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
A few days ago my system crashed, reinstalled everything and got the code running again. However, my previous installation printed (on running main.lua) that it found the cudnn file "libcudnn.so.5" and started training really fast - with great results.
In my new installation this disappeared and training is much slower
th -lcudnn -e "print(cudnn.version)"
gives me 5005
echo $LD_LIBRARY_PATH
gives me /home/user/torch/install/lib:/usr/local/cuda/lib64/:/home/user/torch/install/lib:
my ~./bashrc export PATH=/usr/local/cuda/bin/:$PATH; export LD_LIBRARY_PATH=/usr/local/cuda/$ export CUDNN_ROOT=/usr/local/cuda-8.0/lib64/libcudnn.so.5 . /home/user/torch/install/bin/torch-activate
any idea what's going on? - still in that learning curve - much appreciated!
kubmin.
The text was updated successfully, but these errors were encountered:
Hi there,
A few days ago my system crashed, reinstalled everything and got the code running again. However, my previous installation printed (on running main.lua) that it found the cudnn file "libcudnn.so.5" and started training really fast - with great results.
In my new installation this disappeared and training is much slower
th -lcudnn -e "print(cudnn.version)"
gives me
5005
echo $LD_LIBRARY_PATH
gives me
/home/user/torch/install/lib:/usr/local/cuda/lib64/:/home/user/torch/install/lib:
my ~./bashrc
export PATH=/usr/local/cuda/bin/:$PATH; export LD_LIBRARY_PATH=/usr/local/cuda/$ export CUDNN_ROOT=/usr/local/cuda-8.0/lib64/libcudnn.so.5 . /home/user/torch/install/bin/torch-activate
any idea what's going on? - still in that learning curve - much appreciated!
kubmin.
The text was updated successfully, but these errors were encountered: