LoRA of Laura B. (For Illustrious)

Description

This LoRA trained using kohya_ss on 60 images 1024×1024 up to 10800 steps without captions and with regularization images.
Source model: ‘IllustriousXL_v01.safetensors’, so this LoRA should works somehow with most of Illustrious family checkpoints. In my tests it performs best with semi-realistic Illustrious checkpoints like ilustmix_v9.safetensors and PerfectDeliberateXL.safetensors.
Learning rate: 0.00003.
Precision: fp16.
LR Scheduler: constant_with_warmup.
LR warmup: 30%.

Square preview image generation data:
Positive prompt: 1girl, girl, AIDA_ILLR_LauraB2, lips, bangs, ponytails, pigtails, pink ruffled blouse, black background, black_background, upper body, female child, 7yo
Negative prompt: worst quality, low quality, censored, furry, 3d, anime, extra legs, extra hands, mutated legs, mutated hands, closed eyes, adult, closeup, close up, close_up
Steps: 36,
Sampler: Euler a,
Schedule type: Automatic,
CFG scale: 3.5,
Seed: 2956290413,
Size: 1024×1024,
Model hash: 0826c8c748, Model: ilustmix_v9.safetensors,
VAE hash: 235745af8d, VAE: sdxl-vae-fp16-fix.safetensors,
Clip skip: 2,
Hires upscale: 1.5,
Hires steps: 18,
Hires upscaler: 4x_NMKD-Siax_200k.pth,
Lora value: 1, Lora name and hash: “AIDA_ILLR_LauraB2: 3bfe1bad2f94”.

Download link to AIDA_ILLR_LauraB2.safetensors will be available for you after purchase. For your comfort I packed AIDA_ILLR_LauraB2.safetensors into .zip archive with AIDA_ILLR_LauraB2.png and AIDA_ILLR_LauraB2.txt. Preview and description are compatible with AUTOMATIC1111 and Forge extra networks tabs.

Purchasing this asset you will achieve rights for selling images generated with AIDA_ILLR_LauraB2.safetensors LoRA. If you want say special thanks to me, you can add additional amount to your payment.

Please contact me if you need some extended rights for sharing or selling AIDA_ILLR_LauraB2.safetensors, merging AIDA_ILLR_LauraB2.safetensors with other models, using AIDA_ILLR_LauraB2.safetensors in generation services and so on.