Which TensorFlow and CUDA/cuDNN version combinations are compatible?

[tini jini]

Hello,

I want to install tensorflow-gpu on windows. I searched the internet. I found these steps:

1- install nvidia driver

2- install cuda

3- install cudnn

4- install tensorflow-gpu

Is it correct?

I have noticed that some newer TensorFlow versions are incompatible with older CUDA and cuDNN versions.

There is a list of tested source configurations at this link.

Should I select the versions from that link and then install them?

Thank you

https://developer.nvidia.com/rdp/cudnn-archive

https://developer.nvidia.com/cuda-toolkit-archive

[Laurence Moroney]

Check out these articles with instructions for Linux or Windows:

https://medium.com/@lmoroney_40129/tensorflow-install-for-gpu-on-linux-902b404b6b30

https://medium.com/@lmoroney_40129/installing-tensorflow-with-gpu-on-windows-10-3309fec55a00

--
You received this message because you are subscribed to the Google Groups "Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to discuss+u...@tensorflow.org.
To post to this group, send email to dis...@tensorflow.org.
To view this discussion on the web visit https://groups.google.com/a/tensorflow.org/d/msgid/discuss/10467e92-f58f-4f38-a882-c721732f1800%40tensorflow.org.

--

Laurence Moroney | Developer Advocate | lmor...@google.com | 

[tini jini]

Thank you very much.

[Prasad Gupta]

Best way I suggest is:

1) Install VMWare

2) Instal/Update Nvidia drivers

3) Install nvidia-docker

4) Download and run the docker image

5) inside docker container start jupyter notebook

start tf development

This is the best way to start TF