Xingang Pan

@chenyiman Ok, I know the problem. You should rename the 'driver_23_30frame_part1' to 'driver_23_30frame'. Since BaiduYun do not accept file larger than 20G. I have to split the 'driver_23_30_frame' into two...

@chenyiman Try this. 

@chenyiman My mistake, we should modify 'laneTest.lua' 

@chenyiman This is odd. Your image path has an extra '/data/CULane'. Try to remove it either in data list test.txt or in script test.sh. In my case the path is...

@chenyiman Also, if you have changed your data list, you should remove the files in folder 'gen' before run experiments.

It is already provided at https://drive.google.com/open?id=1Wv3r3dCYNBwJdKl_WPEfrEOt-XGaROKu

Please check https://github.com/XingangPan/SCNN/blob/master/experiments/models/SCNN-gen.lua or https://github.com/XingangPan/SCNN-prototxt-generator

1. Both are the trained model. After each epoch, the code would do preliminary evaluation on the validation set, and model_best.t7 is the one with the highest validation performance, while...

@cuicuizhang1989 @ytzhao I have provided a script for generating SCNN caffe prototxt file: https://github.com/XingangPan/SCNN-prototxt-generator. You can use http://ethereon.github.io/netscope/#/editor to visualize the architecture. Hope it would be helpful.

@FrancisGee The lane detection model has two branches. The problem could be viewed as a multi-task problem. The last convolution layer of the model would output feature maps of 5...