train error with my own dataset

[OneMondy]

Hi,XingangPan@XingangPan,I trained with my own labeled data set. I have run the train.sh with the following question. Can you help me see it?

   ... -> output

} } load file(lane-gen.lua): ./adata/CULane/list/train_gt.txt /home/hfut/torch/install/bin/luajit: /home/hfut/Documents/qiao/ScnnNew/datasets/lane-gen.lua:22: attempt to perform arithmetic on local 'impath' (a string value) stack traceback: /home/hfut/Documents/qiao/ScnnNew/datasets/lane-gen.lua:22: in function 'getPaths' /home/hfut/Documents/qiao/ScnnNew/datasets/lane-gen.lua:104: in function 'exec' ./datasets/init.lua:28: in function 'create' ./dataloader.lua:31: in function 'create' main.lua:33: in main chunk [C]: in function 'dofile' ...hfut/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk [C]: at 0x00405d50

[OneMondy]

@XingangPan

[juntawu]

@OneMondy have you modified the nn.View and nn.Linear of the pre-trained VGG model ?