Install TensorFlow GPU in Anaconda on Windows 11 | TensorFlow GPU and Cuda Installation

hi everyone welcome to this brand new lecture in this lecture I'm gonna show you how you can install a tensorflow with the Anaconda and how you can use a tensorflow GPU if you have a GPU all right so for that you have to have the Anaconda as a pre-installed software in previous lecture I had shown you how you can install the Anaconda thereafter you need to open here Anaconda prompt in a administrator mode all right so either you can click on here a right click and then run as the administrator all right so you will run here Anaconda into administrator mode so this is the default path for the Anaconda into the administrator mode all right thereafter you need to search here tensorflow all right let me Zoom this a little for you so that you can see it clearly all right so here you need to search here tensorflow install all right or install tensorflow whatever you prefer to search there then you will be getting this particular page here install tensorflow 2. thereafter you will be getting here these pip installation instructions but I would request you to click on read the PIP install guide there you will be redirected to something like this so this one you need to install on your Windows computer so you need to click on Windows wsl2 here all right if you are using Linux you can use this one or Mac OS for this one or Windows wsl2 and then you need to copy first line from here just copy this and there you need to go back into your anaconda administrator Anaconda prompt and then paste that here then you need to run this so you will run this the conda will automatically find out the Q dot toolkit 11.2 and qdnn 8.1 which will be installed on your computer under the environment base here all right all right so once conda identify all the downloadable packages it will ask you whether you want to proceed or not I'm gonna say yes I want to proceed then depending on your internet speed it is going to take some time to download Cuda toolkit and qdnn so you need to wait for some time so that it can finish download and then installation all right after some time you will find out that this installation is done thereafter you need to come here and then you need to copy this whole thing from here all right and then just copy these thereafter you need to paste that here all right so it says that the export LTE Library path LD Library path the conda prefix and then leave you just run this you will see that the export is not recognized because we are using here a Windows machine so you need to sit here you need to replace this export with set so we do here we we set our you know the loaded Library path thereafter you need to install a tensorflow so you need to just copy it from there and then paste it here then we are going to run here python MP install tensorflow do remember the first process where we installed Q dot toolkit and qdnn that is required only when you want to use tensorflow with the GPU if you want to use tensorflow just for the purpose of the CPU processing I mean to say that on a computer which don't have a GPU then you don't need to install a Cuda toolkit and the qdnn all right so while installing tensorflow it will also install other necessary libraries and packages on the Fly once this tensorflow will be installed then we will be copying this whole thing to test whether tensorflow is using GPU or not all right so I'm just gonna copy this whole thing then I'll replace this Python 3 with the python there all right now you see there the tensorflow has been successfully installed currently we are using here a tensorflow 2.10 let's go ahead and run this one uh there I need to replace let me just show you here this Python 3 with the python there alright let's go ahead and just run this all right so the first time it takes some time to initialize the tensorflow and you know the Cuda packages that's why it is taking some time otherwise it could have given you the tensorflow is using GPU all right let's wait for some time all right after some time it has given you a result where it says that the device type is the GPU so if GPU was not being used it would have given you as no GPU found or something like that result since we are already using here uh I mean to say that since I already have GPU in my computer and I also installed the qdot toolkit and the qdnn in a correct procedure so my tensorflow is using here GPU now I can develop the algorithm a deep learning algorithm which will be using a GPU but still we have not installed the pi torch yet but suppose that if you want to install the pi torch and you want to use the GPU with the pi torch which we will be seeing in the coming lectures I'll show you how you can install the pi torch and how you can use the GPU with the pi torch all right this is all about in this lecture thanks for watching I'll see you in next one

Channel: KGP Talkie

Views: 40,386

Keywords: kgp talkie, kgp talkie videos, machine learning tutorials, kgp talkie data science, kgp talkie ml, kgp talkie for machine learning, install tensorflow gpu, install tensorflow gpu windows 10, install tensorflow, install tensorflow in anaconda, install tensorflow in jupyter notebook, tensorflow gpu in windows, tensorflow gpu windows 11, tensorflow gpu windows, install tensorflow gpu and cuda, install cuda on windows machine, install cudatoolkit, install cudnn, cudatoolkit, cudnn

Id: 7KN9q31OXBs

Length: 6min 21sec (381 seconds)

Published: Sun Oct 09 2022