Portrait of a blonde woman in a black lace turtleneck dress with detailed eyes and dark red lipstick, set in a dark atmospheric background with artistic splatters.
Close-up portrait of a girl with long light green hair, blue petals on her cheeks, and realistic skin texture, surrounded by dripping watercolor splashes in green and blue tones.

Recommended Prompts

high quality, beautiful, highly detailed, 8k

Recommended Negative Prompts

fastnegative2, clutter, blurry, extremely blurry, elongated, stretched

Recommended Parameters

samplers

DPM++ 2M Karras, Euler a

steps

30 - 40

cfg

7 - 7.5

resolution

888x1152, 832x1152, 892x1152

vae

SDXL-VAE_0.9.safetensors (63aeecb90f)

other models

FaeTasticSDXL1.6PRUNED (7762ec03a3), crystalClearXL_ccxl (0b76532e03)

Recommended Hires (High-Resolution) Parameters

denoising strength

0.25 - 0.3

Tips

The model works beautifully at max strength, but users are encouraged to experiment with different weights to find what suits them best.

The LoRa was trained on SDXL FaeTastic 1.6; compatibility with other SDXL checkpoints is not guaranteed but has shown success in tests.

For training SDXL LoRAs, following the recommended guidelines linked in the description is advised.

Users are asked not to upload the creator's resources on third-party websites.

For questions regarding Civitai, Stable Diffusion, or the creator's resources, users can message the creator on Discord via the official Civitai server.

I've been meticulously refining this LoRa since the inception of my initial SDXL FaeTastic version. Drawing inspiration from two of my cherished creations, x and x I've trained to craft something capable of generating exquisite, vibrant fantasy letter/manuscript pages adorned with exaggerated ink stains, alongside whimsical desk accessories like pens and ink blotters.

It works beautifully at max strength and it's what I used to test this. But check out other weights to see what suits you best. :)

This LoRA was trained on SDXL FaeTastic 1.6, I can't guarantee best use with other SDXL checkpoints...although I did test on others and it seemed to work well. :)

I have been having great success training SDXL LoRAs following these guidelines

Please don't upload my resources on third-party websites, if I wanted them there I would upload them myself.

Feel free to always message me on Discord (you can find me in the official Civitai discord) for any questions regarding Civitai, Stable Diffusion, or any of my resources!

Previous
ta86artstyle - EXP_INK
Next
DonM - Machine World Tech [SDXL] - v1.0

Model Details

Model type

LORA

Base model

SDXL 1.0

Model version

v1.0

Model hash

0b66f17a3d

Trained words

Inkstains

Creator

Discussion

Please log in to leave a comment.

Images by SDXL Ink Stains - v1.0

fantasy Images

ink Images

paper Images

parchment Images

style Images