So I'm trying to set up a GPU profiler on tensorboard but I am getting this error:
W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcupti.so.11.2'; dlerror: libcupti.so.11.2: cannot open shared object file: No such file or directory
This makes sense because the libcupti version I have is 11.6, so it's trying to get 11.2 and I don't have it. How can I specify to use libcupti.so.11.6
?
Machine specifications:
TensorFlow 2.8 doesn't support CUDA 11.6. but requires 11.2 see docs
Seems you need to get in touch with the VM's owner to update the dependencies