JellyML is an open-source tool (python API and command line) for effortlessly embedding a snapshot of your code into a checkpoint of a pytorch (and pytorch lightning) model.
LGPL-3.0 License
JellyML is an open-source tool (python API and command line) for effortlessly embedding a snapshot of your code | into a checkpoint of a pytorch model. Learn more at jellyml.com
pip install jellyml
(Note that the jellyml repository is a monorepo. If you are reading this from the python package source code, go to github.com/mmulet/jellyml to see the whole repository)
python3 -m venv venv
# activate the venv ( depends on your shell and OS)
# see https://docs.python.org/3/library/venv.html
# bash
source venv/bin/activate
pip install build
cd jellyml
python -m build
pip install dist/jellyml-0.0.1-py3-none-any.whl
cd jellyml-lightning
python -m build
pip install dist/jellyml-lightning-0.0.1-py3-none-any.whl
cd client
npm install .
cd ../dev_server
npm install .
npm run build
cd client
npm install .
cd ../dev_server
npm install .
npm run build
Located in the source files in src/jellyml. Have the prefix test_. Run them as a module
cd src;
python3 -m jellyml.test_all