A monitoring dashboard for Nvidia GPUs
Switch branches/tags
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
gpudash
.gitignore
.isort.cfg
.travis.yml
LICENCE
MANIFEST.in
Pipfile
Pipfile.lock
README.md
contributing.json
mypy.ini
pylintrc
pyproject.toml
setup.cfg
setup.py

README.md

gpudash Build status GitHub tag PyPI release

gpudash is a dashboard for monitoring Nvidia GPUs, aimed at data scientists and machine learning engineers training models on their workstations.

Installation

gpudash is a system utility and as such shouldn't be installed into the virtual environment for a given project. Instead, it is recommended to use pipsi to install the latest release from PyPI into ~/.local/bin:

pipsi install gpudash

gpudash requires Python 3.6 or later.

Usage

To start gpudash, run:

gpudash

By default the dashboard will be served on port 8050: you can specify a different port with the -p/--port flag:

gpudash --port 8000

The interval between plot updates defaults to 1 second, and can be configured with the -i/--interval flag:

gpudash --interval 2

The number of points shown on each plot defaults to 60, and can be configured with the -n/--num-points flag:

gpudash --num-points 30

Copyright

Copyright © 2018 Scott Stevenson.

gpudash is distributed under the terms of the ISC licence.