Tensorflow error on GeForce 1030 ?


I’m setting up CUDA programming environment on Ubuntu 18.04.
But after installing CUDA, cuDNN, Tensorflow-GPU, the following error is occurred when running conv2d operation.

tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding=‘SAME’)
“Failed to get convolution algorithm. This is probably because cuDNN failed to initialize”

My CUDA environments are as follows:
GPU: GeForce 1030
CUDA: 10.0
cuDNN: 7.3.1
Tensorflow-GPU: 1.13.1

How can I fix the error message?

Thanks in advance.


Please see this thread for a possible solution: https://devtalk.nvidia.com/default/topic/1043867/failed-to-get-convolution-algorithm-this-is-probably-because-cudnn-failed-to-initialize

Let us know how it goes for you.