Data- and change specialist
A list of free to use AI models that I made. When using, adapting or publishing my work you only have to credit me.
The models are based of datasets I collect myself. This can range from historic images to pictures I made myself. The data and models are publicly available.
If you want to request a model or want to publish or collaborate on a dataset/model, contact me to discuss the possibilities.
A LORA model based of post cards from the 1900-1980. Collected on Wikimedia via manual collection.
Color range: both monochrome and color.
Triggers: WYWH, monochrome, caves, mountains, animals, buildings
Dataset size: 656 (1.6 Gb)
A LORA specialized in simulating damage on photography and paintings.The dataset mainly consists out of paintings so it has a significant bias towards art-styles from that era.
Color range: both monochrome and color.
Triggers: katadb, painting, damage, yellowed, mold, torn, pierced, cut, tear, cut-out, framing.
Dataset size: 320 (1.6 Gb)
A dataset of post cards from the 1900-1980. Collected on Wikimedia via manual collection.
Color range: both monochrome and color.
Dataset size: 656 (1.6 Gb)
A specialized dataset for simulating damage to paper and canvas. The goal is to train a LORA model to simulate damage to artworks and paper with a focus on vintage images.
Color range: both monochrome and color
Dataset size: 320 (1.2 Gb)
Quick question, what do you imagine when you think about photographs of antarctic? Penguins, seals? How about several military outposts? This model focuses on providing a dataset to build a LORA to, generate unique imagery on the harsh reality of Antarctic.
Color range: both monochrome and color
Dataset size: TBD