Sign In

Mafuyu — LoGG_Realism_2.5D

0

Not loaded

Download

1 variant available

fp16 SafeTensor

Mafuyu_V1_HassakuXLIllustrious.safetensors

Half precision, best balance • 377.33 MB

Verified:

Type
LoRA
Stats

44

0

Generation

Not loaded

Reviews
Published

Oct 5, 2026

Base Model

Illustrious

Training
Steps: 1,320
Epochs: 22
Usage Tips
Strength: 0.8
Hash
AutoV2
69EED50554
Trigger Words
mafuyu
default creator card background decoration
Followers - 30

30

Likes - 141

141

Mafuyu_20261005_094200_314359837.png

Mafuyu — LoGG_Realism_2.5D

This is my Mafuyu LoRA, trained on HassakuXL with the main focus being on pushing Mafuyu toward a 2.5D realistic/semi-realistic appearance while keeping her recognizable features and overall identity consistent.

The LoRA was trained with a variety of outfits, poses, environments, lighting conditions and compositions, including both fantasy and more contemporary styles. The goal was to make the character work beyond a single outfit or setting while maintaining the same facial identity, pink hair, blue eyes and characteristic braided hairstyle.

Because this LoRA is specifically designed around a realistic 2.5D aesthetic, it works especially well with detailed skin, realistic hair strands, natural lighting and high-detail fabric rendering.

Trigger word: Mafuyu

Base Model: HassakuXL

For the best results, use the same HassakuXL checkpoint used during training, as other checkpoints may change the character's facial features, proportions and overall rendering style.

  • Trigger: Mafuyu

  • Base: HassakuXL

  • Style: 2.5D / semi-realistic

  • Best results: Use the same HassakuXL checkpoint

  • Strength: Adjust according to how strongly you want the character identity to appear

The LoRA is intended to be flexible enough for character portraits, full-body compositions, cosplay, fantasy outfits and different visual settings while retaining Mafuyu's core appearance.