Open ssh10032 opened 2 years ago
Hi,
Thanks for your interest in our work.
I hypothesize that changing min_val
and max_val
on pretrained model would just change units the output depth map uses. By default, the units are meters but if you change the max_val
to new_max_val
, then output units would be scaled by (new_max_val - new_min_val) / (max_val - min_val)
.
I'm curious if you experimented with this and your concluding insights about this situation.
Hi, I'm going to train Adabins to my custom dataset. At first, there is a part where the maximum and minimum values are specified in the part where the model is loaded, does this have an effect on training?
model = UnetAdaptiveBins.build(n_bins=256, min_val=0, max_val=2.5, norm='linear').to(device)
And is it a problem if I use these values differently when I load the weights later?Thank you for providing your work. It is very helpful for studying.