Mimi Pearlbaton (Re:Zero) — Illustrious SDXL Character LoRA

中文说明 | English (current)

An unofficial character LoRA for Mimi Pearlbaton from Re:Zero − Starting Life in Another World, trained on waiIllustriousSDXL v17. Publishes formal E1–E7. Recommended default: E6 (mimi_pearlbaton-000006.safetensors) at strength 0.8. Dataset: abbuibuibui/MimiPearlbaton-Dataset.

Fan-made derivative; character rights remain with their owners.

Files

Label File Role
E1 mimi_pearlbaton-000001.safetensors Alternate. Early.
E2 mimi_pearlbaton-000002.safetensors Alternate. Early.
E3 mimi_pearlbaton-000003.safetensors Alternate.
E4 mimi_pearlbaton-000004.safetensors Alternate.
E5 mimi_pearlbaton-000005.safetensors Alternate. Matrix candidate.
E6 mimi_pearlbaton-000006.safetensors Recommended default at 0.8.
E7 mimi_pearlbaton.safetensors Alternate final formal epoch (unnumbered; unique SHA).

E6 SHA-256: 6852b02dc80beac7136c4cf37f73947a2c0d79d91947dd9532396ec9f1298d12

Showcase

Direct E6 @ 0.8 outputs. Euler a, 28 steps, CFG 4.5, CLIP skip 2, 832×1216.

Portrait White outfit full body
Face Main

Night

Trigger Words and Settings

Role Token
Identity mimi_pearlbaton (orange hair, animal ears, green eyes, short hair)
Outfit mimi_white_outfit (53/53)

Recommended strength 0.8. Example:

<lora:mimi_pearlbaton-000006:0.8>,
masterpiece, best quality, very aesthetic, absurdres, anime screencap,
1girl, solo, mimi_pearlbaton, mimi_white_outfit,
orange hair, animal ears, green eyes, short hair,
full body, looking at viewer, outdoors

Why E6

Seven formal epochs screened @ 0.8 (face + white outfit). E5–E7 entered the weight matrix (9 scenes × 0.4/0.6/0.8/1.0). E6 best overall balance at 0.8. No alt outfit in this set.

Full notes: evaluation/横向测评记录.md.

Public Horizontal Evaluation

Round 1 @ 0.8

Front portrait White outfit
R1 face R1 outfit

Round 3: E5 / E6 / E7 × weights

Face Three-quarter
01 02

03

04

07

08

09

10

11

Training

  • 53 stills, single outfit folder; captions start with mimi_pearlbaton, mimi_white_outfit, …; keep_tokens=2, repeats=3.
  • LoRA dim 32 / alpha 16; AdamW8bit; TE + UNet LoRA; bf16, 1024 buckets.
  • 7 epochs / 1113 steps.

Related repositories

Compatibility and Limits

Best on waiIllustriousSDXL / Illustrious family. Single outfit token.

License

CreativeML OpenRAIL-M for weights. Character © original rights holders. Non-commercial fan use.

Downloads last month
101
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Dataset used to train abbuibuibui/MimiPearlbaton-Lora