INFO

Reccomended strength: 1

Only the first image is upscaled. The rest are generated with base anima, nothing else.

Sampler

Er_sde gives a more sketchy style while Euler A is smoother, both good.

Useful tags

Use tags like vibrant color, fullcolor for colored images. Use "cross-hatching" and "spot color" to strengthen the style (optional and spot color can make the image black and white sometimes). Use Monochrome, Greyscale, - For black and white (looks good as most of training data was monochrome). Fullcolor still looks good.

Glowing eyes, - For the Eye glow effect

No trigger word

v2.0

Trained with more images at almost double the steps. higher quality.

v1.0

Less steps, fewer images used to train.