I've decided to lean into being the single image dataset LoRA guy and make some fun single image LoRAs. Training data is included.
This LoRA will make something really similar to the original image, but this is intentional. Use lower strengths (0.6 and below) to break out of the original and make something inspired by it, or higher strengths to stick close to the original (0.8 and above).
The LoRA was regularized and will respond to the regularization dataset captions as well, most notably "galena" but also (much more weakly) "oatd". Repeating these style words will make their respective effects stronger. You can use higher strengths with these alternate styles.
Long form captioning: woman, flowers, holding flowers, dress, sweater, blonde hair, hair bun, blue eyes, butterflies, hanging lanterns, night, smile, garden, looking at viewer
Medium form captioning: woman, flowers, dress, butterflies, hanging lanterns, night, garden
This one doesn't have a specific trigger word, it's basically the entire medium form captioning, or any portion of it you want to keep.
Put any details you don't want from the original into the negative prompt.
Description
Initial Version, Resized from a rank 128 LoRA using sv_fro 0.9, max rank 128, bringing it down from nearly 1GB to just 67.9 MB
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.