I have a notebook with NVIDIA GeForce MX150 display adapter, and tried to install CUDA 9.0 or 8.0 in order to run TensorFlow GPU version. Always failed, something incompatible. So I wonder: 1. which version of CUDA I should use for my machine. 2. I heard that TensorFlow 1.7 can support CUDA 8.0 upmost, if that is true, which version of TensorFlow I should download for this notebook. Thanks!
You can use CUDA 9.0 with that GPU. Most recent TF binaries expect CUDA 9.0
If you’re on a dell pc then go to the dell website and download the Nvidia dedicated dell drivers from their website.
Then install the CUDA 11.0 from the nvidia site
then go for the link: python - Why is Tensorflow not recognizing my GPU after conda install? - Stack Overflow
Try the codes suggested by all the users example geometrical Nicolas and one would surely work for you.
For me geometrical’s code worked but for my friend with the same laptop it didn’t work…