/remoteDL

Prepare your remote machine to DeepLearning

Prepare your remote machine to DeepLearning(PyTorch)

Checking

Check the availability GPU in the system.

lspci | grep -i nvidia

Update system and install required packets

sudo apt-get update

sudo apt-get dist-upgrade

sudo apt install -y build-essential git libgfortran3

Install CUDA

https://developer.nvidia.com/cuda-toolkit-archive

Checking installation

nvidia-smi

Install Conda

Anaconda

Miniconda

Checking installation

conda -V

If conda: command not found then add env var to ~/.bashrc export PATH="$HOME/anaconda3/bin:$PATH".

Install PyTorch

conda install pytorch torchvision cudatoolkit=10.0 -c pytorch

Install Jupyter

conda install jupyter or conda install jupyterlab

If need prepare server to external connection

Configure Jupyter

echo c.NotebookApp.ip = '*' > ~/.jupyter/jupyter_notebook_config.py

echo c.NotebookApp.open_browser = False >> ~/.jupyter/jupyter_notebook_config.py

echo c.NotebookApp.port = 8888 >> ~/.jupyter/jupyter_notebook_config.py

jupyter notebook password

Run Jupyter

jupyter-lab --ip=0.0.0.0 --port=8888 --no-browser &

Enter in browser on client side <ip>:8888.

OR

jupyter-lab --port=9000 --no-browser &

ssh -N -f -L 8888:localhost:9000 name@server