search

TuSimple / TuSimple-DUC

Understanding Convolution for Semantic Segmentation
https://arxiv.org/abs/1702.08502
Apache License 2.0
605 stars 118 forks source link

how to set the val.lst #15

Closed ZzzjzzZ closed 6 years ago

ZzzjzzZ commented 6 years ago

Hi,_ when i use predict_full_image.py , I make my val.lst like this:

111 /home/xxx/MXnetProjects/TuSimple-DUC/Data/cityscapes/gtFine/test/berlin/berlin_000000_000019leftImg8bit.png` But these is always a error, i debug it and found that this is a error when use im = cv.imread(img_path)[:, :, ::-1] in predict_single, for i set the vla.lst in wrong format. can you help me. THANKS!

GrassSunFlower commented 6 years ago

The list can be something like index \t relative_img_path \t relative_label_path And the full paths of images of ground truths are concatenated by 'test_img_dir' and 'gt_dir' in the config file respectively.

ZzzjzzZ commented 6 years ago

thanks! it works! and your head portrait of "liu xing" is interesting

shipengai commented 6 years ago

Hi，@CarryJzzZ。 Can you share me your code for making val.lst? Thanks. A baiduyun link.

shipengai commented 6 years ago

I have solve it. Thanks。