Sakana XL (Illustrious) is a fine-tune RouWei base model and merge with my some models, designed to produce high-quality, clean, and visually striking anime 2.5D illustrations and newer dataset.

Key Features

• Supports character tags and artist-style tags

• Delivers clean lines, expressive faces, and dynamic poses

• Balanced for both soft pastel and 2.5D style

Recommended Parameters

• Steps: 20–30

• CFG Scale: 3–7

• Sampler: Euler A

• Upscale: R-ESRGAN 4x+ Anime6B

Quality Prompt

masterpiece, best quality, amazing quality, very aesthetic, high resolution, ultra-detailed, absurdres, newest,

Negative Prompt

modern, recent, oldest, cartoon, graphic, text, painting, crayon, graphite, glitch, deformed, mutated, ugly, disfigured, long body, lowres, bad anatomy, missing fingers, extra digit, fewer digits, cropped, very displeasing, (lowres:1.2, worst quality:1.2, bad quality:1.2), censored, jpeg artifacts, signature, watermark, username, chibi, loli, simple background, conjoined, ai-generated

You’re encouraged to experiment with different parameters and combine LoRA models to fine-tune character traits, outfits, expressions, or stylistic effects.

For inquiries or feedback:

Discord: @LucyMocchi