ltshan

change picture as log: bc311@bc311-ai1:/work/xxx/BiSeNet$ python3 tools/train.py --model bisenetv2 loss_pre= tensor(1.2688, device='cuda:0', grad_fn=) loss_aux= [tensor(9.9736, device='cuda:0', grad_fn=), tensor(3.0100, device='cuda:0', grad_fn=), tensor(5.4277, device='cuda:0', grad_fn=), tensor(3.6156, device='cuda:0', grad_fn=)] sum= tensor(22.0268, device='cuda:0', grad_fn=)...

> i have the same problems after first epoch. i found i do not change the val path in config file. make sure your Bisenetv1.py file has been changed correctly....

> Hi, > > are you using your own dataset or dataset of cityscapes ? It's my own dataset

there are 3 classes, including background for my dataset. it's pascal voc format. how to set class number in config file? and by my check, self.lb_map is NOT none, how...

only one picuture should be ok

It should be B * f / disparity