Sakana’s Latest AI Models Are Recreating Historic Japanese Ukiyo-e Art

The AI startup Sakana AI has released two new image generation models focused on recreating Japan’s historic ukiyo-e art style, aiming to spread appreciation for Japanese culture and find applications in education and literature.

Bringing ukiyo-e art to the modern era with AI: Sakana AI’s new models, Evo-Ukiyoe and Evo-Nishikie, generate images closely resembling the Japanese ukiyo-e art style that flourished between the 17th and 19th centuries:

• Evo-Ukiyoe is a text-to-image model that generates ukiyo-e style images based on text prompts, even allowing for the inclusion of modern elements like hamburgers or laptops, though results may sometimes deviate from the traditional style.
• Evo-Nishikie is an image-to-image model that colorizes monochrome ukiyo-e prints, adding color to historical book illustrations or giving new looks to existing multi-colored prints based on user-provided source images and instructions.

Localization efforts in the AI space: The release of these models is part of a growing trend of AI localization, with companies in countries like South Korea, India, and China developing models tailored to their respective cultures and dialects.

Building on Sakana’s novel AI techniques: The models leverage Sakana’s evolutionary model merging technique and build upon the company’s previous work:

• Evo-Ukiyoe is based on Evo-SDXL-JP, which Sakana developed using its evolutionary model merging technique on top of Stability AI’s SDXL and other open diffusion models, fine-tuned on a dataset of over 24,000 captioned ukiyo-e artworks.
• Evo-Nishikie was created by performing ControlNet training on Evo-Ukiyoe using fixed prompts and condition images.

Research and development goals: While the models are still in the early stages and only support Japanese prompts, Sakana hopes this work will spread appreciation for Japanese culture worldwide and find applications in education and new ways of enjoying classical literature.

Analyzing deeper: Sakana AI’s release of these models showcases the potential for AI to not only generate new content but also to revive and reinterpret historical art forms for modern audiences. As AI continues to advance, we may see more efforts to use the technology to preserve and promote cultural heritage. However, it remains to be seen how well these AI-generated artworks capture the essence and nuances of the original ukiyo-e style and whether they will truly resonate with audiences in the same way as the traditional art form.