faster_rcnn_pytorch icon indicating copy to clipboard operation
faster_rcnn_pytorch copied to clipboard

debug so lowlly

Open tengshaofeng opened this issue 8 years ago • 2 comments

hi, @longcw , i am so appreciated with your great job. and i can run train.py and demo.py correctly. but when i remote debug the code, at first it is fast, but when it comes to the code in train.py as follows:

if not os.path.exists(output_dir):
    os.makedirs(output_dir)

it is so slowly for steping over, after that. what is the problem? can you help me? thanks.

tengshaofeng avatar Dec 04 '17 10:12 tengshaofeng

os.makedirs() shouldn't be slow. Maybe you can check if the output_dir is a reasonable path in your computer.

longcw avatar Dec 04 '17 11:12 longcw

@longcw , thank you for your reply. not only that code slow, the following is also slow, for example 'train_loss = 0' also take few minutes.

tengshaofeng avatar Dec 05 '17 01:12 tengshaofeng