I am currently trying to install tensorflow-gpu to speed up the training of my CNN but after different tries (compatibilities/versions CUDA/Cudnn/tensorflow) I could not manage to successfully run the epochs. I get the error:
The errormessage i get is: UnknownError: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above. [Op:Conv2D]
Now, I’ve noticed that my gpu is not mentioned in the CUDA-enabled gpu:
NVIDIA GeForce GTX 1650 GPU 0: Intel® UHG Graphics 630 CPU: Intel® Core™ i7-9750H CPU @ 2.60GHz
So, I would like to know if the error I have is caused by this and if there is a way around so that I can still use my current gpu for tensorflow.