Sign In

Ethereal Paint (Style) PonyXL LoRA

81
426
2.6k
59
Updated: Jul 20, 2024
stylepaintpainterlystyles
Verified:
SafeTensor
Type
LoRA
Stats
426
2,552
441
Reviews
Published
Jul 19, 2024
Base Model
Pony
Training
Steps: 6,556
Epochs: 160
Usage Tips
Clip Skip: 2
Strength: 0.9
Hash
AutoV2
91089A4197

Want a LoRA like this one for your favourite character or original design? Check out my Fiverr gig (not currently taking style LoRA commissions but send me a DM if you'd be interested).

This LoRA does not mimic any individual artist style, but is instead trained on a range of concept art/painterly art to achieve a 'painterly' rough style. This is my first attempt at a style LoRA so please let me know what you think.

I really like how this LoRA looks with SDE & Ancestral samplers at CFGs between 8 and 10. It doesn't need any prompt keywords as it's trained to implicitly achieve a style. It plays well with other LoRAs for the most part.

Trained on PonyXL with 169 images over 160 epochs with adafactor optimizer & cosine scheduler on 4x NVIDIA Tesla P40 GPUs.

This LoRA will not work with SD 1.5 or SDXL. You need to use it with PonyXL or a PonyXL-based checkpoint.

Suggested strength: 0.8-1.1

Suggested CFG: 8+

This LoRA does not require any prompting, it's implicitly trained. Sample images also use this detailer LoRA at 0.45 weight & 0.45 clip strength.