Hey,
I’m trying to get TensorFlow to work on my new GPU, so I downloaded and installed cuDNN 8.6.0, CUDA Toolkit 11.8 on my computer, and added the paths to the environment variable PATH. On Conda, I installed Python 3.10 with pip 20.3, along with cudatoolkit 11.2 and cudnn 8.1.0. I didn’t get any errors of any kind, but when I run print(tf.config.list_physical_devices(‘GPU’)), I get an empty list. What can I do?
PS : I have an windows 11 and a rtx 4070
Hi @RedDefaltos, I have tried to follow these steps for installing Tensorflow with GPU on windows 11 and can able to detect the GPU
conda create --name tensorflow-gpu Python=3.9
conda install -c conda-forge cudatoolkit=11.2 cudnn=8.1.0
python -m pip install "tensorflow<2.11"
python -c "import tensorflow as tf; print('Tensorflow_Version: ',TF.__version__); print(tf.config.list_physical_devices('GPU'))"
From the steps you have given I can see that you are installing Python versions 3.9 and 3.10 in the same environment that might be causing the conflict. Could you please try to create a new environment and follow the mentioned steps. Thank You.