Personal lighting preferenceVersion 2 should perform better in darker areas than Version 1 due to being trained on more low-light images. It also features higher contrast.
lighting.safetensors