This my first DoRA.
Trained with batch size 4 and 8 gradient accumulation steps.
capSleevesDoraVersion_animaginexlV4OptV2.safetensors