 768x512 (Landscape). Hires. fix is highly recommended for detailed shots.
Prompter's Tip
This model responds very well to descriptive artistic tags. You don't need "masterpiece" or "best quality" as much as other models, but adding tags like flat color, scenic, or detailed background will help steer the aesthetic effectively.
License & Usage
This model is focused on adult-oriented artistic content (NSFW capable) but is not intended for generating illegal or non-consensual imagery. Please respect the terms of use of the original parent models (AnyLoRa and NothingV4) when using this mix.
Show me your generations in the reviews! I’d love to see what you create.