So in our case because we are installing CUDA 9.0 the command will looks like this: sudo apt-key add /var/cuda-repo-9-0-local/7fa2af80.pub Sudo apt-key add /var/cuda-repo-/7fa2af80.pubĭon’t forget to replace - in the second command with your installed version.
deb file execute the following commands (As you also see on the website): sudo dpkg -i cuda-repo-ubuntu-local_9.0.176-1_b deb local installation method.Ĭhoose the architecture that suits your operating System, in my case I use x86_64, Ubuntu, 16.04 (they don’t have a version yet for 18.04, but this works fine with 18.04) and. Then go to download page of CUDA Toolkit 9.0 here Then execute the following command in your terminal: sudo apt install nvidia-390 Installing CUDA-9.0 Add the repository firstly sudo add-apt-repository ppa:graphics-drivers/ppa In this blog post I use nvidia-390 driver. In addition, you have to install (almost) the latest nVidia driver. The first step to be able to use Cuda and cuDNN is having a nVidia graphic card. Tensorflow for example, took 10 to 15 seconds to perform recognition tasks when running on cpu, while it took 2 to 5 seconds for the same recognition tasks when running on a GPU with Cuda installed. Based on my little experience in this field. Hi everyone, Cuda and cuDNN are must-have tools for everyone who wants to start with Computer Vision, Deep Learning, Machine Learning using GPU (which is way much faster than using the CPU even if it’s core i7).
0 kommentar(er)
