CentOS 7.6.1810 has a very old version of Git (v 1.8.3.1 from ~ 2013). To install a current version of Git, you have a few options:
sudo yum -y install wget perl-CPAN gettext-devel perl-devel openssl-devel zlib-devel
export VER="2.22.1"
wget https://github.com/git/git/archive/v${VER}.tar.gz
tar -xvf v${VER}.tar.gz
cd git-*
make
make install
git 2.22.1 will now be installed to ~/bin
- if you want it to be used instead of the system git, modify your .bash_profile
to have ~/bin come before /usr/bin
conda install git
As long as the conda environment is active in which git has been installed, the new git version will be used over the system git.
First make sure prerequisites are installed (not sure why yum doesn't install these as they are required):
sudo yum install gcc kernel-devel kernel-headers
Note, these steps should build kernel drivers which will take several minutes; if the install
step is quick then it likely didn't work right.
- Nvidia GPU drivers: https://www.nvidia.com/Download/driverResults.aspx/150862/en-us - download apropriate GPU driver and run the following:
sudo rpm -i nvidia-diag-driver-local-repo-rhel7-410.129-1.0-1.x86_64.rpm
sudo yum clean all
sudo yum install cuda-drivers
sudo reboot
- Install CUDA 10.0 toolkit: https://developer.nvidia.com/cuda-10.0-download-archive?target_os=Linux&target_arch=x86_64&target_distro=CentOS&target_version=7&target_type=rpmlocal - download 2 files and run the following:
sudo rpm -i cuda-repo-rhel7-10-0-local-10.0.130-410.48-1.0-1.x86_64.rpm
sudo yum clean all
sudo yum install cuda
sudo reboot
- Install CUDA 10.0 patch from CUDA URL above:
sudo rpm -i cuda-repo-rhel7-10-0-local-nvjpeg-update-1-1.0-1.x86_64.rpm
sudo yum clean all
sudo yum install cuda-nvjpeg-10-0
sudo reboot
- Install CUDNN, requires free Nvidia Account. (download from: https://developer.nvidia.com/compute/machine-learning/cudnn/secure/7.6.4.38/Production/10.0_20190923/cudnn-10.0-linux-x64-v7.6.4.38.tgz) (installation instructions: https://docs.nvidia.com/deeplearning/sdk/cudnn-install/index.html)
tar -xzvf cudnn-9.0-linux-x64-v7.tgz
sudo cp cuda/include/cudnn.h /usr/local/cuda/include
sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64
sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*
- Install TensorRT: https://developer.nvidia.com/tensorrt, installation instructions: https://docs.nvidia.com/deeplearning/sdk/tensorrt-install-guide/index.html
sudo rpm -Uvh nv-tensorrt-repo-rhel7-cuda10.0-trt5.0.2.6-ga-20181010-1-1.x86_64.rpm
sudo yum clean expire-cache
sudo yum install --nogpgcheck tensorrt
Install Anaconda:
wget https://repo.anaconda.com/archive/Anaconda3-2019.10-Linux-x86_64.sh
bash Anaconda3-2019.10-Linux-x86_64.sh
Create an environment - envname and python version as desired:
conda create -n envname python=3.6
Then switch to new environment - name as desired:
conda activate envname
Items to install in each conda environment:
conda install seaborn pytest statsmodels scikit-learn git
pip install gpustat
conda install -c conda-forge altair vega_datasets jupyterlab
conda install -c conda-forge jupytext nodejs
Requires all the GPU libraries and drivers to be first installed - see https://www.tensorflow.org/install/gpu Once that is completed:
conda create -n tensorflow python=3.6
conda activate tensorflow
pip install tensorflow
Follow instructions at https://keras.io/#installation
By default Keras uses TensorFlow as the tensor backend. If you successfully get TensorFlow installed and working, you just have to then run:
conda create -n keras python=3.6
conda activate keras
pip install tensorflow
pip install keras
See instructions and additional options available at https://pytorch.org/get-started/locally/
conda create -n pytorch python=3.7
conda activate pytorch
conda install pytorch torchvision cudatoolkit=10.1 -c pytorch