LAI calculation with trained neural network weights

Hi there,

I am new to SNAP. Currently, I want to convert the JS custom script of LAI calculation provided by SNAP to Python script. I was wondering whether it is applicable to completely copy the trained weights and biases for each band which were obtained from a neural network. Are those weights and biases common to whatever Sentinel-2 image (S2A or S2B) I intend to use?

Many thanks,
Alex