Open YanisLalou opened 5 months ago
It sounds like a functionality of DomainAwareDataset
, right?
Not totally, if you have access to X,y,sample_domain
but want to plot individual domains (or estimate their) or the reverse sucjh functions make sens but now that I see it probably it is better to have domain_merge
and domain_split
function that can also return/take a list of domain numbers corresponding to the arrays in the lists. For the moment we have no multidomain functiosn but it will be necessary to have such helpers inside the classes also (going back to DomainAwareDataset and doing a list of pack is not friendly for this )
Add the ability to have:
source_target_merge
from utils.py to accept*List[arrays]
with each list corresponding to one sample domain. Ex: List[X_domain_1, y_domain_1], List[X_domain_2, y_domain_2] ...source_target_split
from utils.py to return*List[arrays]
with each list corresponding to one sample domain.