Scott Brown
Scott Brown
@aduriseti I did not implement this, no. My use case was.... unusual.... so I got the impression that they didn't want such a feature here. I ended up just making...
Any support for this?
I am having the same problem, except I cannot detect the dog for any combination of weights and architectures. CUDA 8.0 K80 CUDNN 7.1 Nvidia drivers 375.66 GPU=1, CUDNN=1
Upgraded my nvidia drivers and cuda. No better. CUDA 9.1 K80 CUDNN 7.1 Nvidia drivers 390.67 GPU=1, CUDNN=1
Lowered the `-thresh` option to 0.01, and it detects everything, everywhere at a confidence of 5%. This is weird, but I suppose this is because 0%s cause issues with cross-entropy...
Turned off CUDNN and it seems to work. What version of CUDNN is everyone using?
@tomMoral you say "No python error". Are you saying that this problem is not caused by a python error in the worker? I am getting this error and am not...
Oh, yes the quick reference looks very handy. Thanks for the reference.
@danwild still getting failures on 10.22.0 and 6.14.6
As an aside, if this seems like a bad plan or if there is a better plan, I would be grateful for literally any advice. It's not immediately clear to...