Diffusion64 (SDXL): v2.0: diffusion64-v2-merge.safetensors

Slightly larger synthetic data augmentation set, more training steps. Merge of a 20k step low constant lr Adam run and a variable lr 11.9k step Prodigy run.

A Stable Diffusion XL LoRA that generates Ocarina of Time / Majora's Mask style images. Training set is 1,500+ hand-tagged screenshots.

Uses sd_xl_base_1.0_0.9vae as base model.