You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, thanks a lot for your work! To make it more accessible, can you make it 100% compatible with Windows? Almost everywhere the paths are dealt correctly with os, but line 81 in create_hdf5.py does a label = int(file_d.split('/')[-2]) that does not work with Windows. If you use os.path.split it will work on any platform. Be careful also with statements like
Hi, thanks a lot for your work! To make it more accessible, can you make it 100% compatible with Windows? Almost everywhere the paths are dealt correctly with os, but line 81 in create_hdf5.py does a
label = int(file_d.split('/')[-2])
that does not work with Windows. If you use os.path.split it will work on any platform. Be careful also with statements likedigit_train = glob.glob(os.path.join(dataset_path, 'Train/{}/*.bin'.format(digit)))
I think it's safer to write it as
digit_train = glob.glob(os.path.join(dataset_path, 'Train', str(digit), '*.bin'))
I could add those changes if you give me permission, since I'm using your dataset for my research, and I'm on Windows.
Have a good day,
Luca
The text was updated successfully, but these errors were encountered: