Raehoshi Illust XL

an enhanced iteration built upon the Illustrious XL model. It aims to elevate the visual style by addressing some of the limitations in the original, such as oversaturation and artifact noise. While these issues are not entirely eliminated, noticeable improvements have been made, and further refinements will continue. The goal is to deliver a more polished, balanced output while staying true to the strengths of the base model.

Enjoying my work? Consider supporting the development of my models by buying me a coffee on Ko-fi

Recommended setting

Positive prompt :

masterpiece, best quality, good quality

Negative prompt :

lowres, bad quality, worst quality, bad anatomy, sketch, jpeg artifacts, signature, watermark

Steps : 20-28

CFG : 5-6

Sampler : euler_a

Supported resolution :

1024 x 1024, 1152 x 896, 896 x 1152, 1216 x 832, 832 x 1216, 1344 x 768, 768 x 1344, 1536 x 640, 640 x 1536

Training Details

The model was developed using a two-stage fine-tuning process. In Stage 1, new series and characters were introduced into the model. Stage 2 focused on fixing issues and enhancing the overall style for improved output.

Stage 1

Dataset : 31k
Hardware : 2x A100 80gb
Batch size : 32
Gradient accumulation steps : 2
Learning rate : 6e-6
Text encoder : 3e-6
Epoch : 15

Stage 2

Dataset : 2.5k
Hardware : 1x A100 80gb
Batch size : 48
Gradient accumulation steps : 1
Learning rate : 3e-6
Text encoder : disable
Epoch : 15

List of New Series/Characters Trained:

zenless zone zero series character
wuthering waves series character
hololive from justice and dev is
HSR character from 2.2 - 2.5 version
tokidoki bosotto roshia-go de dereru tonari no alya-san series

For character prompts, please refer to the Danbooru site for accurate tags and references.

License

Fair AI Public License 1.0-SD

Raehoshi illust XL - v4.0

Raehoshi illust XL - v4.0

Sampler

Steps

CFG Scale

Seed

Clip Skip

Size