NASLib is a Neural Architecture Search (NAS) library. Its purpose is to facilitate NAS research for the community by providing interfaces to several state-of-the-art NAS search spaces
⚠️ This library is under construction and there is no official release yet. Feel free to play around and have a look but be aware that the APIs will be changed until we have a first release.
It is designed to be modular, extensible and easy to use.
search_space = SimpleCellSearchSpace()
optimizer = DARTSOptimizer(config)
optimizer.adapt_search_space(search_space)
trainer = Trainer(optimizer, config)
trainer.search() # Search for an architecture
trainer.evaluate() # Evaluate the best architectureFor an example file see demo.py.
For more extensive documentation see docs.
Make sure you use the latest version of pip. It makes sense to set up a virtual environment, too.
python3 -m venv naslib
source naslib/bin/activate
pip install --upgrade pip setuptools wheel
pip install cython
Clone and install.
If you plan to modify naslib consider adding the -e option for pip install.
git clone ...
cd naslib
pip install .
To validate the installation, you can run tests:
cd tests
coverage run -m unittest discover
The test coverage can be seen with coverage report.
If you use this code in your own work, please cite NASLib using the following bibtex entry:
@misc{naslib-2020,
title={NASLib: A Modular and Flexible Neural Architecture Search Library},
author={Ruchte, Michael and Zela, Arber and Siems, Julien and Grabocka, Josif and Hutter, Frank},
year={2020}, publisher={GitHub},
howpublished={\url{https://github.com/automl/NASLib}} }