ProGamerGov's picture
Improve ReadMe description formatting
d4f9725
|
raw
history blame
1.62 kB
metadata
license: creativeml-openrail-m

Min-Illust-Background-Diffusion

This fine-tuned Stable Diffusion v1.5 model was trained for 2250 iterations with a batch size of 4, on a selection of artistic works by Sin jong hun. Training was performed using ShivamShrirao/diffusers with full precision, prior-preservation loss, the train-text-encoder feature, and the new 1.5 MSE VAE from Stability AI. A total of 4120 regularization / class images were used from here. Regularization images were generated using the prompt "artwork style", 50 DDIM steps, and a CFG of 7.

Use the tokens sjh style in your prompts for the effect. Note that the effect also appears to occur at a much weaker strength on prompts that steer the output towards specific artistic styles.

Example images were generated using 50 steps of DPM++ 2M Karras using:

<prompt>, sjh style