A problem in running facialDetails.py

[seven-sent]

Thanks for publishing the project! But some problems occur when I run the code on my computer.

When I try python facialDetails.py -i ./samples/details/019615.jpg -o ./results --batchSize 10 it returns RuntimeError: cuDNN error: CUDNN_STATUS_BAD_PARAM How can I solve this problem？

[apchenstu]

sorry for replying late, did you resolve this issue?

[seven-sent]

Thank you for your reply. The problem appears on

 File "\released_v0.1\DFDN\models\networks.py", line 253, in forward
    netOut = nn.parallel.data_parallel(self.model, input, self.gpu_ids)

I tried to solve the problem through the Internet. Some people said it's because of data type problems, but I don't see any corresponding problems.

I wonder if it's because I only have one GPU. Then I tried again, but still failed. Could you give me some advice?

[apchenstu]

one gpu should be fine, I didn't occur this issue before, could you post your pytorch, cuda version and GPU types?

[seven-sent]

Thank you for responding so quickly! My running environment: torch==1.2.0 CUDA==10.0 GPU==GTX 1650Ti

[apchenstu]

Thank you for responding so quickly! My running environment: torch==1.2.0 CUDA==10.0 GPU==GTX 1650Ti

so weird and sorry I couldn't figure out the problem.