README.md

JupyterLab Notebook with Pytorch

This docker image for jupyterlab, pytorch and cuda.

Run the container

Basic usage:

docker run --rm \                       # remove the container when it exits
           -it \                        # pseudo-TTY
           -p 8888:8888 \               # port forwarding: <Host>:<Container>
           --gpus all \                 # support all gpus
           -v /local_vol:/docker_vol \  # volume: mapping local folder to container
           -e JUPYTER_TOKEN=passwd \    # Jupyter password: passwd
           -e SYS_MEM_LIMIT=16 \        # Set the monitoring memory limit
           -e SYS_CPU_LIMIT=8 \         # Set the monitoring cpu limit
           -d rodrigobaron/torch-notebook:latest

Start CPU only container:

docker run --rm -it  \
           -p 8888:8888  \
           rodrigobaron/torch-notebook:latest

Start the container with GPUs support:

docker run --rm -it  \
           --gpus all  \
           -p 8888:8888  \
           rodrigobaron/torch-notebook:latest

Extensions enabled

The containe contains simple but powerfull extensions enablind productivity boosting.
See the extension list above:

• jupyterlab-lsp: Turn jupyter notebook into IDE like, documentation inspecting, diagnostics, code complection..
• jupyterlab-git: Brings git commands and visualization tools.
• jupyterlab-system-monitor: display system infomation (memory and cpu usage).
• jupyterlab/debugger: Enable debugging and tracing.
• aquirdturtle_collapsible_headings: Make headings collapsible like Mathematica notebooks.
• jupyterlab_execute_time: Display cell timings.

Using darcula theme as default.