I’m aware that when TensorFlow is installed, you choose between the CPU and GPU variants, and I am currently working on Linux. I need to determine which version has been set up on my system. Additionally, if the GPU variant has been installed but a GPU isn’t available, will the system automatically switch to the CPU or will it generate an error? Moreover, if a GPU is present, is there a particular configuration or parameter that must be adjusted to ensure it is actively used?
hey, ive been using tf.config.list_physical_devices(‘gpu’) too. its cool when it falls back to cpu if no gpu’s available, but im curious… did u notice any issues with env vars or drivers? how did u check your versions?