- Look into using a environment file for all of this
# ./environment.yml
# To use:
# $ conda env create -f environment.yml
# $ conda activate d2l
name: d2l
dependencies:
- python=3.9
- pip
- pip:
- d2l==0.17.0
- torch==1.8.1
- torchvision==0.9.1
ssh -i '.path/to/key' ubuntu@<ip>
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh && chmod u+x Miniconda3-latest-Linux-x86_64.sh && ./Miniconda3-latest-Linux-x86_64.sh && source .bashrc
conda install -y pytorch torchvision torchaudio cudatoolkit pandas scikit-learn matplotlib tqdm seaborn tensorboard -c pytorch
conda install -c conda-forge jupyterlab
conda clean -a
server-configurations:
touch .jupyter/jupyter_server_config.py
nano .jupyter/jupyter_server_config.py
then inside you can add this
c.ServerApp.open_browser = False
c.ServerApp.port = 8888 # or any other ports you'd like
c.ServerApp.ip = '*' # bind to any network interface
c.ServerApp.password =
then to get the password we
jupyter server password
nohup jupyter lab --no-browser &
then
cat .jupyter/jupyter_server_config.json
copy the hashed password into the .py file you created above
tmux new-session
tmux new -s mysession
> jupyter lab
Ctrl + b w
Session and Window Preview
Ctrl + b d
Detach from session
tmux ls
Ctrl + b s
Show all sessions
tmux a
Attach to last session
tmux a -t mysession
Attach to a session with the name mysession
go here
:8888/lab
You would have installed tensorboardX and tensorboard
pip install tensorboard
pip install tensorboardX
Then in a local terminal run
ssh -i '.\Documents\Other docs\Key\Moayad.pem' -NL 6006:localhost:6006 ubuntu@
replace the key file if needed and add the public IP address. Then on a browser navigate to
http://localhost:6006/