-1

When I run tf.test.is_gpu_availabe() it returns false and the only problem that it is showing that cudnn64_7.dll not found

I am using python 3.8.17 for tensorflow 2.3.0 Cuda-10.1, CudNN-8.0.5 GPU-Nvidia Geforce MX130, cuda compute compatibility-5

enter image description here

talonmies
  • 70,661
  • 34
  • 192
  • 269
Yeasin Al Yeasa
  • 3
  • 3
  • cudnn64_7 is from CuDNN 7. You have CuDNN 8. – Elmir Jul 21 '23 at 09:45
  • I believe you can find the answer here - [Could not find 'cudnn64_7.dll' - Stack Overflow](https://stackoverflow.com/questions/48698536/tensorflow-gpu-import-tensorflow-importerror-could-not-find-cudnn64-7-dll) – AsafH Jul 21 '23 at 09:46

1 Answers1

0

According to this table (scroll down to GPU, it's for Linux but should be the same for Windows) TF 2.3 works with CUDA 10.1 and CuDNN version 7.6. As @Elmir pointed out the error indicates that TF searches for cudnn libraries from version 7. To fix it, replace your CuDNN files with the right version.

mhenning
  • 426
  • 4
  • 8