search

MenghaoGuo / AutoDeeplab

Pytorch Implementation the paper Auto-DeepLab Hierarchical Neural Architecture Search for Semantic Image Segmentation
https://arxiv.org/abs/1901.02985
410 stars 97 forks source link

what is train_loader1 and train_loader2? #12

Open lqycrystal opened 5 years ago

lqycrystal commented 5 years ago

self.train_loader1, self.train_loader2, self.val_loader, self.test_loader, self.nclass = make_data_loader(args, **kwargs) ValueError: not enough values to unpack (expected 5, got 4)

NoamRosenberg commented 5 years ago

I get this error too

dagongji10 commented 5 years ago

self.train_loader1, self.train_loader2, self.val_loader, self.test_loader, self.nclass = make_data_loader(args, **kwargs) ValueError: not enough values to unpack (expected 5, got 4)

you need to modify dataloaders/init.py, the function make_data_loader() should return train_loader twice.

NoamRosenberg commented 5 years ago

I just change the output to output a single train_loader, @dagongji10 do you see any reason we need two?

dagongji10 commented 5 years ago

@NoamRosenberg I guess train_loader1 is for image and train_loader2 is for image

I have not finish the code-reading, so I'm not sure.

NoamRosenberg commented 5 years ago

@dagongji10 which train do you use for calculate_weigths_labels in the init function in train_autodeeplab.py ??

Randylcy commented 5 years ago

trainA and trainB is not same in the original paper, while I think the code make trainA and trainB same.