I installed a graphics card into my machine, and the necessary nvidia drivers were installed when I installed TensorFlow-GPU through conda. I’ve been training networks on the GPU and it’s worked fine. After rebooting for the first time, the GPU is not detected. In Settings → Details → About, Graphics is listed as llvmpipe (LLVM 10.0.0, 256 bits), not my GTX Titan X that was listed before the reboot. I’m also unable to modify the screen resolution. I’ve attached a bug report here:
nvidia-bug-report.log.gz (93.9 KB)