question about calculating albedo map

[Torment123]

Hi, I'm new to intrinsic image decomposition and have some confusions of the calculation of albedo maps. and I'd be appreciated if you could help me:

1. Based on the code in nerf_network.py and line 147 of ddp_model.py, as I understand it, the albedo map is calculated like the color computation of vanilla nerf, i.e., get the view dependent rgb color from each sampled points along the ray from MLPNet, and then weighted combine with the alpha values. This albedo map is later modulated by shadow map and irradiance (line 208, 209) to form the final rendered color
2. But according to the definition of albedo, it is the overall reflectance rate of a surface point, which is not view dependent, so why is it can be calculated through the view-dependent rgb colors from 1?

Sorry if my description is not clear enough