Description
Initial.
The LoRA was trained on 100 images from 2025 and 2026.
The dataset consists of:
year 2025: 64 imagesyear 2026: 36 images
Only one image per variant set was used during training to encourage greater compositional diversity.
The dataset includes both cen/decen images, with a greater proportion of decen images.
dim32/alpha32, CAME, UNet LR 3e-05, cosine 900 steps, 100 images (1024²)