GoGoDuck912/Self-Correction-Human-Parsing

Always get black image output (all zero label parsing)

hyeokhyen opened this issue · 4 comments

hyeokhyen commented

Thank you for the sharing code and the model.

I was running the evaluation on LIP with the shared pertaining dataset.
However, for any image, I get the black result output. (even for the images from the training set)

Could you help me with it?
I will appreciate your reply.

pasta99 commented

Have you been able to fix this problem? I'm experiencing the same problem.

Karenou commented

Have you solved this problem? I also encounter save issue.

DATOULEE commented

使用CPU推理就可正常

Karenou commented