This my first DoRA.
Trained with batch size 2 and 16 gradient accumulation steps
capSleevesDoraVersion_naiEpsV3.safetensors