Then, choose “Download cuDNN”, and you’ll be asked to login or create an NVIDIA account. Installing cuDNNĪfter installing CUDA 10.1, you can now install cuDNN 7.6.5 by downloading it from this link. In Ubuntu 18.04, as you might know, CUDA is installed in /usr/local/cuda or in /usr/local/cuda-10.1. usr/lib/cuda - which you can verify by running, $ whereis cuda cuda: /usr/lib/cuda /usr/include/cuda.h Unlike in Ubuntu 18.04 (where I was from), CUDA is installed in a different path in 20.04, i.e. Then you will get an output similar to the following to verify if you had a successful installation, nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2019 NVIDIA Corporation Built on Sun_Jul_28_19:07:16_PDT_2019 Cuda compilation tools, release 10.1, V10.1.243 So, you can install CUDA 10.1 in Ubuntu 20.04 by running, $ sudo apt install nvidia-cuda-toolkitĪfter installing CUDA 10.1, run nvcc -V. The CUDA Toolkit Archive.įor the sake of being verbose, do not try to use 18.10 or 18.04 CUDA 10.1 for Ubuntu 20.04. We are installing CUDA 10.1 because it is the compatible version with TensorFlow GPU. We use the proprietary version over the open source one since CUDA can only operate with the proprietary driver. for CUDA 10.1, the required driver version is ≥ 418.39. Install TensorFlow 2.0 with GPU support.įirst, ensure that you are using the NVIDIA proprietary driver by going to “Additional Drivers”, and then choosing the appropriate driver, i.e.Install the cuDNN version compatible with CUDA 10.1.Install CUDA 10.1 by installing nvidia-cuda-toolkit.
The installation of TensorFlow GPU in Ubuntu 20.04 can be summarized in the following points,
So, this post is actually based on the answers given by meetnick and singrium in this related question posted in Ask Ubuntu. Fortunately, I saw some helpful answers online, and now I’m expanding on their answers by adding a bit more explanations. As I expected, it took me a while to work things out in my new system.
0 kommentar(er)
