Trained on around 60 images, seems to output mostly decently.
You will need to select a best image and upscale.
l1lb00tz0.safetensors
Same model published on other platforms. May have additional downloads or version variants.
