Help needed in installation of cuda, cudnn, tf for GTX 1650 Ti in gpu support for training on jupyter notebook

TensorFlow General Discussion
,
1

I’m currently using the following setup:

CUDA: 12.6.77
cuDNN: 9.x for CUDA 12.x (installed via tarball.zip)
TensorFlow: 2.17.0
Python: 3.11.3
GPU: GTX 1650 Ti with driver version 565.90
Operating System: Windows 11 (upgraded from Windows 10)

I’m trying to enable GPU support for model training in a Jupyter notebook running in a Conda environment. Despite multiple attempts, TensorFlow is not recognizing my GPU, as confirmed by the output of tf.config.list_physical_devices('GPU'), which returns an empty list.

Can someone help me identify what I might be doing wrong? Is this a compatibility issue with the versions I’m using, or is there something specific I need to configure on Windows to get TensorFlow to detect the GPU? Any guidance would be greatly appreciated!

2

Caution: TensorFlow 2.10 was the last TensorFlow release that supported GPU on native-Windows. Starting with TensorFlow 2.11, you will need to install TensorFlow in WSL2, or install tensorflow or tensorflow-cpu and, optionally, try the TensorFlow-DirectML-Plugin

1 Like
3

Thank you for your help. Appreciated

1 Like