Another deep learning network to work on that deals with multi-label classification problem is CNN+RNN net

check this paper CNN-RNN: A Unified Framework for Multi-label Image Classification

We can use basic CNN model like VGG16/19 or Res101/152 and change the loss function a bit to suit our project purpose.

ResNet is harder and slower to train than VGG16 tho..

Maybe we have talked about this before but I forgot. Should we use human label or machine label, or both of them for training? Human labels have more unique labels than machine label (18155 vs. 7509). But machine labels have less confidence (say, confidence might be 0.7 instead of 1 in human label.) Moreover, more labels in human label file are not trainable. (there are only 7178 trainable label). But, in total there are fewer human labels. (8036466 vs 15259186), so from the perspective of efficiency and convenience, we should probably start with human label and get its subset of "trainable" ones?

Any better way than random sampling?