Hey guys. sorry for this dumb question I’m really newbie at this.
In tensorflow instruction at software requirements section mentioned to install cuda toolkit =11.8 and cudnn SDK=8.6.0 but in the GPU setup section mentioned this code
conda install -c conda-forge cudatoolkit=11.2 cudnn=8.1.0
and in the build tab cuda toolkit=11.2 and cudnn=8.1 were mentioned for tensorflow=2.10
my question is which version of these packages should be installed. cudatoolkit=11.8 or 11.2
I’m using:
Windows 10
Nvidia Geforce 1050 ti
GPU driver 532.03
1 Like
@Robert_Ford,
Welcome to the Tensorflow Forum,
GPU support on native-Windows is only available for 2.10 or earlier versions, starting in TF 2.11, CUDA build is not supported for Windows.
Hence, you should use CUDA 11.2
to use the GPU on Windows; otherwise, you will need to install TensorFlow in WSL2 or install tensorflow
or tensorflow-cpu
and, optionally, try the TensorFlow-DirectML-Plugin
conda install -c conda-forge cudatoolkit=11.2 cudnn=8.1.0
python -m pip install tensorflow=2.10
python -c "import tensorflow as tf; print(tf.config.list_physical_devices('GPU'))"
Thank you!
3 Likes