Sign In

Ilya Kuvshinov LoRA

699

3.1k

12

Verified:

SafeTensor

Type

LoRA

Stats

3,142

0

Reviews

Published

Mar 12, 2023

Base Model

SD 1.5

Training

Epochs: 4

Trigger Words

kuvshinov

Hash

AutoV2
45A829C8A0

This is my first LoRA that's focused more towards artistic generation rather than the more 'tasteless' models on my profile. Given that I expect a lot of people to gravitate towards this as Ilya Kuvshinov is a very good illustrator, I find it necessary to give generation details which will be further down below.

For the generations seen, I used a strength of 1, but this isn't necessary. As usual for my LoRA's, a strength between 0.6 and 0.8 is good.

This LoRA was trained on the NovelAI model with 300 images, 4 epochs, 2 batches, and 10 restarts. I can't remember the amount of training steps specifically, but it numbered 6000~.

All images generated were on Anything V4.5, AOM3, with Orangemix VAE and Hires Fix..