About training and inference

[q16131130-dev]

After training with a LUT built from my own data, how can the generated .npz file be applied in inference?
Or, can I directly use my own LUT for inference?

Thanks