Use:

As the name suggests this is meant to be an enhancer of style. You can use it in combination with other styles and artists to generate some interesting results or you can use it on its own but it's not a dedicated ukiyo-e model, only a textual inversion.

Enhancing:

Use it anywhere in the prompt, the further it is to the start the more influence it will have. Best used with other styles and artists.

On its own:

Start with "ukiyoestyle-1500 masterpiece"

Look at the sample images prompts to get an idea. I also included comparison of the two different versions with otherwise same settings.

Training:

This Textual Inversion was trained on 32 high quality ukiyo-e woodblock prints from Hokusai and Hiroshige.

The training images contained as many different elements as possible like people, landscape, animals, ships, buildings to make this embedding versatile.

Versions:

There are two versions included:

ukiyoestyle-1500 (regular) and ukiyoestyle-3000 (strong).

The "strong" version was trained more and has an interesing side effect:

Instead of making the style stronger it will make the images slightly more photorealistic if no specific style is prompted. Also it will occasionally bleed the trainings images content instead of applying the style. Which means it will tend to force asian faces and kimono. For example you will get a photorealistic woman wearing kimono instead of a woman wearing modern clothes in the style of ukiyo-e. See examples.

Suggestions:

Works well in combination with Ohara Koson Style Lora and specifically prompting ukiyo-e style.