CivArchive
    Preview 1Preview 2Preview 3Preview 4Preview 5Preview 6Preview 7Preview 8Preview 9Preview 10

    MidJourney checkpoints (like OpenJourney) and LoRAs are typically made with many different MidJourney images. But early MidJourney images had a variety of distinctive styles and textures. MadeJourney was trained on a set of MidJourney v1 images, but I'm also including variations that were trained on a subset of MidJourney images that exhibited a specific style or texture.

    Description

    MadeJourney was trained on MidJourney v1 images that exhibit a variety of textures and styles, including all of the subsets used to train the other models. Base results are often good, but the best results come from a second pass with 1.5× or 2× upscaling.