Examples: Image Trainer

This model is trained on the images of Tipoo's Tiger, a 200+ year old automaton from time of the Southern India ruler Tipu Sultan, which is now housed at the V&A Museum, London. For non-commercial use only.

Who owns it?
Victoria and Albert Museum, London

IMAGE ATTRIBUTION: https://collections.vam.ac.uk/item/O61949/tippoos-tiger-mechanical-organ-unknown/

AI image generation models can ultimately only create images using the styles and objects on which they were trained. This often leaves out digitally underrepresented visual cultures, such as folk art or the styles and characters of lesser-known artists. With this tool, you can upload your own image data and fine-tune an AI image generation model, enabling new styles and creative possibilities for you and those you choose to share your model with. Here, we've added Creative Commons images of Mughal Miniature Paintings to create a model that can output any image scene using the keyword "mugh4!".

To create your own:
Clear this image set
Upload your own images
Describe the contents of the images in words
Tap Run

Note: this process is computationally intensive, costing about $4 each time you hit Run and consuming significant amounts of water and electricity.

IMAGE ATTRIBUTION:
https://www.metmuseum.org/art/collection/search?q=Govardhan&sortBy=Relevance
(Part of the Shahjahan Album in the MET collection)

This image model is trained with Lascaux Cave Painting.
Image Attribution:
https://commons.wikimedia.org/wiki/Category:Lascaux
Image 1 -https://commons.wikimedia.org/wiki/File:Lacsaux_VOA_2021_09.jpg
Image 2 - https://commons.wikimedia.org/wiki/File:Lascaux2.jpg
Image 3 - https://commons.wikimedia.org/wiki/File:Lascaux_painting.jpg
(This image is published under the GNU Free Documentation style. The official position taken by the Wikimedia Foundation is that "faithful reproductions of two-dimensional public domain works of art are public domain.)
Image 4 - https://commons.wikimedia.org/wiki/File:Lascaux,_replica_04.JPG
Image 5 - https://commons.wikimedia.org/wiki/File:Lascaux,_replica_03.JPG

Ever wanted to visualize a character doing (almost) anything? Below are concept images of a single character sheet made in Gooey's Image Generator that we've then used as the input image data set for our image trainer. The result is a fine-tuned LORA model that you can use to then render the character (referred to as DORALORA in your prompt) in any scene.