voldemortX / pytorch-auto-drive

PytorchAutoDrive: Segmentation models (ERFNet, ENet, DeepLab, FCN...) and Lane detection models (SCNN, RESA, LSTR, LaneATT, BézierLaneNet...) based on PyTorch with fast training, visualization, benchmarking & deployment help
BSD 3-Clause "New" or "Revised" License
840 stars 138 forks source link

cannot identify image file 'culane/test/00004.png.lines.txt' #127

Closed Mind-s closed 1 year ago

Mind-s commented 1 year ago

Thanks for posting your excellent code! When I wanted to predict directly on a few of my own pictures, I meet the issue: PIL.UnidentifiedImageError: cannot identify image file ' culane/test/00004.png.lines.txt'.

Howerver, there is no '*.lines.txt'files in my folder.I don't understand why this is happening.My source code is:

_CUDA_VISIBLE_DEVICES=1 python ./tools/vis/lane_img_dir.py --image-path=/culane/test --image-suffix=.png --save-path=/culane/test_result --pred --config=configs/lane_detection/baseline/resnet101_culane.py --checkpoint=/configs/lane_detection/baseline/resnet101_baseline_culane20210312.pt

I would appreciate it if you can help me solve this problem

voldemortX commented 1 year ago

@Transcients Let me try reproduce this. Can you provide your culane/test folder's filenames?

Mind-s commented 1 year ago

OHHHHHHHH, it looks like I generated these '*.line.txt' files while running other programs. I'll delete them and try it again. Thanks!

voldemortX commented 1 year ago

OHHHHHHHH, it looks like I generated these '*.line.txt' files while running other programs. I'll delete them and try it again. Thanks!

hhhhh