Attempting to fetch value instead of handling error Internal: failed to get device attribute 13 for device 0: CUDA_ERROR_UNKNOWN: unknown error
Source of the problem: A problem occurred when Python installed tensorflow.
The cause of the problem: the driver does not match the cuda, just update the nvidia driver.
Solution: Go to the cuda-gpu matching page: http://developer.nvidia.com/cuda-gpus , click on your computer’s gpu model (as shown in Figure 1, such as GeForce), the GPU list will be displayed below it (such as Figure 2)
Figure 1 NVIDIA graphics card type selection
Figure 2 GPU graphics card model selection
Click on the graphics card model of your computer in Figure 2 to enter the detailed page of the graphics card. There is a "Drivers" link under the Support column at the bottom of the page (as shown in Figure 3), click it to enter the next page ( As shown in Figure 4)
Figure 3 Graphics card details page
Figure 4 Graphics card driver download page
Click the "Download" button in Figure 4 to download the GeForce_Experience software, and the problem will be solved after installation.
Very easy to use, if you have any questions, you can leave a comment.
Reference blog: https://www.cnblogs.com/ziyu-trip/p/12663978.html