Closed fusionlove closed 6 years ago
If anyone else is having this issue, the code appears to work OK without cudnn, if you disable it by modifying the theano flags.
Hi @fusionlove ,
The version of CUDA is 7.5 and CuDNN is v5.0. Best,
Junting
Thanks for your speedy response! F
On 21 February 2017 at 15:03, junting notifications@github.com wrote:
Hi @fusionlove https://github.com/fusionlove ,
The version of CUDA is 7.5 and CuDNN is v5.0. Best,
Junting
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/imatge-upc/saliency-salgan-2017/issues/10#issuecomment-281369418, or mute the thread https://github.com/notifications/unsubscribe-auth/AAK_zyOcNlb17XYOBzKasCi7Sci6oMMtks5revzFgaJpZM4MHaDk .
Hi,@fusionlove, I have the same problem in the process of training, can you tell me how to solve this problem? Thanks a lot.
Best, Wei Shan
Hello, thank you for making your code available. I'm having some issues with cudnn (not able to obtain a cudnn handle when running the predict script).
Could you provide some more details about which version of CUDA and cudnn you are using? Thanks Fintan Attention lab, UCL