| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. To something like Wed Jun 6 14:53:35 2018 Make sure that the latest NVIDIA driver is installed and running. The output of the nvidia-smi command changes from NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. You can also try running watch -n 1 nvidia-smi on a different terminal and see what happens when you run the above command.
#NVIDIA GEFORCE GT 930M TENSORFLOW DRIVERS#
I tried several times to uninstall and reinstall the NVIDIA drivers and Tensorflow differently onlly to realise in the end that they had both been working fine all the time and all I really needed to do was to prefix the optirun command as follows: Otherwise your GPU is switched off thanks to Bumblebee to save battery power. What took me a while to realise what that in order to use the GPU on my laptoip with a Python script (which calls Tensorflow internally), I had to prefix it with the optirun command.
#NVIDIA GEFORCE GT 930M TENSORFLOW INSTALL#
Then I installed Tensorflow GPU from Pip ( sudo pip install tensorflow_gpu). Since I'm using a laptop, I also installed the Bumblebee package. All I did was follow the right set of instructions to install the NVIDIA drivers from the official repositories for my hardware. I use Debian 9 with GPU enabled Tensorflow.
0 kommentar(er)
