Skip to content

Commit

Permalink
updates torchani-ase-jupyter Dockerfile and metadata
Browse files Browse the repository at this point in the history
  • Loading branch information
SinaMostafanejad committed Jul 11, 2023
1 parent 05b58e9 commit f92b35c
Show file tree
Hide file tree
Showing 2 changed files with 23 additions and 8 deletions.
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,20 @@ RUN mamba install numpy \
scipy \
matplotlib \
ase=3.22.1 \
-c conda-forge -yq \
jupyterlab \
-c conda-forge -y \
&& pip install --no-cache-dir torchani==2.2.3 \
&& conda clean -afy
&& conda clean -afy \
&& find ${CONDA_PREFIX} -follow -type f -name '*.a' -delete \
&& find ${CONDA_PREFIX} -follow -type f -name '*.pyc' -delete \
&& find ${CONDA_PREFIX} -follow -type f -name '*.js.map' -delete

CMD [ \
"jupyter", \
"lab", \
"--notebook-dir=/home", \
"--no-browser", \
"--allow-root", \
"--ip=*", \
"--port=8888" \
]
Original file line number Diff line number Diff line change
Expand Up @@ -17,19 +17,20 @@
}
],
"docker_pull_command": "docker pull molssi/torchani223-cu117-ase-mamba141-jupyter:5.3.2023",
"docker_run_command": "docker run -it --name torchani -v $(pwd):/home molssi/torchani223-cu117-ase-mamba141-jupyter:5.3.2023 /bin/bash",
"note": "By default, the ``-v $(pwd):/home`` option mounts the current working directory to ``/home`` in your running container. Doing so, the contents of your current working directory become available in your running container. If you do not wish this to happen, you may simply remove or change this option. For NVIDIA GPU support with nvidia containers, add the ``--runtime nvidia --gpus all`` flags to the previous container run command and then run ``nvidia-smi`` to make sure all available GPUs on the docker host are visible inside the docker container.",
"docker_run_command": "docker run -it --name torchani -v $(pwd):/home -p 8888:8888 molssi/torchani223-cu117-ase-mamba141-jupyter:5.3.2023",
"note": "By default, the ``-v $(pwd):/home`` option mounts the current working directory to ``/home`` in your running container. Doing so, the contents of your current working directory become available in your running container. If you do not wish this to happen, you may simply remove or change this option. By default, Jupyter server communicates to your computed through port ``8888``. Make sure to include ``-p 8888:8888`` in the ``docker run`` command. For NVIDIA GPU support with nvidia containers, add the ``--runtime nvidia --gpus all`` flags to the previous container run command and then run ``nvidia-smi`` to make sure all available GPUs on the docker host are visible inside the docker container.",
"image_specifications":[
{
"OS/Arch": "debian:bullseye-slim (linux/amd64)",
"Users (UID)": "molssi (1000)",
"Groups (GID)": "molssi (1000)",
"Users (UID)": "root (0)",
"Groups (GID)": "root (0)",
"Environment variables": [
{
"PATH":"/opt/conda/bin"
"CONDA_PREFIX": "/opt/conda",
"PATH": "${CONDA_PREFIX}/bin:$PATH"
}
],
"Volumes": "$(pwd):$HOME",
"Volumes": "None",
"Network": "None",
"Extras": [
{
Expand Down

0 comments on commit f92b35c

Please sign in to comment.