Hello,
I understood that the CUDA & cuDNN framework seems to show incompatibility effects when used together with Tensorflow on Windows.
Others have already created elaborate compatibility lists, respectively:
Can anybody here maybe help me to confirm if my below setup is known to be working:
- CUDA 10.2
- cuDNN 7.6.5
- tensorflow-gpu-2.2.0rc3
- Windows 10
- NVidia GeForce RTX 2070
The reason I am asking is that I still I get the below error message in my current environment when running a Python script that would run OK in previous environments:
`Failed to load the native TensorFlow runtime.`
Thanks in advance and with best regards,
Robert