Check the highest cuda version supported by this computer: nvidia-smi
Modify the conda source in ~/.condarc:
show_channel_urls: true
ssl_verify: false
channels:
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/bioconda/
- defaults#The following command can view the relationship between each sub-version of cudnn8 and the cuda version: conda search cudnn=8 --info
conda create -n tensorflow_gpu3 python=3.7 cudnn=7.6.5.32 cudatoolkit=10.2.89
conda activate tensorflow_gpu3
pip install tensorflow-gpu==2.11.0
Verify that the installation was successful
Enter the python environment:
import tensorflow as tf
tf.config.list_physical_devices('GPU')